Arrow Research search

Author name cluster

Razvan Pascanu

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

69 papers
2 author rows

Possible papers

69

TMLR Journal 2026 Journal Article

A Survey on Over-smoothing and Over-squashing: Unified Propagation Perspectives on Graph Neural Networks and Transformers

  • Alvaro Arroyo
  • Federico Barbero
  • Hugh Blayney
  • Michael M. Bronstein
  • Xiaowen Dong
  • Pietro Lio
  • Razvan Pascanu
  • Pierre Vandergheynst

Decoder-Transformers have achieved remarkable success and have laid the groundwork for the development of Large Language Models (LLMs). At the core of these models is the self-attention matrix, which allows different tokens to interact with each other. This process is remarkably similar to the message-passing mechanism used in Graph Neural Networks (GNNs), and as such decoder-Transformers suffer many of the optimization difficulties studied extensively in the GNN literature. In this paper, we present a unified graph perspective that bridges the theoretical understanding of decoder-Transformers and GNNs. We systematically examine how well-known phenomena in GNNs, such as over-smoothing and over-squashing, directly manifest as analogous issues like rank collapse and representational collapse in deep Transformer architectures. By interpreting Transformers' self-attention as a learned adjacency operator, we reveal shared underlying principles governing signal propagation and demonstrate how insights from one field can illuminate challenges and solutions in the other. We analyze the role of architectural components like residual connections, normalization, and causal masking in these issues. We aim to provide a framework for understanding how information flows through deep learning models that perform sequence mixing through an adjacency operator, and to highlight areas for cross-pollination of research, as well as to provide a comprehensive reference for researchers interested in the underpinnings of these architectures.

PDF Details

TMLR Journal 2026 Journal Article

TRecViT: A Recurrent Video Transformer

  • Viorica Patraucean
  • Xu Owen He
  • Joseph Heyward
  • Chuhan ZHANG
  • Mehdi S. M. Sajjadi
  • George-Cristian Muraru
  • Artem Zholus
  • Mahdi Karami

We propose a novel block for causal video modelling. It relies on a time - space - channel factorisation with dedicated blocks for each dimension: gated linear recurrent units (LRUs) perform information mixing over time, self-attention layers perform mixing over space, and MLPs over channels. The resulting architecture TRecViT is causal and shows strong performance on sparse and dense tasks, trained in supervised or self-supervised regimes, being the first causal video model in the state-space models family. Notably, our model outperforms or is on par with the popular (non-causal) ViViT-L model on large scale video datasets (SSv2, Kinetics400), while having 3x less parameters, 12x smaller memory footprint, and 5x lower FLOPs count than the full self-attention ViViT, with an inference throughput of about 300 frames per second, running comfortably in real-time. When compared with causal transformer-based models (TSM, RViT) and other recurrent models like LSTM, TRecViT obtains state-of-the-art results on the challenging SSv2 dataset. Code and checkpoints are available online \url{https://github.com/google-deepmind/trecvit}.

PDF Details

ICML Conference 2025 Conference Paper

A Large Recurrent Action Model: xLSTM enables Fast Inference for Robotics Tasks

  • Thomas Schmied
  • Thomas Adler
  • Vihang Prakash Patil
  • Maximilian Beck
  • Korbinian Pöppel
  • Johannes Brandstetter
  • Günter Klambauer
  • Razvan Pascanu

In recent years, there has been a trend in the field of Reinforcement Learning (RL) towards large action models trained offline on large-scale datasets via sequence modeling. Existing models are primarily based on the Transformer architecture, which results in powerful agents. However, due to slow inference times, Transformer-based approaches are impractical for real-time applications, such as robotics. Recently, modern recurrent architectures, such as xLSTM and Mamba, have been proposed that exhibit parallelization benefits during training similar to the Transformer architecture while offering fast inference. In this work, we study the aptitude of these modern recurrent architectures for large action models. Consequently, we propose a Large Recurrent Action Model (LRAM) with an xLSTM at its core that comes with linear-time inference complexity and natural sequence length extrapolation abilities. Experiments on 432 tasks from 6 domains show that LRAM compares favorably to Transformers in terms of performance and speed.

Details

ICLR Conference 2025 Conference Paper

Attention as a Hypernetwork

  • Simon Schug
  • Seijin Kobayashi
  • Yassir Akram
  • João Sacramento
  • Razvan Pascanu

Transformers can under some circumstances generalize to novel problem instances whose constituent parts might have been encountered during training, but whose compositions have not. What mechanisms underlie this ability for compositional generalization? By reformulating multi-head attention as a hypernetwork, we reveal that a composable, low-dimensional latent code specifies key-query specific operations. We find empirically that this latent code is predictive of the subtasks the network performs on unseen task compositions, revealing that latent codes acquired during training are reused to solve unseen problem instances. To further examine the hypothesis that the intrinsic hypernetwork of multi-head attention supports compositional generalization, we ablate whether making the hypernetwork-generated linear value network nonlinear strengthens compositionality. We find that this modification improves compositional generalization on abstract reasoning tasks. In particular, we introduce a symbolic version of the Raven's Progressive Matrices human intelligence test, which gives us precise control over the problem compositions encountered during training and evaluation. We demonstrate on this task how scaling model size and data enables compositional generalization in transformers and gives rise to a functionally structured latent space.

Details

TMLR Journal 2025 Journal Article

Maxwell's Demon at Work: Efficient Pruning by Leveraging Saturation of Neurons

  • Simon Dufort-Labbé
  • Pierluca D'Oro
  • Evgenii Nikishin
  • Irina Rish
  • Pierre-Luc Bacon
  • Razvan Pascanu
  • Aristide Baratin

When training neural networks, dying neurons —units becoming inactive or saturated— are traditionally seen as harmful. This paper sheds new light on this phenomenon. By exploring the impact of various hyperparameter configurations on dying neurons during training, we gather insights on how to improve upon sparse training approaches to pruning. We introduce Demon Pruning (DemP), a method that controls the proliferation of dead neurons through a combination of noise injection on active units and a one-cycled schedule regularization strategy, dynamically leading to network sparsity. Experiments on CIFAR-10 and ImageNet datasets demonstrate that DemP outperforms existing dense-to-sparse structured pruning methods, achieving better accuracy-sparsity tradeoffs while speeding up training up to 3.56$\times$. These findings provide a novel perspective on dying neurons as a resource for efficient model compression and optimization.

PDF Details

NeurIPS Conference 2025 Conference Paper

Meta-learning how to Share Credit among Macro-Actions

  • Ionel-Alexandru Hosu
  • Traian Rebedea
  • Razvan Pascanu

One proposed mechanism to improve exploration in reinforcement learning is the use of macro-actions, a form of temporal abstractions over actions. Paradoxically though, in many scenarios the naive addition of macro-actions does not lead to better exploration, but rather the opposite. In this work, we argue that the difficulty stems from the trade-offs between reducing the average number of decisions per episode versus increasing the size of the action space. Namely, one typically treats each potential macro-action as independent and atomic, hence strictly increasing the search space and making typical exploration strategies inefficient. To address this problem we propose a novel regularization term that exploits the relationship between actions and macro-actions to improve the credit assignment mechanism reducing the effective dimension of the action space and therefore improving exploration. The term relies on a similarity matrix that is meta-learned jointly with learning the desired policy. We empirically validate our strategy looking at macro-actions in Atari games, and the StreetFighter II environment. Our results show significant improvements over the Rainbow-DQN baseline in all environments. Additionally, we show that the macro-action similarity is transferable to other environments with similar dynamics. We believe this work is a small but important step towards understanding how the similarity-imposed geometry on the action space can be exploited to improve credit assignment and exploration, therefore making learning more efficient.

PDF Details

NeurIPS Conference 2025 Conference Paper

Plasticity as the Mirror of Empowerment

  • David Abel
  • Michael Bowling
  • Andre Barreto
  • Will Dabney
  • Shi Dong
  • Steven Hansen
  • Anna Harutyunyan
  • Khimya Khetarpal

Agents are minimally entities that are influenced by their past observations and act to influence future observations. This latter capacity is captured by empowerment, which has served as a vital framing concept across artificial intelligence and cognitive science. This former capacity, however, is equally foundational: In what ways, and to what extent, can an agent be influenced by what it observes? In this paper, we ground this concept in a universal agent-centric measure that we refer to as plasticity, and reveal a fundamental connection to empowerment. Following a set of desiderata on a suitable definition, we define plasticity using a new information-theoretic quantity we call the generalized directed information. We show that this new quantity strictly generalizes the directed information introduced by Massey (1990) while preserving all of its desirable properties. Under this definition, we find that plasticity is well thought of as the mirror of empowerment: The two concepts are defined using the same measure, with only the direction of influence reversed. Our main result establishes a tension between the plasticity and empowerment of an agent, suggesting that agent design needs to be mindful of both characteristics. We explore the implications of these findings, and suggest that plasticity, empowerment, and their relationship are essential to understanding agency.

PDF Details

NeurIPS Conference 2025 Conference Paper

RAT: Bridging RNN Efficiency and Attention Accuracy via Chunk-based Sequence Modeling

  • Xiuying Wei
  • Anunay Yadav
  • Razvan Pascanu
  • Caglar Gulcehre

Transformers have become the cornerstone of modern large-scale language models, but their reliance on softmax attention poses a computational bottleneck at both training and inference. Recurrent models offer high efficiency, but compressing the full sequence into a fixed-size and holistic representation can suffer from memory degradation in long contexts and limit fine-grained retrieval. To address this, we propose RAT, an intermediate design that bridges the efficiency of RNNs and capacity of attention. RAT partitions the input into chunks, applies recurrence within each chunk for local dependencies, and softmax-based attention across chunks for long-range interactions. This design mitigates memory degradation and enables direct access to distant tokens, while retaining computational efficiency. Empirically, with a chunk size of 16, the RAT block achieves a $7\times$ improvement in training speed for 100K sequence length and $9\times$ in generation at the 4K position, while maintaining similar performance compared to standard attention. We demonstrate this by training 1. 3B parameter models from scratch and performing large-scale evaluations, including short- and long-context benchmarks, as well as supervised fine-tuning (SFT). We further propose a hybrid architecture that interleaves RAT with local attention. By combining efficient long-range modeling with strong local interactions, this hybrid design not only improves inference speed and reduces cache memory usage, but also consistently enhances performance and shows the overall best results. Code is available at \url{https: //github. com/CLAIRE-Labo/RAT}.

PDF Details

ICLR Conference 2025 Conference Paper

Round and Round We Go! What makes Rotary Positional Encodings useful?

  • Federico Barbero
  • Alex Vitvitskyi
  • Christos Perivolaropoulos
  • Razvan Pascanu
  • Petar Velickovic

Positional Encodings (PEs) are a critical component of Transformer-based Large Language Models (LLMs), providing the attention mechanism with important sequence-position information. One of the most popular types of encoding used today in LLMs are Rotary Positional Encodings (RoPE), that rotate the queries and keys based on their relative distance. A common belief is that RoPE is useful because it helps to decay token dependency as relative distance increases. In this work, we argue that this is unlikely to be the core reason. We study the internals of a trained Gemma 7B model to understand how RoPE is being used at a mechanical level. We find that Gemma learns to use RoPE to construct robust `positional' attention patterns by exploiting the highest frequencies. We also find that, in general, Gemma greatly prefers to use the lowest frequencies of RoPE, which we suspect are used to carry semantic information. We mathematically prove interesting behaviours of RoPE and conduct experiments to verify our findings, proposing a modification of RoPE that fixes some highlighted issues and improves performance. We believe that this work represents an interesting step in better understanding PEs in LLMs, which we believe holds crucial value for scaling LLMs to large sizes and context lengths.

Details

ICML Conference 2025 Conference Paper

Softmax is not Enough (for Sharp Size Generalisation)

  • Petar Velickovic
  • Christos Perivolaropoulos
  • Federico Barbero
  • Razvan Pascanu

A key property of reasoning systems is the ability to make sharp decisions on their input data. For contemporary AI systems, a key carrier of sharp behaviour is the softmax function, with its capability to perform differentiable query-key lookups. It is a common belief that the predictive power of networks leveraging softmax arises from "circuits" which sharply perform certain kinds of computations consistently across many diverse inputs. However, for these circuits to be robust, they would need to generalise well to arbitrary valid inputs. In this paper, we dispel this myth: even for tasks as simple as finding the maximum key, any learned circuitry must disperse as the number of items grows at test time. We attribute this to a fundamental limitation of the softmax function to robustly approximate sharp functions with increasing problem size, prove this phenomenon theoretically, and propose adaptive temperature as an ad-hoc technique for improving the sharpness of softmax at inference time.

Details

NeurIPS Conference 2024 Conference Paper

Building on Efficient Foundations: Effective Training of LLMs with Structured Feedforward Layers

  • Xiuying Wei
  • Skander Moalla
  • Razvan Pascanu
  • Caglar Gulcehre

State-of-the-art results in large language models (LLMs) often rely on scale, whichbecomes computationally expensive. This has sparked a research agenda to reducethese models’ parameter counts and computational costs without significantlyimpacting their performance. Our study focuses on transformer-based LLMs, specifically targeting the computationally intensive feedforward networks (FFNs), which are less studied than attention blocks. We consider three structured linearparameterizations of the FFN using efficient low-rank and block-diagonal matrices. In contrast to many previous works that examined these approximations, our studyi) explores these structures from a training-from-scratch perspective, ii) scales upto 1. 3B parameters, and iii) is conducted within recent Transformer-based LLMsrather than convolutional architectures. We demonstrate that these structures canlead to actual computational gains in various scenarios, including online decodingwhen using a pre-merge technique. Additionally, we propose a novel trainingregime, called self-guided training, aimed at improving the poor training dynamicsthat these approximations exhibit when used from initialization. Interestingly, the scaling performance of structured matrices is explored, revealing steepercurves in scaling training FLOPs, along with a favorable scaling trend in theovertraining regime. Specifically, we show that wide and structured networkscan utilize training FLOPs more efficiently, with fewer parameters and lowerloss than dense models at their optimal trade-off. Our code is available athttps: //github. com/CLAIRE-Labo/StructuredFFN/tree/main.

PDF Details DOI

TMLR Journal 2024 Journal Article

Continual Learning: Applications and the Road Forward

  • Eli Verwimp
  • Rahaf Aljundi
  • Shai Ben-David
  • Matthias Bethge
  • Andrea Cossu
  • Alexander Gepperth
  • Tyler L. Hayes
  • Eyke Hüllermeier

Continual learning is a subfield of machine learning, which aims to allow machine learning models to continuously learn on new data, by accumulating knowledge without forgetting what was learned in the past. In this work, we take a step back, and ask: "Why should one care about continual learning in the first place?". We set the stage by examining recent continual learning papers published at four major machine learning conferences, and show that memory-constrained settings dominate the field. Then, we discuss five open problems in machine learning, and even though they might seem unrelated to continual learning at first sight, we show that continual learning will inevitably be part of their solution. These problems are model editing, personalization and specialization, on-device learning, faster (re-)training and reinforcement learning. Finally, by comparing the desiderata from these unsolved problems and the current assumptions in continual learning, we highlight and discuss four future directions for continual learning research. We hope that this work offers an interesting perspective on the future of continual learning, while displaying its potential value and the paths we have to pursue in order to make it successful. This work is the result of the many discussions the authors had at the Dagstuhl seminar on Deep Continual Learning, in March 2023.

PDF Details

ICLR Conference 2024 Conference Paper

Discovering modular solutions that generalize compositionally

  • Simon Schug
  • Seijin Kobayashi
  • Yassir Akram
  • Maciej Wolczyk
  • Alexandra Maria Proca
  • Johannes von Oswald
  • Razvan Pascanu
  • João Sacramento

Many complex tasks can be decomposed into simpler, independent parts. Discovering such underlying compositional structure has the potential to enable compositional generalization. Despite progress, our most powerful systems struggle to compose flexibly. It therefore seems natural to make models more modular to help capture the compositional nature of many tasks. However, it is unclear under which circumstances modular systems can discover hidden compositional structure. To shed light on this question, we study a teacher-student setting with a modular teacher where we have full control over the composition of ground truth modules. This allows us to relate the problem of compositional generalization to that of identification of the underlying modules. In particular we study modularity in hypernetworks representing a general class of multiplicative interactions. We show theoretically that identification up to linear transformation purely from demonstrations is possible without having to learn an exponential number of module combinations. We further demonstrate empirically that under the theoretically identified conditions, meta-learning from finite data can discover modular policies that generalize compositionally in a number of complex environments.

Details

ICML Conference 2024 Conference Paper

Fine-tuning Reinforcement Learning Models is Secretly a Forgetting Mitigation Problem

  • Maciej Wolczyk
  • Bartlomiej Cupial
  • Mateusz Ostaszewski
  • Michal Bortkiewicz
  • Michal Zajac 0005
  • Razvan Pascanu
  • Lukasz Kucinski
  • Piotr Milos

Fine-tuning is a widespread technique that allows practitioners to transfer pre-trained capabilities, as recently showcased by the successful applications of foundation models. However, fine-tuning reinforcement learning (RL) models remains a challenge. This work conceptualizes one specific cause of poor transfer, accentuated in the RL setting by the interplay between actions and observations: forgetting of pre-trained capabilities. Namely, a model deteriorates on the state subspace of the downstream task not visited in the initial phase of fine-tuning, on which the model behaved well due to pre-training. This way, we lose the anticipated transfer benefits. We identify conditions when this problem occurs, showing that it is common and, in many cases, catastrophic. Through a detailed empirical analysis of the challenging NetHack and Montezuma’s Revenge environments, we show that standard knowledge retention techniques mitigate the problem and thus allow us to take full advantage of the pre-trained capabilities. In particular, in NetHack, we achieve a new state-of-the-art for neural models, improving the previous best score from $5$K to over $10$K points in the Human Monk scenario.

Details

ICML Conference 2024 Conference Paper

Improving fine-grained understanding in image-text pre-training

  • Ioana Bica
  • Anastasija Ilic
  • Matthias Bauer
  • Goker Erdogan
  • Matko Bosnjak
  • Christos Kaplanis
  • Alexey A. Gritsenko
  • Matthias Minderer

We introduce SPARse fine-grained Contrastive alignment (SPARC), a simple method for pretraining more fine-grained multimodal representations from image-text pairs. Given that multiple image patches often correspond to single words, we propose to learn a grouping of image patches for every token in the caption. To achieve this, we use a sparse similarity metric between image patches and language tokens and compute for each token a language-grouped vision embedding as the weighted average of patches. The token and language-grouped vision embeddings are then contrasted through a fine-grained sequence-wise loss that only depends on individual samples and does not require other batch samples as negatives, i. e. , more detailed information is encoded in a computationally inexpensive way. SPARC combines this fine-grained loss with a contrastive loss between global image and text embeddings to learn representations that simultaneously encode global and local information. We thoroughly evaluate SPARC and show improved performance over competing approaches both on image-level tasks relying on coarse-grained information, e. g. classification, as well as region-level tasks relying on fine-grained information, e. g. , retrieval, object detection, segmentation while also improving model faithfulness and captioning in foundational vision-language models.

Details

ICLR Conference 2024 Conference Paper

Kalman Filter for Online Classification of Non-Stationary Data

  • Michalis K. Titsias
  • Alexandre Galashov
  • Amal Rannen-Triki
  • Razvan Pascanu
  • Yee Whye Teh
  • Jörg Bornschein

In Online Continual Learning (OCL) a learning system receives a stream of data and sequentially performs prediction and training steps. Key challenges in OCL include automatic adaptation to the specific non-stationary structure of the data and maintaining appropriate predictive uncertainty. To address these challenges we introduce a probabilistic Bayesian online learning approach that utilizes a (possibly pretrained) neural representation and a state space model over the linear predictor weights. Non-stationarity in the linear predictor weights is modelled using a “parameter drift” transition density, parametrized by a coefficient that quantifies forgetting. Inference in the model is implemented with efficient Kalman filter recursions which track the posterior distribution over the linear weights, while online SGD updates over the transition dynamics coefficient allow for adaptation to the non-stationarity observed in the data. While the framework is developed assuming a linear Gaussian model, we extend it to deal with classification problems and for fine-tuning the deep learning representation. In a set of experiments in multi-class classification using data sets such as CIFAR-100 and CLOC we demonstrate the model's predictive ability and its flexibility in capturing non-stationarity.

Details

NeurIPS Conference 2024 Conference Paper

No Representation, No Trust: Connecting Representation, Collapse, and Trust Issues in PPO

  • Skander Moalla
  • Andrea Miele
  • Daniil Pyatko
  • Razvan Pascanu
  • Caglar Gulcehre

Reinforcement learning (RL) is inherently rife with non-stationarity since the states and rewards the agent observes during training depend on its changing policy. Therefore, networks in deep RL must be capable of adapting to new observations and fitting new targets. However, previous works have observed that networks trained under non-stationarity exhibit an inability to continue learning, termed loss of plasticity, and eventually a collapse in performance. For off-policy deep value-based RL methods, this phenomenon has been correlated with a decrease in representation rank and the ability to fit random targets, termed capacity loss. Although this correlation has generally been attributed to neural network learning under non-stationarity, the connection to representation dynamics has not been carefully studied in on-policy policy optimization methods. In this work, we empirically study representation dynamics in Proximal Policy Optimization (PPO) on the Atari and MuJoCo environments, revealing that PPO agents are also affected by feature rank deterioration and capacity loss. We show that this is aggravated by stronger non-stationarity, ultimately driving the actor's performance to collapse, regardless of the performance of the critic. We ask why the trust region, specific to methods like PPO, cannot alleviate or prevent the collapse and find a connection between representation collapse and the degradation of the trust region, one exacerbating the other. Finally, we present Proximal Feature Optimization (PFO), a novel auxiliary loss that, along with other interventions, shows that regularizing the representation dynamics mitigates the performance collapse of PPO agents. Code and run histories are available at https: //github. com/CLAIRE-Labo/no-representation-no-trust.

PDF Details DOI

EWRL Workshop 2024 Workshop Paper

No Representation, No Trust: Connecting Representation, Collapse, and Trust Issues in PPO

  • Skander Moalla
  • Andrea Miele
  • Razvan Pascanu
  • Caglar Gulcehre

Reinforcement learning (RL) is inherently rife with non-stationarity since the states and rewards the agent observes during training depend on its changing policy. Therefore, networks in deep RL must be capable of adapting to new observations and fitting new targets. However, previous works have observed that networks in off-policy deep value-based methods exhibit a decrease in representation rank, often correlated with an inability to continue learning or a collapse in performance. Although this phenomenon has generally been attributed to neural network learning under non-stationarity, it has been overlooked in on-policy policy optimization methods which are often thought capable of training indefinitely. In this work, we empirically study representation dynamics in Proximal Policy Optimization (PPO) on the Atari and MuJoCo environments, revealing that PPO agents are also affected by feature rank deterioration and loss of plasticity. We show that this is aggravated with stronger non-stationarity, ultimately driving the actor's performance to collapse, regardless of the performance of the critic. We ask why the trust region, specific to methods like PPO, cannot alleviate or prevent the collapse. We find that there is a connection between representation collapse and the degradation of the trust region, one exacerbating the other, and present Proximal Feature Optimization (PFO), a novel auxiliary loss that, along with other interventions, shows that regularizing the representation dynamics improves the performance of PPO agents.

PDF

NeurIPS Conference 2024 Conference Paper

Non-Stationary Learning of Neural Networks with Automatic Soft Parameter Reset

  • Alexandre Galashov
  • Michalis Titsias
  • András György
  • Clare Lyle
  • Razvan Pascanu
  • Yee Whye Teh
  • Maneesh Sahani

Neural networks are most often trained under the assumption that data come from a stationary distribution. However, settings in which this assumption is violated are of increasing importance; examples include supervised learning with distributional shifts, reinforcement learning, continual learning and non-stationary contextual bandits. Here, we introduce a novel learning approach that automatically models and adapts to non-stationarity by linking parameters through an Ornstein-Uhlenbeck process with an adaptive drift parameter. The adaptive drift draws the parameters towards the distribution used at initialisation, so the approach can be understood as a form of soft parameter reset. We show empirically that our approach performs well in non-stationary supervised, and off-policy reinforcement learning settings.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Normalization and effective learning rates in reinforcement learning

  • Clare Lyle
  • Zeyu Zheng
  • Khimya Khetarpal
  • James Martens
  • Hado van Hasselt
  • Razvan Pascanu
  • Will Dabney

Normalization layers have recently experienced a renaissance in the deep reinforcement learning and continual learning literature, with several works highlighting diverse benefits such as improving loss landscape conditioning and combatting overestimation bias. However, normalization brings with it a subtle but important side effect: an equivalence between growth in the norm of the network parameters and decay in the effective learning rate. This becomes problematic in continual learning settings, where the resulting learning rate schedule may decay to near zero too quickly relative to the timescale of the learning problem. We propose to make the learning rate schedule explicit with a simple re-parameterization which we call Normalize-and-Project (NaP), which couples the insertion of normalization layers with weight projection, ensuring that the effective learning rate remains constant throughout training. This technique reveals itself as a powerful analytical tool to better understand learning rate schedules in deep reinforcement learning, and as a means of improving robustness to nonstationarity in synthetic plasticity loss benchmarks along with both the single-task and sequential variants of the Arcade Learning Environment. We also show that our approach can be easily applied to popular architectures such as ResNets and transformers while recovering and in some cases even slightly improving the performance of the base model in common stationary benchmarks.

PDF Details DOI

TMLR Journal 2024 Journal Article

Promoting Exploration in Memory-Augmented Adam using Critical Momenta

  • Pranshu Malviya
  • Goncalo Mordido
  • Aristide Baratin
  • Reza Babanezhad Harikandeh
  • Jerry Huang
  • Simon Lacoste-Julien
  • Razvan Pascanu
  • Sarath Chandar

Adaptive gradient-based optimizers, notably Adam, have left their mark in training large-scale deep learning models, offering fast convergence and robustness to hyperparameter settings. However, they often struggle with generalization, attributed to their tendency to converge to sharp minima in the loss landscape. To address this, we propose a new memory-augmented version of Adam that encourages exploration towards flatter minima by incorporating a buffer of critical momentum terms during training. This buffer prompts the optimizer to overshoot beyond narrow minima, promoting exploration. Through comprehensive analysis in simple settings, we illustrate the efficacy of our approach in increasing exploration and bias towards flatter minima. We empirically demonstrate that it can improve model performance for image classification on ImageNet and CIFAR10/100, language modelling on Penn Treebank, and online learning tasks on TinyImageNet and 5-dataset. Our code is available at https://github.com/chandar-lab/CMOptimizer.

PDF Details

NeurIPS Conference 2024 Conference Paper

Transformers need glasses! Information over-squashing in language tasks

  • Federico Barbero
  • Andrea Banino
  • Steven Kapturowski
  • Dharshan Kumaran
  • João G. Araújo
  • Alex Vitvitskyi
  • Razvan Pascanu
  • Petar Veličković

We study how information propagates in decoder-only Transformers, which are the architectural foundation of most existing frontier large language models (LLMs). We rely on a theoretical signal propagation analysis---specifically, we analyse the representations of the last token in the final layer of the Transformer, as this is the representation used for next-token prediction. Our analysis reveals a representational collapse phenomenon: we prove that certain distinct pairs of inputs to the Transformer can yield arbitrarily close representations in the final token. This effect is exacerbated by the low-precision floating-point formats frequently used in modern LLMs. As a result, the model is provably unable to respond to these sequences in different ways---leading to errors in, e. g. , tasks involving counting or copying. Further, we show that decoder-only Transformer language models can lose sensitivity to specific tokens in the input, which relates to the well-known phenomenon of over-squashing in graph neural networks. We provide empirical evidence supporting our claims on contemporary LLMs. Our theory points to simple solutions towards ameliorating these issues.

PDF Details DOI

ICML Conference 2024 Conference Paper

Universality of Linear Recurrences Followed by Non-linear Projections: Finite-Width Guarantees and Benefits of Complex Eigenvalues

  • Antonio Orvieto
  • Soham De
  • Çaglar Gülçehre
  • Razvan Pascanu
  • Samuel L. Smith

Deep neural networks based on linear RNNs interleaved with position-wise MLPs are gaining traction as competitive approaches for sequence modeling. Examples of such architectures include state-space models (SSMs) like S4, LRU, and Mamba: recently proposed models that achieve promising performance on text, genetics, and other data that require long-range reasoning. Despite experimental evidence highlighting these architectures’ effectiveness and computational efficiency, their expressive power remains relatively unexplored, especially in connection to specific choices crucial in practice - e. g. , carefully designed initialization distribution and potential use of complex numbers. In this paper, we show that combining MLPs with both real or complex linear diagonal recurrences leads to arbitrarily precise approximation of regular causal sequence-to-sequence maps. At the heart of our proof, we rely on a separation of concerns: the linear RNN provides a lossless encoding of the input sequence, and the MLP performs non-linear processing on this encoding. While we show that real diagonal linear recurrences are enough to achieve universality in this architecture, we prove that employing complex eigenvalues near unit disk - i. e. , empirically the most successful strategy in S4 - greatly helps the RNN in storing information. We connect this finding with the vanishing gradient issue and provide experiments supporting our claims.

Details

NeurIPS Conference 2023 Conference Paper

Deep Reinforcement Learning with Plasticity Injection

  • Evgenii Nikishin
  • Junhyuk Oh
  • Georg Ostrovski
  • Clare Lyle
  • Razvan Pascanu
  • Will Dabney
  • Andre Barreto

A growing body of evidence suggests that neural networks employed in deep reinforcement learning (RL) gradually lose their plasticity, the ability to learn from new data; however, the analysis and mitigation of this phenomenon is hampered by the complex relationship between plasticity, exploration, and performance in RL. This paper introduces plasticity injection, a minimalistic intervention that increases the network plasticity without changing the number of trainable parameters or biasing the predictions. The applications of this intervention are two-fold: first, as a diagnostic tool — if injection increases the performance, we may conclude that an agent's network was losing its plasticity. This tool allows us to identify a subset of Atari environments where the lack of plasticity causes performance plateaus, motivating future studies on understanding and combating plasticity loss. Second, plasticity injection can be used to improve the computational efficiency of RL training if the agent has to re-learn from scratch due to exhausted plasticity or by growing the agent's network dynamically without compromising performance. The results on Atari show that plasticity injection attains stronger performance compared to alternative methods while being computationally efficient.

PDF Details

NeurIPS Conference 2023 Conference Paper

Learning to Modulate pre-trained Models in RL

  • Thomas Schmied
  • Markus Hofmarcher
  • Fabian Paischer
  • Razvan Pascanu
  • Sepp Hochreiter

Reinforcement Learning (RL) has been successful in various domains like robotics, game playing, and simulation. While RL agents have shown impressive capabilities in their specific tasks, they insufficiently adapt to new tasks. In supervised learning, this adaptation problem is addressed by large-scale pre-training followed by fine-tuning to new down-stream tasks. Recently, pre-training on multiple tasks has been gaining traction in RL. However, fine-tuning a pre-trained model often suffers from catastrophic forgetting. That is, the performance on the pre-training tasks deteriorates when fine-tuning on new tasks. To investigate the catastrophic forgetting phenomenon, we first jointly pre-train a model on datasets from two benchmark suites, namely Meta-World and DMControl. Then, we evaluate and compare a variety of fine-tuning methods prevalent in natural language processing, both in terms of performance on new tasks, and how well performance on pre-training tasks is retained. Our study shows that with most fine-tuning approaches, the performance on pre-training tasks deteriorates significantly. Therefore, we propose a novel method, Learning-to-Modulate (L2M), that avoids the degradation of learned skills by modulating the information flow of the frozen pre-trained model via a learnable modulation pool. Our method achieves state-of-the-art performance on the Continual-World benchmark, while retaining performance on the pre-training tasks. Finally, to aid future research in this area, we release a dataset encompassing 50 Meta-World and 16 DMControl tasks.

PDF Details

JMLR Journal 2023 Journal Article

Nevis'22: A Stream of 100 Tasks Sampled from 30 Years of Computer Vision Research

  • Jorg Bornschein
  • Alexandre Galashov
  • Ross Hemsley
  • Amal Rannen-Triki
  • Yutian Chen
  • Arslan Chaudhry
  • Xu Owen He
  • Arthur Douillard

A shared goal of several machine learning communities like continual learning, meta-learning and transfer learning, is to design algorithms and models that efficiently and robustly adapt to unseen tasks. An even more ambitious goal is to build models that never stop adapting, and that become increasingly more efficient through time by suitably transferring the accrued knowledge. Beyond the study of the actual learning algorithm and model architecture, there are several hurdles towards our quest to build such models, such as the choice of learning protocol, metric of success and data needed to validate research hypotheses. In this work, we introduce the Never-Ending VIsual-classification Stream (NEVIS'22), a benchmark consisting of a stream of over 100 visual classification tasks, sorted chronologically and extracted from papers sampled uniformly from computer vision proceedings spanning the last three decades. The resulting stream reflects what the research community thought was meaningful at any point in time, and it serves as an ideal test bed to assess how well models can adapt to new tasks, and do so better and more efficiently as time goes by. Despite being limited to classification, the resulting stream has a rich diversity of tasks from OCR, to texture analysis, scene recognition, and so forth. The diversity is also reflected in the wide range of dataset sizes, spanning over four orders of magnitude. Overall, NEVIS'22 poses an unprecedented challenge for current sequential learning approaches due to the scale and diversity of tasks, yet with a low entry barrier as it is limited to a single modality and well understood supervised learning problems. Moreover, we provide a reference implementation including strong baselines and an evaluation protocol to compare methods in terms of their trade-off between accuracy and compute. We hope that NEVIS'22 can be useful to researchers working on continual learning, meta-learning, AutoML and more generally sequential learning, and help these communities join forces towards more robust models that efficiently adapt to a never ending stream of data. [abs] [ pdf ][ bib ] [ code ] &copy JMLR 2023. ( edit, beta )

PDF Details

ICLR Conference 2023 Conference Paper

Pre-training via Denoising for Molecular Property Prediction

  • Sheheryar Zaidi
  • Michael Schaarschmidt
  • James Martens
  • Hyunjik Kim
  • Yee Whye Teh
  • Alvaro Sanchez-Gonzalez
  • Peter W. Battaglia
  • Razvan Pascanu

Many important problems involving molecular property prediction from 3D structures have limited data, posing a generalization challenge for neural networks. In this paper, we describe a pre-training technique based on denoising that achieves a new state-of-the-art in molecular property prediction by utilizing large datasets of 3D molecular structures at equilibrium to learn meaningful representations for downstream tasks. Relying on the well-known link between denoising autoencoders and score-matching, we show that the denoising objective corresponds to learning a molecular force field -- arising from approximating the Boltzmann distribution with a mixture of Gaussians -- directly from equilibrium structures. Our experiments demonstrate that using this pre-training objective significantly improves performance on multiple benchmarks, achieving a new state-of-the-art on the majority of targets in the widely used QM9 dataset. Our analysis then provides practical insights into the effects of different factors -- dataset sizes, model size and architecture, and the choice of upstream and downstream datasets -- on pre-training.

Details

ICML Conference 2023 Conference Paper

Resurrecting Recurrent Neural Networks for Long Sequences

  • Antonio Orvieto
  • Samuel L. Smith
  • Albert Gu
  • Anushan Fernando
  • Çaglar Gülçehre
  • Razvan Pascanu
  • Soham De

Recurrent Neural Networks (RNNs) offer fast inference on long sequences but are hard to optimize and slow to train. Deep state-space models (SSMs) have recently been shown to perform remarkably well on long sequence modeling tasks, and have the added benefits of fast parallelizable training and RNN-like fast inference. However, while SSMs are superficially similar to RNNs, there are important differences that make it unclear where their performance boost over RNNs comes from. We show that careful design of deep RNNs using standard signal propagation arguments can recover the impressive performance of deep SSMs on long-range reasoning tasks, while matching their training speed. To achieve this, we analyze and ablate a series of changes to standard RNNs including linearizing and diagonalizing the recurrence, using better parameterizations and initializations, and ensuring careful normalization of the forward pass. Our results provide new insights on the origins of the impressive performance of deep SSMs, and introduce an RNN block called the Linear Recurrent Unit (or LRU) that matches both their performance on the Long Range Arena benchmark and their computational efficiency.

Details

ICLR Conference 2023 Conference Paper

SemPPL: Predicting Pseudo-Labels for Better Contrastive Representations

  • Matko Bosnjak
  • Pierre Harvey Richemond
  • Nenad Tomasev
  • Florian Strub
  • Jacob C. Walker
  • Felix Hill
  • Lars Buesing
  • Razvan Pascanu

Learning from large amounts of unsupervised data and a small amount of supervision is an important open problem in computer vision. We propose a new semi-supervised learning method, Semantic Positives via Pseudo-Labels (SEMPPL), that combines labelled and unlabelled data to learn informative representations. Our method extends self-supervised contrastive learning—where representations are shaped by distinguishing whether two samples represent the same underlying datum (positives) or not (negatives)—with a novel approach to selecting positives. To enrich the set of positives, we leverage the few existing ground-truth labels to predict the missing ones through a k-nearest neighbors classifier by using the learned embeddings of the labelled data. We thus extend the set of positives with datapoints having the same pseudo-label and call these semantic positives. We jointly learn the representation and predict bootstrapped pseudo-labels. This creates a reinforcing cycle. Strong initial representations enable better pseudo-label predictions which then improve the selection of semantic positives and lead to even better representations. SEMPPL outperforms competing semi-supervised methods setting new state-of-the-art performance of 76% and 68.5% top-1accuracy when using a ResNet-50 and training on 10% and 1% of labels on ImageNet, respectively. Furthermore, when using selective kernels, SEMPPL significantly outperforms previous state-of-the-art achieving 72.3% and 78.3% top-1accuracy on ImageNet with 1% and 10% labels, respectively, which improves absolute +7.8% and +6.2% over previous work. SEMPPL also exhibits state-of-the-art performance over larger ResNet models as well as strong robustness, out-of-distribution and transfer performance. We release the checkpoints and the evaluation code at https://github.com/deepmind/semppl.

Details

NeurIPS Conference 2023 Conference Paper

The Tunnel Effect: Building Data Representations in Deep Neural Networks

  • Wojciech Masarczyk
  • Mateusz Ostaszewski
  • Ehsan Imani
  • Razvan Pascanu
  • Piotr Miłoś
  • Tomasz Trzcinski

Deep neural networks are widely known for their remarkable effectiveness across various tasks, with the consensus that deeper networks implicitly learn more complex data representations. This paper shows that sufficiently deep networks trained for supervised image classification split into two distinct parts that contribute to the resulting data representations differently. The initial layers create linearly-separable representations, while the subsequent layers, which we refer to as \textit{the tunnel}, compress these representations and have a minimal impact on the overall performance. We explore the tunnel's behavior through comprehensive empirical studies, highlighting that it emerges early in the training process. Its depth depends on the relation between the network's capacity and task complexity. Furthermore, we show that the tunnel degrades out-of-distribution generalization and discuss its implications for continual learning.

PDF Details

ICML Conference 2023 Conference Paper

Understanding Plasticity in Neural Networks

  • Clare Lyle
  • Zeyu Zheng
  • Evgenii Nikishin
  • Bernardo Ávila Pires
  • Razvan Pascanu
  • Will Dabney

Plasticity, the ability of a neural network to quickly change its predictions in response to new information, is essential for the adaptability and robustness of deep reinforcement learning systems. Deep neural networks are known to lose plasticity over the course of training even in relatively simple learning problems, but the mechanisms driving this phenomenon are still poorly understood. This paper conducts a systematic empirical analysis into plasticity loss, with the goal of understanding the phenomenon mechanistically in order to guide the future development of targeted solutions. We find that loss of plasticity is deeply connected to changes in the curvature of the loss landscape, but that it often occurs in the absence of saturated units. Based on this insight, we identify a number of parameterization and optimization design choices which enable networks to better preserve plasticity over the course of training. We validate the utility of these findings on larger-scale RL benchmarks in the Arcade Learning Environment.

Details

TMLR Journal 2022 Journal Article

An empirical study of implicit regularization in deep offline RL

  • Caglar Gulcehre
  • Srivatsan Srinivasan
  • Jakub Sygnowski
  • Georg Ostrovski
  • Mehrdad Farajtabar
  • Matthew Hoffman
  • Razvan Pascanu
  • Arnaud Doucet

Deep neural networks are the most commonly used function approximators in offline reinforcement learning. Prior works have shown that neural nets trained with TD-learning and gradient descent can exhibit implicit regularization that can be characterized by under-parameterization of these networks. Specifically, the rank of the penultimate feature layer, also called effective rank, has been observed to drastically collapse during the training. In turn, this collapse has been argued to reduce the model's ability to further adapt in later stages of learning, leading to the diminished final performance. Such an association between the effective rank and performance makes effective rank compelling for offline RL, primarily for offline policy evaluation. In this work, we conduct a careful empirical study on the relation between effective rank and performance on three offline RL datasets: bsuite, Atari, and DeepMind lab. We observe that a direct association exists only in restricted settings and disappears in the more extensive hyperparameter sweeps. Also, we empirically identify three phases of learning that explain the impact of implicit regularization on the learning dynamics and found that bootstrapping alone is insufficient to explain the collapse of the effective rank. Further, we show that several other factors could confound the relationship between effective rank and performance and conclude that studying this association under simplistic assumptions could be highly misleading.

PDF Details

JMLR Journal 2022 Journal Article

Behavior Priors for Efficient Reinforcement Learning

  • Dhruva Tirumala
  • Alexandre Galashov
  • Hyeonwoo Noh
  • Leonard Hasenclever
  • Razvan Pascanu
  • Jonathan Schwarz
  • Guillaume Desjardins
  • Wojciech Marian Czarnecki

As we deploy reinforcement learning agents to solve increasingly challenging problems, methods that allow us to inject prior knowledge about the structure of the world and effective solution strategies becomes increasingly important. In this work we consider how information and architectural constraints can be combined with ideas from the probabilistic modeling literature to learn behavior priors that capture the common movement and interaction patterns that are shared across a set of related tasks or contexts. For example the day-to day behavior of humans comprises distinctive locomotion and manipulation patterns that recur across many different situations and goals. We discuss how such behavior patterns can be captured using probabilistic trajectory models and how these can be integrated effectively into reinforcement learning schemes, e.g. to facilitate multi-task and transfer learning. We then extend these ideas to latent variable models and consider a formulation to learn hierarchical priors that capture different aspects of the behavior in reusable modules. We discuss how such latent variable formulations connect to related work on hierarchical reinforcement learning (HRL) and mutual information and curiosity based objectives, thereby offering an alternative perspective on existing ideas. We demonstrate the effectiveness of our framework by applying it to a range of simulated continuous control domains, videos of which can be found at the following url: https://sites.google.com/view/behavior-priors. [abs] [ pdf ][ bib ] &copy JMLR 2022. ( edit, beta )

PDF Details

NeurIPS Conference 2022 Conference Paper

Disentangling Transfer in Continual Reinforcement Learning

  • Maciej Wolczyk
  • Michał Zając
  • Razvan Pascanu
  • Łukasz Kuciński
  • Piotr Miłoś

The ability of continual learning systems to transfer knowledge from previously seen tasks in order to maximize performance on new tasks is a significant challenge for the field, limiting the applicability of continual learning solutions to realistic scenarios. Consequently, this study aims to broaden our understanding of transfer and its driving forces in the specific case of continual reinforcement learning. We adopt SAC as the underlying RL algorithm and Continual World as a suite of continuous control tasks. We systematically study how different components of SAC (the actor and the critic, exploration, and data) affect transfer efficacy, and we provide recommendations regarding various modeling options. The best set of choices, dubbed ClonEx-SAC, is evaluated on the recent Continual World benchmark. ClonEx-SAC achieves 87% final success rate compared to 80% of PackNet, the best method in the benchmark. Moreover, the transfer grows from 0. 18 to 0. 54 according to the metric provided by Continual World.

PDF Details

ICML Conference 2022 Conference Paper

The CLRS Algorithmic Reasoning Benchmark

  • Petar Velickovic
  • Adrià Puigdomènech Badia
  • David Budden
  • Razvan Pascanu
  • Andrea Banino
  • Misha Dashevskiy
  • Raia Hadsell
  • Charles Blundell

Learning representations of algorithms is an emerging area of machine learning, seeking to bridge concepts from neural networks with classical algorithms. Several important works have investigated whether neural networks can effectively reason like algorithms, typically by learning to execute them. The common trend in the area, however, is to generate targeted kinds of algorithmic data to evaluate specific hypotheses, making results hard to transfer across publications, and increasing the barrier of entry. To consolidate progress and work towards unified evaluation, we propose the CLRS Algorithmic Reasoning Benchmark, covering classical algorithms from the Introduction to Algorithms textbook. Our benchmark spans a variety of algorithmic reasoning procedures, including sorting, searching, dynamic programming, graph algorithms, string algorithms and geometric algorithms. We perform extensive experiments to demonstrate how several popular algorithmic reasoning baselines perform on these tasks, and consequently, highlight links to several open challenges. Our library is readily available at https: //github. com/deepmind/clrs.

Details

ICML Conference 2022 Conference Paper

Wide Neural Networks Forget Less Catastrophically

  • Seyed-Iman Mirzadeh
  • Arslan Chaudhry
  • Dong Yin
  • Huiyi Hu
  • Razvan Pascanu
  • Dilan Görür
  • Mehrdad Farajtabar

A primary focus area in continual learning research is alleviating the "catastrophic forgetting" problem in neural networks by designing new algorithms that are more robust to the distribution shifts. While the recent progress in continual learning literature is encouraging, our understanding of what properties of neural networks contribute to catastrophic forgetting is still limited. To address this, instead of focusing on continual learning algorithms, in this work, we focus on the model itself and study the impact of "width" of the neural network architecture on catastrophic forgetting, and show that width has a surprisingly significant effect on forgetting. To explain this effect, we study the learning dynamics of the network from various perspectives such as gradient orthogonality, sparsity, and lazy training regime. We provide potential explanations that are consistent with the empirical results across different architectures and continual learning benchmarks.

Details

NeurIPS Conference 2021 Conference Paper

Continual World: A Robotic Benchmark For Continual Reinforcement Learning

  • Maciej Wołczyk
  • Michał Zając
  • Razvan Pascanu
  • Łukasz Kuciński
  • Piotr Miłoś

Continual learning (CL) --- the ability to continuously learn, building on previously acquired knowledge --- is a natural requirement for long-lived autonomous reinforcement learning (RL) agents. While building such agents, one needs to balance opposing desiderata, such as constraints on capacity and compute, the ability to not catastrophically forget, and to exhibit positive transfer on new tasks. Understanding the right trade-off is conceptually and computationally challenging, which we argue has led the community to overly focus on catastrophic forgetting. In response to these issues, we advocate for the need to prioritize forward transfer and propose Continual World, a benchmark consisting of realistic and meaningfully diverse robotic tasks built on top of Meta-World as a testbed. Following an in-depth empirical evaluation of existing CL methods, we pinpoint their limitations and highlight unique algorithmic challenges in the RL setting. Our benchmark aims to provide a meaningful and computationally inexpensive challenge for the community and thus help better understand the performance of existing and future solutions. Information about the benchmark, including the open-source code, is available at https: //sites. google. com/view/continualworld.

PDF Details

ICLR Conference 2021 Conference Paper

Linear Mode Connectivity in Multitask and Continual Learning

  • Seyed-Iman Mirzadeh
  • Mehrdad Farajtabar
  • Dilan Görür
  • Razvan Pascanu
  • Hassan Ghasemzadeh 0001

Continual (sequential) training and multitask (simultaneous) training are often attempting to solve the same overall objective: to find a solution that performs well on all considered tasks. The main difference is in the training regimes, where continual learning can only have access to one task at a time, which for neural networks typically leads to catastrophic forgetting. That is, the solution found for a subsequent task does not perform well on the previous ones anymore. However, the relationship between the different minima that the two training regimes arrive at is not well understood. What sets them apart? Is there a local structure that could explain the difference in performance achieved by the two different schemes? Motivated by recent work showing that different minima of the same task are typically connected by very simple curves of low error, we investigate whether multitask and continual solutions are similarly connected. We empirically find that indeed such connectivity can be reliably achieved and, more interestingly, it can be done by a linear path, conditioned on having the same initialization for both. We thoroughly analyze this observation and discuss its significance for the continual learning process. Furthermore, we exploit this finding to propose an effective algorithm that constrains the sequentially learned minima to behave as the multitask solution. We show that our method outperforms several state of the art continual learning algorithms on various vision benchmarks.

Details

NeurIPS Conference 2021 Conference Paper

LiRo: Benchmark and leaderboard for Romanian language tasks

  • Stefan Dumitrescu
  • Petru Rebeja
  • Beata Lorincz
  • Mihaela Gaman
  • Andrei Avram
  • Mihai Ilie
  • Andrei Pruteanu
  • Adriana Stan

Recent advances in NLP have been sustained by the availability of large amounts of data and standardized benchmarks, which are not available for many languages. As a small step towards addressing this we propose LiRo, a platform for benchmarking models on the Romanian language on nine standard tasks: text classification, named entity recognition, machine translation, sentiment analysis, POS tagging, dependency parsing, language modelling, question-answering, and semantic textual similarity. We also include a less standard task of embedding debiasing, to address the growing concerns related to gender bias in language models. The platform exposes per-task leaderboards populated with baseline results for each task. In addition, we create three new datasets: one from Romanian Wikipedia and two by translating the Semantic Textual Similarity (STS) benchmark and the Cross-lingual Question Answering Dataset (XQuAD) into Romanian. We believe LiRo will not only add to the growing body of benchmarks covering various languages, but can also enable multi-lingual research by augmenting parallel corpora, and hence is of interest for the wider NLP community. LiRo is available at https: //lirobenchmark. github. io/.

PDF Details

NeurIPS Conference 2021 Conference Paper

On the Role of Optimization in Double Descent: A Least Squares Study

  • Ilja Kuzborskij
  • Csaba Szepesvari
  • Omar Rivasplata
  • Amal Rannen-Triki
  • Razvan Pascanu

Empirically it has been observed that the performance of deep neural networks steadily improves with increased model size, contradicting the classical view on overfitting and generalization. Recently, the double descent phenomenon has been proposed to reconcile this observation with theory, suggesting that the test error has a second descent when the model becomes sufficiently overparameterized, as the model size itself acts as an implicit regularizer. In this paper we add to the growing body of work in this space, providing a careful study of learning dynamics as a function of model size for the least squares scenario. We show an excess risk bound for the gradient descent solution of the least squares objective. The bound depends on the smallest non-zero eigenvalue of the sample covariance matrix of the input features, via a functional form that has the double descent behaviour. This gives a new perspective on the double descent curves reported in the literature, as our analysis of the excess risk allows to decouple the effect of optimization and generalization error. In particular, we find that in the case of noiseless regression, double descent is explained solely by optimization-related quantities, which was missed in studies focusing on the Moore-Penrose pseudoinverse solution. We believe that our derivation provides an alternative view compared to existing works, shedding some light on a possible cause of this phenomenon, at least in the considered least squares setting. We empirically explore if our predictions hold for neural networks, in particular whether the spectrum of the sample covariance of features at intermediary hidden layers has a similar behaviour as the one predicted by our derivations in the least squares setting.

PDF Details

NeurIPS Conference 2021 Conference Paper

Powerpropagation: A sparsity inducing weight reparameterisation

  • Jonathan Schwarz
  • Siddhant Jayakumar
  • Razvan Pascanu
  • Peter E Latham
  • Yee Teh

The training of sparse neural networks is becoming an increasingly important tool for reducing the computational footprint of models at training and evaluation, as well enabling the effective scaling up of models. Whereas much work over the years has been dedicated to specialised pruning techniques, little attention has been paid to the inherent effect of gradient based training on model sparsity. Inthis work, we introduce Powerpropagation, a new weight-parameterisation for neural networks that leads to inherently sparse models. Exploiting the behaviour of gradient descent, our method gives rise to weight updates exhibiting a “rich get richer” dynamic, leaving low-magnitude parameters largely unaffected by learning. Models trained in this manner exhibit similar performance, but have a distributionwith markedly higher density at zero, allowing more parameters to be pruned safely. Powerpropagation is general, intuitive, cheap and straight-forward to implement and can readily be combined with various other techniques. To highlight its versatility, we explore it in two very different settings: Firstly, following a recent line of work, we investigate its effect on sparse training for resource-constrained settings. Here, we combine Powerpropagation with a traditional weight-pruning technique as well as recent state-of-the-art sparse-to-sparse algorithms, showing superior performance on the ImageNet benchmark. Secondly, we advocate the useof sparsity in overcoming catastrophic forgetting, where compressed representations allow accommodating a large number of tasks at fixed model capacity. In all cases our reparameterisation considerably increases the efficacy of the off-the-shelf methods.

PDF Details

ICML Conference 2021 Conference Paper

Spectral Normalisation for Deep Reinforcement Learning: An Optimisation Perspective

  • Florin Gogianu
  • Tudor Berariu
  • Mihaela Rosca
  • Claudia Clopath
  • Lucian Busoniu
  • Razvan Pascanu

Most of the recent deep reinforcement learning advances take an RL-centric perspective and focus on refinements of the training objective. We diverge from this view and show we can recover the performance of these developments not by changing the objective, but by regularising the value-function estimator. Constraining the Lipschitz constant of a single layer using spectral normalisation is sufficient to elevate the performance of a Categorical-DQN agent to that of a more elaborated agent on the challenging Atari domain. We conduct ablation studies to disentangle the various effects normalisation has on the learning dynamics and show that is sufficient to modulate the parameter updates to recover most of the performance of spectral normalisation. These findings hint towards the need to also focus on the neural component and its learning dynamics to tackle the peculiarities of Deep Reinforcement Learning.

Details

ICLR Conference 2020 Conference Paper

Functional Regularisation for Continual Learning with Gaussian Processes

  • Michalis K. Titsias
  • Jonathan Richard Schwarz
  • Alexander G. de G. Matthews
  • Razvan Pascanu
  • Yee Whye Teh

We introduce a framework for Continual Learning (CL) based on Bayesian inference over the function space rather than the parameters of a deep neural network. This method, referred to as functional regularisation for Continual Learning, avoids forgetting a previous task by constructing and memorising an approximate posterior belief over the underlying task-specific function. To achieve this we rely on a Gaussian process obtained by treating the weights of the last layer of a neural network as random and Gaussian distributed. Then, the training algorithm sequentially encounters tasks and constructs posterior beliefs over the task-specific functions by using inducing point sparse Gaussian process methods. At each step a new task is first learnt and then a summary is constructed consisting of (i) inducing inputs – a fixed-size subset of the task inputs selected such that it optimally represents the task – and (ii) a posterior distribution over the function values at these inputs. This summary then regularises learning of future tasks, through Kullback-Leibler regularisation terms. Our method thus unites approaches focused on (pseudo-)rehearsal with those derived from a sequential Bayesian inference perspective in a principled way, leading to strong results on accepted benchmarks.

Details

ICML Conference 2020 Conference Paper

Improving the Gating Mechanism of Recurrent Neural Networks

  • Albert Gu
  • Çaglar Gülçehre
  • Thomas Paine
  • Matthew Hoffman 0002
  • Razvan Pascanu

Gating mechanisms are widely used in neural network models, where they allow gradients to backpropagate easily through depth or time. However, their saturation property introduces problems of its own. For example, in recurrent models these gates need to have outputs near 1 to propagate information over long time-delays, which requires them to operate in their saturation regime and hinders gradient-based learning of the gate mechanism. We address this problem by deriving two synergistic modifications to the standard gating mechanism that are easy to implement, introduce no additional hyperparameters, and improve learnability of the gates when they are close to saturation. We show how these changes are related to and improve on alternative recently proposed gating mechanisms such as chrono-initialization and Ordered Neurons. Empirically, our simple gating mechanisms robustly improve the performance of recurrent models on a range of applications, including synthetic memorization tasks, sequential image classification, language modeling, and reinforcement learning, particularly when long-term dependencies are involved.

Details

ICLR Conference 2020 Conference Paper

Meta-Learning with Warped Gradient Descent

  • Sebastian Flennerhag
  • Andrei A. Rusu
  • Razvan Pascanu
  • Francesco Visin
  • Hujun Yin
  • Raia Hadsell

Learning an efficient update rule from data that promotes rapid learning of new tasks from the same distribution remains an open problem in meta-learning. Typically, previous works have approached this issue either by attempting to train a neural network that directly produces updates or by attempting to learn better initialisations or scaling factors for a gradient-based update rule. Both of these approaches pose challenges. On one hand, directly producing an update forgoes a useful inductive bias and can easily lead to non-converging behaviour. On the other hand, approaches that try to control a gradient-based update rule typically resort to computing gradients through the learning process to obtain their meta-gradients, leading to methods that can not scale beyond few-shot task adaptation. In this work, we propose Warped Gradient Descent (WarpGrad), a method that intersects these approaches to mitigate their limitations. WarpGrad meta-learns an efficiently parameterised preconditioning matrix that facilitates gradient descent across the task distribution. Preconditioning arises by interleaving non-linear layers, referred to as warp-layers, between the layers of a task-learner. Warp-layers are meta-learned without backpropagating through the task training process in a manner similar to methods that learn to directly produce updates. WarpGrad is computationally efficient, easy to implement, and can scale to arbitrarily large meta-learning problems. We provide a geometrical interpretation of the approach and evaluate its effectiveness in a variety of settings, including few-shot, standard supervised, continual and reinforcement learning.

Details

ICLR Conference 2020 Conference Paper

Multiplicative Interactions and Where to Find Them

  • Siddhant M. Jayakumar
  • Wojciech Czarnecki 0001
  • Jacob Menick
  • Jonathan Richard Schwarz
  • Jack W. Rae
  • Simon Osindero
  • Yee Whye Teh
  • Tim Harley

We explore the role of multiplicative interaction as a unifying framework to describe a range of classical and modern neural network architectural motifs, such as gating, attention layers, hypernetworks, and dynamic convolutions amongst others. Multiplicative interaction layers as primitive operations have a long-established presence in the literature, though this often not emphasized and thus under-appreciated. We begin by showing that such layers strictly enrich the representable function classes of neural networks. We conjecture that multiplicative interactions offer a particularly powerful inductive bias when fusing multiple streams of information or when conditional computation is required. We therefore argue that they should be considered in many situation where multiple compute or information paths need to be combined, in place of the simple and oft-used concatenation operation. Finally, we back up our claims and demonstrate the potential of multiplicative interactions by applying them in large-scale complex RL and sequence modelling tasks, where their use allows us to deliver state-of-the-art results, and thereby provides new evidence in support of multiplicative interactions playing a more prominent role when designing new neural network architectures.

Details

NeurIPS Conference 2020 Conference Paper

Pointer Graph Networks

  • Petar Veličković
  • Lars Buesing
  • Matthew Overlan
  • Razvan Pascanu
  • Oriol Vinyals
  • Charles Blundell

Graph neural networks (GNNs) are typically applied to static graphs that are assumed to be known upfront. This static input structure is often informed purely by insight of the machine learning practitioner, and might not be optimal for the actual task the GNN is solving. In absence of reliable domain expertise, one might resort to inferring the latent graph structure, which is often difficult due to the vast search space of possible graphs. Here we introduce Pointer Graph Networks (PGNs) which augment sets or graphs with additional inferred edges for improved model generalisation ability. PGNs allow each node to dynamically point to another node, followed by message passing over these pointers. The sparsity of this adaptable graph structure makes learning tractable while still being sufficiently expressive to simulate complex algorithms. Critically, the pointing mechanism is directly supervised to model long-term sequences of operations on classical data structures, incorporating useful structural inductive biases from theoretical computer science. Qualitatively, we demonstrate that PGNs can learn parallelisable variants of pointer-based data structures, namely disjoint set unions and link/cut trees. PGNs generalise out-of-distribution to 5x larger test inputs on dynamic graph connectivity tasks, outperforming unrestricted GNNs and Deep Sets.

PDF Details

ICML Conference 2020 Conference Paper

Stabilizing Transformers for Reinforcement Learning

  • Emilio Parisotto
  • H. Francis Song
  • Jack W. Rae
  • Razvan Pascanu
  • Çaglar Gülçehre
  • Siddhant M. Jayakumar
  • Max Jaderberg
  • Raphaël Lopez Kaufman

Owing to their ability to both effectively integrate information over long time horizons and scale to massive amounts of data, self-attention architectures have recently shown breakthrough success in natural language processing (NLP). Harnessing the transformer’s ability to process long time horizons of information could provide a similar performance boost in partially observable reinforcement learning (RL) domains, but the large-scale transformers used in NLP have yet to be successfully applied to the RL setting. In this work we demonstrate that the standard transformer architecture is difficult to optimize, which was previously observed in the supervised learning setting but becomes especially pronounced with RL objectives. We propose architectural modifications that substantially improve the stability and learning speed of the original Transformer and XL variant. The proposed architecture, the Gated Transformer-XL (GTrXL), surpasses LSTMs on challenging memory environments and achieves state-of-the-art results on the multi-task DMLab-30 benchmark suite, exceeding the performance of an external memory architecture. We show that the GTrXL has stability and performance that consistently matches or exceeds a competitive LSTM baseline, including on more reactive tasks where memory is less critical.

Details

NeurIPS Conference 2020 Conference Paper

Top-KAST: Top-K Always Sparse Training

  • Siddhant Jayakumar
  • Razvan Pascanu
  • Jack Rae
  • Simon Osindero
  • Erich Elsen

Sparse neural networks are becoming increasingly important as the field seeks to improve the performance of existing models by scaling them up, while simultaneously trying to reduce power consumption and computational footprint. Unfortunately, most existing methods for inducing performant sparse models still entail the instantiation of dense parameters, or dense gradients in the backward-pass, during training. For very large models this requirement can be prohibitive. In this work we propose Top-KAST, a method that preserves constant sparsity throughout training (in both the forward and backward-passes). We demonstrate the efficacy of our approach by showing that it performs comparably to or better than previous works when training models on the established ImageNet benchmark, whilst fully maintaining sparsity. In addition to our ImageNet results, we also demonstrate our approach in the domain of language modeling where the current best performing architectures tend to have tens of billions of parameters and scaling up does not yet seem to have saturated performance. Sparse versions of these architectures can be run with significantly fewer resources, making them more widely accessible and applicable. Furthermore, in addition to being effective, our approach is straightforward and can easily be implemented in a wide range of existing machine learning frameworks with only a few additional lines of code. We therefore hope that our contribution will help enable the broader community to explore the potential held by massive models, without incurring massive computational cost.

PDF Details

NeurIPS Conference 2020 Conference Paper

Understanding the Role of Training Regimes in Continual Learning

  • Seyed Iman Mirzadeh
  • Mehrdad Farajtabar
  • Razvan Pascanu
  • Hassan Ghasemzadeh

Catastrophic forgetting affects the training of neural networks, limiting their ability to learn multiple tasks sequentially. From the perspective of the well established plasticity-stability dilemma, neural networks tend to be overly plastic, lacking the stability necessary to prevent the forgetting of previous knowledge, which means that as learning progresses, networks tend to forget previously seen tasks. This phenomenon coined in the continual learning literature, has attracted much attention lately, and several families of approaches have been proposed with different degrees of success. However, there has been limited prior work extensively analyzing the impact that different training regimes -- learning rate, batch size, regularization method-- can have on forgetting. In this work, we depart from the typical approach of altering the learning algorithm to improve stability. Instead, we hypothesize that the geometrical properties of the local minima found for each task play an important role in the overall degree of forgetting. In particular, we study the effect of dropout, learning rate decay, and batch size on forming training regimes that widen the tasks' local minima and consequently, on helping it not to forget catastrophically. Our study provides practical insights to improve stability via simple yet effective techniques that outperform alternative baselines.

PDF Details

NeurIPS Conference 2019 Conference Paper

Continual Unsupervised Representation Learning

  • Dushyant Rao
  • Francesco Visin
  • Andrei Rusu
  • Razvan Pascanu
  • Yee Whye Teh
  • Raia Hadsell

Continual learning aims to improve the ability of modern learning systems to deal with non-stationary distributions, typically by attempting to learn a series of tasks sequentially. Prior art in the field has largely considered supervised or reinforcement learning tasks, and often assumes full knowledge of task labels and boundaries. In this work, we propose an approach (CURL) to tackle a more general problem that we will refer to as unsupervised continual learning. The focus is on learning representations without any knowledge about task identity, and we explore scenarios when there are abrupt changes between tasks, smooth transitions from one task to another, or even when the data is shuffled. The proposed approach performs task inference directly within the model, is able to dynamically expand to capture new concepts over its lifetime, and incorporates additional rehearsal-based techniques to deal with catastrophic forgetting. We demonstrate the efficacy of CURL in an unsupervised learning setting with MNIST and Omniglot, where the lack of labels ensures no information is leaked about the task. Further, we demonstrate strong performance compared to prior art in an i. i. d setting, or when adapting the technique to supervised tasks such as incremental class learning.

PDF Details

ICML Conference 2018 Conference Paper

Been There, Done That: Meta-Learning with Episodic Recall

  • Samuel Ritter
  • Jane X. Wang
  • Zeb Kurth-Nelson
  • Siddhant M. Jayakumar
  • Charles Blundell
  • Razvan Pascanu
  • Matthew M. Botvinick

Meta-learning agents excel at rapidly learning new tasks from open-ended task distributions; yet, they forget what they learn about each task as soon as the next begins. When tasks reoccur {–} as they do in natural environments {–} meta-learning agents must explore again instead of immediately exploiting previously discovered solutions. We propose a formalism for generating open-ended yet repetitious environments, then develop a meta-learning architecture for solving these environments. This architecture melds the standard LSTM working memory with a differentiable neural episodic memory. We explore the capabilities of agents with this episodic LSTM in five meta-learning environments with reoccurring tasks, ranging from bandits to navigation and stochastic sequential decision problems.

Details

ICML Conference 2018 Conference Paper

Mix & Match Agent Curricula for Reinforcement Learning

  • Wojciech Czarnecki 0001
  • Siddhant M. Jayakumar
  • Max Jaderberg
  • Leonard Hasenclever
  • Yee Whye Teh
  • Nicolas Heess
  • Simon Osindero
  • Razvan Pascanu

We introduce Mix and match (M&M) – a training framework designed to facilitate rapid and effective learning in RL agents that would be too slow or too challenging to train otherwise. The key innovation is a procedure that allows us to automatically form a curriculum over agents. Through such a curriculum we can progressively train more complex agents by, effectively, bootstrapping from solutions found by simpler agents. In contradistinction to typical curriculum learning approaches, we do not gradually modify the tasks or environments presented, but instead use a process to gradually alter how the policy is represented internally. We show the broad applicability of our method by demonstrating significant performance gains in three different experimental setups: (1) We train an agent able to control more than 700 actions in a challenging 3D first-person task; using our method to progress through an action-space curriculum we achieve both faster training and better final performance than one obtains using traditional methods. (2) We further show that M&M can be used successfully to progress through a curriculum of architectural variants defining an agents internal state. (3) Finally, we illustrate how a variant of our method can be used to improve agent performance in a multitask setting.

Details

ICML Conference 2018 Conference Paper

Progress & Compress: A scalable framework for continual learning

  • Jonathan Richard Schwarz
  • Wojciech Czarnecki 0001
  • Jelena Luketina
  • Agnieszka Grabska-Barwinska
  • Yee Whye Teh
  • Razvan Pascanu
  • Raia Hadsell

We introduce a conceptually simple and scalable framework for continual learning domains where tasks are learned sequentially. Our method is constant in the number of parameters and is designed to preserve performance on previously encountered tasks while accelerating learning progress on subsequent problems. This is achieved by training a network with two components: A knowledge base, capable of solving previously encountered problems, which is connected to an active column that is employed to efficiently learn the current task. After learning a new task, the active column is distilled into the knowledge base, taking care to protect any previously acquired skills. This cycle of active learning (progression) followed by consolidation (compression) requires no architecture growth, no access to or storing of previous data or tasks, and no task-specific parameters. We demonstrate the progress & compress approach on sequential classification of handwritten alphabets as well as two reinforcement learning domains: Atari games and 3D maze navigation.

Details

NeurIPS Conference 2018 Conference Paper

Relational recurrent neural networks

  • Adam Santoro
  • Ryan Faulkner
  • David Raposo
  • Jack Rae
  • Mike Chrzanowski
  • Theophane Weber
  • Daan Wierstra
  • Oriol Vinyals

Memory-based neural networks model temporal data by leveraging an ability to remember information for long periods. It is unclear, however, whether they also have an ability to perform complex relational reasoning with the information they remember. Here, we first confirm our intuitions that standard memory architectures may struggle at tasks that heavily involve an understanding of the ways in which entities are connected -- i. e. , tasks involving relational reasoning. We then improve upon these deficits by using a new memory module -- a Relational Memory Core (RMC) -- which employs multi-head dot product attention to allow memories to interact. Finally, we test the RMC on a suite of tasks that may profit from more capable relational reasoning across sequential information, and show large gains in RL domains (BoxWorld & Mini PacMan), program evaluation, and language modeling, achieving state-of-the-art results on the WikiText-103, Project Gutenberg, and GigaWord datasets.

PDF Details

NeurIPS Conference 2017 Conference Paper

A simple neural network module for relational reasoning

  • Adam Santoro
  • David Raposo
  • David Barrett
  • Mateusz Malinowski
  • Razvan Pascanu
  • Peter Battaglia
  • Timothy Lillicrap

Relational reasoning is a central component of generally intelligent behavior, but has proven difficult for neural networks to learn. In this paper we describe how to use Relation Networks (RNs) as a simple plug-and-play module to solve problems that fundamentally hinge on relational reasoning. We tested RN-augmented networks on three tasks: visual question answering using a challenging dataset called CLEVR, on which we achieve state-of-the-art, super-human performance; text-based question answering using the bAbI suite of tasks; and complex reasoning about dynamical physical systems. Then, using a curated dataset called Sort-of-CLEVR we show that powerful convolutional networks do not have a general capacity to solve relational questions, but can gain this capacity when augmented with RNs. Thus, by simply augmenting convolutions, LSTMs, and MLPs with RNs, we can remove computational burden from network components that are not well-suited to handle relational reasoning, reduce overall network complexity, and gain a general ability to reason about the relations between entities and their properties.

PDF Details

NeurIPS Conference 2017 Conference Paper

Distral: Robust multitask reinforcement learning

  • Yee Teh
  • Victor Bapst
  • Wojciech Czarnecki
  • John Quan
  • James Kirkpatrick
  • Raia Hadsell
  • Nicolas Heess
  • Razvan Pascanu

Most deep reinforcement learning algorithms are data inefficient in complex and rich environments, limiting their applicability to many scenarios. One direction for improving data efficiency is multitask learning with shared neural network parameters, where efficiency may be improved through transfer across related tasks. In practice, however, this is not usually observed, because gradients from different tasks can interfere negatively, making learning unstable and sometimes even less data efficient. Another issue is the different reward schemes between tasks, which can easily lead to one task dominating the learning of a shared model. We propose a new approach for joint training of multiple tasks, which we refer to as Distral (DIStill & TRAnsfer Learning). Instead of sharing parameters between the different workers, we propose to share a distilled policy that captures common behaviour across tasks. Each worker is trained to solve its own task while constrained to stay close to the shared policy, while the shared policy is trained by distillation to be the centroid of all task policies. Both aspects of the learning process are derived by optimizing a joint objective function. We show that our approach supports efficient transfer on complex 3D environments, outperforming several related methods. Moreover, the proposed learning process is more robust and more stable---attributes that are critical in deep reinforcement learning.

PDF Details

NeurIPS Conference 2017 Conference Paper

Imagination-Augmented Agents for Deep Reinforcement Learning

  • Sébastien Racanière
  • Theophane Weber
  • David Reichert
  • Lars Buesing
  • Arthur Guez
  • Danilo Jimenez Rezende
  • Adrià Puigdomènech Badia
  • Oriol Vinyals

We introduce Imagination-Augmented Agents (I2As), a novel architecture for deep reinforcement learning combining model-free and model-based aspects. In contrast to most existing model-based reinforcement learning and planning methods, which prescribe how a model should be used to arrive at a policy, I2As learn to interpret predictions from a trained environment model to construct implicit plans in arbitrary ways, by using the predictions as additional context in deep policy networks. I2As show improved data efficiency, performance, and robustness to model misspecification compared to several strong baselines.

PDF Details

ICML Conference 2017 Conference Paper

Sharp Minima Can Generalize For Deep Nets

  • Laurent Dinh
  • Razvan Pascanu
  • Samy Bengio
  • Yoshua Bengio

Despite their overwhelming capacity to overfit, deep learning architectures tend to generalize relatively well to unseen data, allowing them to be deployed in practice. However, explaining why this is the case is still an open area of research. One standing hypothesis that is gaining popularity, e. g. \ Hochreiter \& Schmidhuber (1997); Keskar et al. \ (2017), is that the flatness of minima of the loss function found by stochastic gradient based methods results in good generalization. This paper argues that most notions of flatness are problematic for deep models and can not be directly applied to explain generalization. Specifically, when focusing on deep networks with rectifier units, we can exploit the particular geometry of parameter space induced by the inherent symmetries that these architectures exhibit to build equivalent models corresponding to arbitrarily sharper minima. Or, depending on the definition of flatness, it is the same for any given minimum. Furthermore, if we allow to reparametrize a function, the geometry of its parameters can change drastically without affecting its generalization properties.

Details

NeurIPS Conference 2017 Conference Paper

Sobolev Training for Neural Networks

  • Wojciech Czarnecki
  • Simon Osindero
  • Max Jaderberg
  • Grzegorz Swirszcz
  • Razvan Pascanu

At the heart of deep learning we aim to use neural networks as function approximators - training them to produce outputs from inputs in emulation of a ground truth function or data creation process. In many cases we only have access to input-output pairs from the ground truth, however it is becoming more common to have access to derivatives of the target output with respect to the input -- for example when the ground truth function is itself a neural network such as in network compression or distillation. Generally these target derivatives are not computed, or are ignored. This paper introduces Sobolev Training for neural networks, which is a method for incorporating these target derivatives in addition the to target values while training. By optimising neural networks to not only approximate the function’s outputs but also the function’s derivatives we encode additional information about the target function within the parameters of the neural network. Thereby we can improve the quality of our predictors, as well as the data-efficiency and generalization capabilities of our learned function approximation. We provide theoretical justifications for such an approach as well as examples of empirical evidence on three distinct domains: regression on classical optimisation datasets, distilling policies of an agent playing Atari, and on large-scale applications of synthetic gradients. In all three domains the use of Sobolev Training, employing target derivatives in addition to target values, results in models with higher accuracy and stronger generalisation.

PDF Details

NeurIPS Conference 2017 Conference Paper

Visual Interaction Networks: Learning a Physics Simulator from Video

  • Nicholas Watters
  • Daniel Zoran
  • Theophane Weber
  • Peter Battaglia
  • Razvan Pascanu
  • Andrea Tacchetti

From just a glance, humans can make rich predictions about the future of a wide range of physical systems. On the other hand, modern approaches from engineering, robotics, and graphics are often restricted to narrow domains or require information about the underlying state. We introduce the Visual Interaction Network, a general-purpose model for learning the dynamics of a physical system from raw visual observations. Our model consists of a perceptual front-end based on convolutional neural networks and a dynamics predictor based on interaction networks. Through joint training, the perceptual front-end learns to parse a dynamic visual scene into a set of factored latent object representations. The dynamics predictor learns to roll these states forward in time by computing their interactions, producing a predicted physical trajectory of arbitrary length. We found that from just six input video frames the Visual Interaction Network can generate accurate future trajectories of hundreds of time steps on a wide range of physical systems. Our model can also be applied to scenes with invisible objects, inferring their future states from their effects on the visible objects, and can implicitly infer the unknown mass of objects. This work opens new opportunities for model-based decision-making and planning from raw sensory observations in complex physical environments.

PDF Details

NeurIPS Conference 2016 Conference Paper

Interaction Networks for Learning about Objects, Relations and Physics

  • Peter Battaglia
  • Razvan Pascanu
  • Matthew Lai
  • Danilo Jimenez Rezende
  • Koray Kavukcuoglu

Reasoning about objects, relations, and physics is central to human intelligence, and a key goal of artificial intelligence. Here we introduce the interaction network, a model which can reason about how objects in complex systems interact, supporting dynamical predictions, as well as inferences about the abstract properties of the system. Our model takes graphs as input, performs object- and relation-centric reasoning in a way that is analogous to a simulation, and is implemented using deep neural networks. We evaluate its ability to reason about several challenging physical domains: n-body problems, rigid-body collision, and non-rigid dynamics. Our results show it can be trained to accurately simulate the physical trajectories of dozens of objects over thousands of time steps, estimate abstract quantities such as energy, and generalize automatically to systems with different numbers and configurations of objects and relations. Our interaction network implementation is the first general-purpose, learnable physics engine, and a powerful general framework for reasoning about object and relations in a wide variety of complex real-world domains.

PDF Details

NeurIPS Conference 2015 Conference Paper

Natural Neural Networks

  • Guillaume Desjardins
  • Karen Simonyan
  • Razvan Pascanu
  • Koray Kavukcuoglu

We introduce Natural Neural Networks, a novel family of algorithms that speed up convergence by adapting their internal representation during training to improve conditioning of the Fisher matrix. In particular, we show a specific example that employs a simple and efficient reparametrization of the neural network weights by implicitly whitening the representation obtained at each layer, while preserving the feed-forward computation of the network. Such networks can be trained efficiently via the proposed Projected Natural Gradient Descent algorithm (PRONG), which amortizes the cost of these reparametrizations over many parameter updates and is closely related to the Mirror Descent online learning algorithm. We highlight the benefits of our method on both unsupervised and supervised learning tasks, and showcase its scalability by training on the large-scale ImageNet Challenge dataset.

PDF Details

NeurIPS Conference 2014 Conference Paper

Identifying and attacking the saddle point problem in high-dimensional non-convex optimization

  • Yann Dauphin
  • Razvan Pascanu
  • Caglar Gulcehre
  • Kyunghyun Cho
  • Surya Ganguli
  • Yoshua Bengio

A central challenge to many fields of science and engineering involves minimizing non-convex error functions over continuous, high dimensional spaces. Gradient descent or quasi-Newton methods are almost ubiquitously used to perform such minimizations, and it is often thought that a main source of difficulty for these local methods to find the global minimum is the proliferation of local minima with much higher error than the global minimum. Here we argue, based on results from statistical physics, random matrix theory, neural network theory, and empirical evidence, that a deeper and more profound difficulty originates from the proliferation of saddle points, not local minima, especially in high dimensional problems of practical interest. Such saddle points are surrounded by high error plateaus that can dramatically slow down learning, and give the illusory impression of the existence of a local minimum. Motivated by these arguments, we propose a new approach to second-order optimization, the saddle-free Newton method, that can rapidly escape high dimensional saddle points, unlike gradient descent and quasi-Newton methods. We apply this algorithm to deep or recurrent neural network training, and provide numerical evidence for its superior optimization performance.

PDF Details

NeurIPS Conference 2014 Conference Paper

On the Number of Linear Regions of Deep Neural Networks

  • Guido Montufar
  • Razvan Pascanu
  • Kyunghyun Cho
  • Yoshua Bengio

We study the complexity of functions computable by deep feedforward neural networks with piecewise linear activations in terms of the symmetries and the number of linear regions that they have. Deep networks are able to sequentially map portions of each layer's input-space to the same output. In this way, deep models compute functions that react equally to complicated patterns of different inputs. The compositional structure of these functions enables them to re-use pieces of computation exponentially often in terms of the network's depth. This paper investigates the complexity of such compositional maps and contributes new theoretical results regarding the advantage of depth for neural networks with piecewise linear activation functions. In particular, our analysis is not specific to a single family of models, and as an example, we employ it for rectifier and maxout networks. We improve complexity bounds from pre-existing work and investigate the behavior of units in higher layers.

PDF Details

ICLR Conference 2014 Conference Paper

Revisiting Natural Gradient for Deep Networks

  • Razvan Pascanu
  • Yoshua Bengio

The aim of this paper is three-fold. First we show that Hessian-Free (Martens, 2010) and Krylov Subspace Descent (Vinyals and Povey, 2012) can be described as implementations of natural gradient descent due to their use of the extended Gauss-Newton approximation of the Hessian. Secondly we re-derive natural gradient from basic principles, contrasting the difference between two versions of the algorithm found in the neural network literature, as well as highlighting a few differences between natural gradient and typical second order methods. Lastly we show empirically that natural gradient can be robust to overfitting and particularly it can be robust to the order in which the training data is presented to the model.

Details

ICML Conference 2013 Conference Paper

On the difficulty of training recurrent neural networks

  • Razvan Pascanu
  • Tomás Mikolov
  • Yoshua Bengio

There are two widely known issues with properly training recurrent neural networks, the vanishing and the exploding gradient problems detailed in Bengio et al. (1994). In this paper we attempt to improve the understanding of the underlying issues by exploring these problems from an analytical, a geometric and a dynamical systems perspective. Our analysis is used to justify a simple yet effective solution. We propose a gradient norm clipping strategy to deal with exploding gradients and a soft constraint for the vanishing gradients problem. We validate empirically our hypothesis and proposed solutions in the experimental section.

Details

JMLR Journal 2012 Journal Article

Learning Algorithms for the Classification Restricted Boltzmann Machine

  • Hugo Larochelle
  • Michael Mandel
  • Razvan Pascanu
  • Yoshua Bengio

Recent developments have demonstrated the capacity of restricted Boltzmann machines (RBM) to be powerful generative models, able to extract useful features from input data or construct deep artificial neural networks. In such settings, the RBM only yields a preprocessing or an initialization for some other model, instead of acting as a complete supervised model in its own right. In this paper, we argue that RBMs can provide a self-contained framework for developing competitive classifiers. We study the Classification RBM (ClassRBM), a variant on the RBM adapted to the classification setting. We study different strategies for training the ClassRBM and show that competitive classification performances can be reached when appropriately combining discriminative and generative training objectives. Since training according to the generative objective requires the computation of a generally intractable gradient, we also compare different approaches to estimating this gradient and address the issue of obtaining such a gradient for problems with very high dimensional inputs. Finally, we describe how to adapt the ClassRBM to two special cases of classification problems, namely semi-supervised and multitask learning. [abs] [ pdf ][ bib ] &copy JMLR 2012. ( edit, beta )

PDF Details

ICRA Conference 2010 Conference Paper

Extraction of quadrics from noisy point-clouds using a sensor noise model

  • Narunas Vaskevicius
  • Kaustubh Pathak
  • Razvan Pascanu
  • Andreas Birk 0002

Fitting optimum quadrics on segmented noisy range-image is a challenging task. The algebraic least-squares fit usually considered in the literature is biased, although it is fast to compute. We extend our previous work in planar patch extraction using a detailed sensor noise model to the extraction of quadrics. First, the fast least-squares method is modified to detect degeneracy automatically and it is used to aid segmentation. The final segmented quadric patch is then refined using a numerical maximum likelihood approach which employs the sensor range error model. Experimental results for artificial data and two different 3D sensors are provided to show the feasibility of our approach.

Details
v2026.09.13