Author name cluster

Loic Matthey

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.

11 papers

2 author rows

ICML Conference 2024 Conference Paper

Leveraging VLM-Based Pipelines to Annotate 3D Objects

Rishabh Kabra
Loic Matthey
Alexander Lerchner
Niloy J. Mitra

Pretrained vision language models (VLMs) present an opportunity to caption unlabeled 3D objects at scale. The leading approach to summarize VLM descriptions from different views of an object (Luo et al. , 2023) relies on a language model (GPT4) to produce the final output. This text-based aggregation is susceptible to hallucinations as it merges potentially contradictory descriptions. We propose an alternative algorithm to marginalize over factors such as the viewpoint that affect the VLM’s response. Instead of merging text-only responses, we utilize the VLM’s joint image-text likelihoods. We show our probabilistic aggregation is not only more reliable and efficient, but sets the SoTA on inferring object types with respect to human-verified labels. The aggregated annotations are also useful for conditional inference; they improve downstream predictions (e. g. , of object material) when the object’s type is specified as an auxiliary text-based input. Such auxiliary inputs allow ablating the contribution of visual reasoning over visionless reasoning in an unsupervised setting. With these supervised and unsupervised evaluations, we show how a VLM-based pipeline can be leveraged to produce reliable annotations for 764K objects from the Objaverse dataset.

Details

NeurIPS Conference 2023 Conference Paper

Combining Behaviors with the Successor Features Keyboard

Wilka Carvalho Carvalho
Andre Saraiva
Angelos Filos
Andrew Lampinen
Loic Matthey
Richard L Lewis
Honglak Lee
Satinder Singh

The Option Keyboard (OK) was recently proposed as a method for transferring behavioral knowledge across tasks. OK transfers knowledge by adaptively combining subsets of known behaviors using Successor Features (SFs) and Generalized Policy Improvement (GPI). However, it relies on hand-designed state-features and task encodings which are cumbersome to design for every new environment. In this work, we propose the "Successor Features Keyboard" (SFK), which enables transfer with discovered state-features and task encodings. To enable discovery, we propose the "Categorical Successor Feature Approximator" (CSFA), a novel learning algorithm for estimating SFs while jointly discovering state-features and task encodings. With SFK and CSFA, we achieve the first demonstration of transfer with SFs in a challenging 3D environment where all the necessary representations are discovered. We first compare CSFA against other methods for approximating SFs and show that only CSFA discovers representations compatible with SF&GPI at this scale. We then compare SFK against transfer learning baselines and show that it transfers most quickly to long-horizon tasks.

PDF Details

NeurIPS Conference 2021 Conference Paper

Alchemy: A benchmark and analysis toolkit for meta-reinforcement learning agents

Jane Wang
Michael King
Nicolas Porcel
Zeb Kurth-Nelson
Tina Zhu
Charles Deck
Peter Choy
Mary Cassin

There has been rapidly growing interest in meta-learning as a method for increasing the flexibility and sample efficiency of reinforcement learning. One problem in this area of research, however, has been a scarcity of adequate benchmark tasks. In general, the structure underlying past benchmarks has either been too simple to be inherently interesting, or too ill-defined to support principled analysis. In the present work, we introduce a new benchmark for meta-RL research, emphasizing transparency and potential for in-depth analysis as well as structural richness. Alchemy is a 3D video game, implemented in Unity, which involves a latent causal structure that is resampled procedurally from episode to episode, affording structure learning, online inference, hypothesis testing and action sequencing based on abstract domain knowledge. We evaluate a pair of powerful RL agents on Alchemy and present an in-depth analysis of one of these agents. Results clearly indicate a frank and specific failure of meta-learning, providing validation for Alchemy as a challenging benchmark for meta-RL. Concurrent with this report, we are releasing Alchemy as public resource, together with a suite of analysis tools and sample agent trajectories.

PDF Details

NeurIPS Conference 2021 Conference Paper

SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decomposition

Rishabh Kabra
Daniel Zoran
Goker Erdogan
Loic Matthey
Antonia Creswell
Matt Botvinick
Alexander Lerchner
Chris Burgess

To help agents reason about scenes in terms of their building blocks, we wish to extract the compositional structure of any given scene (in particular, the configuration and characteristics of objects comprising the scene). This problem is especially difficult when scene structure needs to be inferred while also estimating the agent’s location/viewpoint, as the two variables jointly give rise to the agent’s observations. We present an unsupervised variational approach to this problem. Leveraging the shared structure that exists across different scenes, our model learns to infer two sets of latent representations from RGB video input alone: a set of "object" latents, corresponding to the time-invariant, object-level contents of the scene, as well as a set of "frame" latents, corresponding to global time-varying elements such as viewpoint. This factorization of latents allows our model, SIMONe, to represent object attributes in an allocentric manner which does not depend on viewpoint. Moreover, it allows us to disentangle object dynamics and summarize their trajectories as time-abstracted, view-invariant, per-object properties. We demonstrate these capabilities, as well as the model's performance in terms of view synthesis and instance segmentation, across three procedurally generated video datasets.

PDF Details

ICLR Conference 2020 Conference Paper

Unsupervised Model Selection for Variational Disentangled Representation Learning

Sunny Duan
Loic Matthey
Andre Saraiva 0001
Nick Watters
Chris Burgess 0001
Alexander Lerchner
Irina Higgins

Disentangled representations have recently been shown to improve fairness, data efficiency and generalisation in simple supervised and reinforcement learning tasks. To extend the benefits of disentangled representations to more complex domains and practical applications, it is important to enable hyperparameter tuning and model selection of existing unsupervised approaches without requiring access to ground truth attribute labels, which are not available for most datasets. This paper addresses this problem by introducing a simple yet robust and reliable method for unsupervised disentangled model selection. We show that our approach performs comparably to the existing supervised alternatives across 5400 models from six state of the art unsupervised disentangled representation learning model classes. Furthermore, we show that the ranking produced by our approach correlates well with the final task performance on two different domains.

Details

ICML Conference 2019 Conference Paper

Multi-Object Representation Learning with Iterative Variational Inference

Klaus Greff
Raphaël Lopez Kaufman
Rishabh Kabra
Nick Watters
Chris Burgess 0001
Daniel Zoran
Loic Matthey
Matthew M. Botvinick

Human perception is structured around objects which form the basis for our higher-level cognition and impressive systematic generalization abilities. Yet most work on representation learning focuses on feature learning without even considering multiple objects, or treats segmentation as an (often supervised) preprocessing step. Instead, we argue for the importance of learning to segment and represent objects jointly. We demonstrate that, starting from the simple assumption that a scene is composed of multiple entities, it is possible to learn to segment images into interpretable objects with disentangled representations. Our method learns – without supervision – to inpaint occluded parts, and extrapolates to scenes with more objects and to unseen objects with novel feature combinations. We also show that, due to the use of iterative variational inference, our system is able to learn multi-modal posteriors for ambiguous inputs and extends naturally to sequences.

Details

NeurIPS Conference 2018 Conference Paper

Life-Long Disentangled Representation Learning with Cross-Domain Latent Homologies

Alessandro Achille
Tom Eccles
Loic Matthey
Chris Burgess
Nicholas Watters
Alexander Lerchner
Irina Higgins

Intelligent behaviour in the real-world requires the ability to acquire new knowledge from an ongoing sequence of experiences while preserving and reusing past knowledge. We propose a novel algorithm for unsupervised representation learning from piece-wise stationary visual data: Variational Autoencoder with Shared Embeddings (VASE). Based on the Minimum Description Length principle, VASE automatically detects shifts in the data distribution and allocates spare representational capacity to new knowledge, while simultaneously protecting previously learnt representations from catastrophic forgetting. Our approach encourages the learnt representations to be disentangled, which imparts a number of desirable properties: VASE can deal sensibly with ambiguous inputs, it can enhance its own representations through imagination-based exploration, and most importantly, it exhibits semantically meaningful sharing of latents between different datasets. Compared to baselines with entangled representations, our approach is able to reason beyond surface-level statistics and perform semantically meaningful cross-domain inference.

PDF Details

ICML Conference 2017 Conference Paper

DARLA: Improving Zero-Shot Transfer in Reinforcement Learning

Irina Higgins
Arka Pal
Andrei A. Rusu
Loic Matthey
Chris Burgess 0001
Alexander Pritzel
Matthew M. Botvinick
Charles Blundell

Domain adaptation is an important open problem in deep reinforcement learning (RL). In many scenarios of interest data is hard to obtain, so agents may learn a source policy in a setting where data is readily available, with the hope that it generalises well to the target domain. We propose a new multi-stage RL agent, DARLA (DisentAngled Representation Learning Agent), which learns to see before learning to act. DARLA’s vision is based on learning a disentangled representation of the observed environment. Once DARLA can see, it is able to acquire source policies that are robust to many domain shifts – even with no access to the target domain. DARLA significantly outperforms conventional baselines in zero-shot domain adaptation scenarios, an effect that holds across a variety of RL environments (Jaco arm, DeepMind Lab) and base RL algorithms (DQN, A3C and EC).

Details

ICRA Conference 2009 Conference Paper

Stochastic strategies for a swarm robotic assembly system

Loic Matthey
Spring Berman
Vijay Kumar 0001

We present a decentralized, scalable approach to assembling a group of heterogeneous parts into different products using a swarm of robots. While the assembly plans are predetermined, the exact sequence of assembly of parts and the allocation of subassembly tasks to robots are determined by the interactions between robots in a decentralized fashion in real time. Our approach is based on developing a continuous abstraction of the system derived from models of chemical reactions and formulating the strategy as a problem of selecting rates of assembly and disassembly. These rates are mapped onto probabilities that determine stochastic control policies for individual robots, which then produce the desired aggregate behavior. This top-down approach to determining robot controllers also allows us to optimize the rates at the abstract level to achieve fast convergence to the specified target numbers of products. Because the method incorporates programs for assembly and disassembly, changes in demand can lead to reconfiguration in a seamless fashion. We illustrate the methodology using a physics-based simulator with examples involving 15 robots and two types of final products.

Details

ICRA Conference 2008 Conference Paper

A comparison of casting and spiraling algorithms for odor source localization in laminar flow

Thomas Lochmatter
Xavier Raemy
Loic Matthey
Saurabh Indra
Alcherio Martinoli

We compare two well-known algorithms for locating odor sources in environments with a main wind flow. Their plume tracking performance is tested through systematic experiments with real robots in a wind tunnel under laminar flow condition. We present the system setup and show the wind and odor profiles. The results are then compared in terms of time and distance to reach the source, as well as speed in upwind direction. We conclude that the spiral-surge algorithm yields significantly better results than the casting algorithm, and discuss possible rationales behind this performance difference.

Details

IROS Conference 2008 Conference Paper

Experimental study of limit cycle and chaotic controllers for the locomotion of centipede robots

Loic Matthey
Ludovic Righetti
Auke Jan Ijspeert

In this contribution we present a CPG (central pattern generator) controller based on coupled Rössler systems. It is able to generate both limit cycle and chaotic behaviors through bifurcation. We develop an experimental test bench to measure quantitatively the performance of different controllers on unknown terrains of increasing difficulty. First, we show that for flat terrains, open loop limit cycle systems are the most efficient (in terms of speed of locomotion) but that they are quite sensitive to environmental changes. Second, we show that sensory feedback is a crucial addition for unknown terrains. Third, we show that the chaotic controller with sensory feedback outperforms the other controllers in very difficult terrains and actually promotes the emergence of short synchronized movement patterns. All that is done using an unified framework for the generation of limit cycle and chaotic behaviors, where a simple parameter change can switch from one behavior to the other through bifurcation. Such flexibility would allow the automatic adaptation of the robot locomotion strategy to the terrain uncertainty.

Details