Arrow Research search

Author name cluster

Chenghao Liu

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.

29 papers
2 author rows

Possible papers

29

AAAI Conference 2026 Conference Paper

TTF-VLA: Temporal Token Fusion via Pixel-Attention Integration for Vision-Language-Action Models

  • Chenghao Liu
  • Jiachen Zhang
  • Chengxuan Li
  • Zhimu Zhou
  • Shixin Wu
  • Songfang Huang
  • Huiling Duan

Vision-Language-Action (VLA) models process visual inputs independently at each timestep, discarding valuable temporal information inherent in robotic manipulation tasks. This frame-by-frame processing makes models vulnerable to visual noise while ignoring the substantial coherence between consecutive frames in manipulation sequences. We propose Temporal Token Fusion (TTF), a training-free approach that intelligently integrates historical and current visual representations to enhance VLA inference quality. Our method employs dual-dimension detection combining efficient grayscale pixel difference analysis with attention-based semantic relevance assessment, enabling selective temporal token fusion through hard fusion strategies and keyframe anchoring to prevent error accumulation. Comprehensive experiments across LIBERO, SimplerEnv, and real robot tasks demonstrate consistent improvements: 4.0 percentage points average on LIBERO (72.4% vs 68.4% baseline), cross-environment validation on SimplerEnv (4.8% relative improvement), and 8.7% relative improvement on real robot tasks. Our approach proves model-agnostic, working across OpenVLA and VLA-Cache architectures. Notably, TTF reveals that selective Query matrix reuse in attention mechanisms enhances rather than compromises performance, suggesting promising directions for direct KQV matrix reuse strategies that achieve computational acceleration while improving task success rates.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

Fast Projection-Free Approach (without Optimization Oracle) for Optimization over Compact Convex Set

  • Chenghao Liu
  • Enming Liang
  • Minghua Chen

Projection-free first-order methods, e. g. , the celebrated Frank-Wolfe (FW) algorithms, have emerged as powerful tools for optimization over simple convex sets such as polyhedra, because of their scalability, fast convergence, and iteration-wise feasibility without costly projections. However, extending these methods effectively to general compact convex sets remains challenging and largely open, as FW methods rely on expensive linear optimization oracles (LOO), while penalty-based methods often struggle with poor feasibility. We tackle this open challenge by presenting **Hom-PGD**, a novel projection-free method without expensive (optimization) oracles. Our method constructs a homeomorphism between the convex constraint set and a unit ball, transforming the original problem into an equivalent ball-constrained formulation, thus enabling efficient gradient-based optimization while preserving the original problem structure. We prove that Hom-PGD attains *optimal* convergence rates matching gradient descent with constant step-size to find an $\epsilon$-approximate (stationary) solution: $\mathcal{O}(\log (1/\epsilon))$ for strongly convex objectives, $\mathcal{O}(\epsilon^{-1})$ for convex objectives, and $\mathcal{O}(\epsilon^{-2})$ for non-convex objectives. Meanwhile, Hom-PGD enjoys a low per-iteration complexity of $\mathcal{O}(n^2)$, without expensive oracles like LOO or projection, where $n$ is the input size. Our framework further extends to certain non-convex sets, broadening its applicability in practical optimization scenarios with complex constraints. Extensive numerical experiments demonstrate that Hom-PGD achieves comparable convergence rates to state-of-the-art projection-free methods, while significantly reducing per-iteration runtime (up to 5 orders of magnitude faster) and thus the total problem-solving time.

PDF Details

ICML Conference 2025 Conference Paper

Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Experts

  • Xu Liu 0014
  • Juncheng Liu
  • Gerald Woo
  • Taha Aksu
  • Yuxuan Liang 0002
  • Roger Zimmermann
  • Chenghao Liu
  • Junnan Li 0001

Achieving effective unified pretraining on large time series corpora remains an open challenge in developing time series foundation models. Existing methods, such as Moirai, introduce multiple projection layers for time series of different frequencies to account for high data heterogeneity. We identify major drawbacks to this human-imposed frequency-level model specialization. First, frequency is not a reliable indicator for grouping pretraining data. Second, time series can display varied distributions even within a short window. Frequency-level specialization overlooks the diversity at this granularity. To address these issues, this paper introduces Moirai-MoE, excluding human-defined data groupings while delegating the modeling of diverse time series patterns to the sparse mixture of experts (MoE) within Transformers. With this design, Moirai-MoE eliminates reliance on heuristics and enables automatic token-level specialization. Extensive evaluations on 39 datasets demonstrate the superiority of Moirai-MoE over state-of-the-art foundation models. This study also conducts comprehensive model analyses to explore the inner workings of time series MoE foundation models.

Details

NeurIPS Conference 2025 Conference Paper

Multi-Scale Finetuning for Encoder-based Time Series Foundation Models

  • Zhongzheng Qiao
  • Chenghao Liu
  • Yiming Zhang
  • Ming Jin
  • Quang Pham
  • Qingsong Wen
  • Ponnuthurai Suganthan
  • Xudong Jiang

Time series foundation models (TSFMs) demonstrate impressive zero-shot performance for time series forecasting. However, an important yet underexplored challenge is how to effectively finetune TSFMs on specific downstream tasks. While naive finetuning can yield performance gains, we argue that it falls short of fully leveraging TSFMs' capabilities, often resulting in overfitting and suboptimal performance. Given the diverse temporal patterns across sampling scales and the inherent multi-scale forecasting capabilities of TSFMs, we adopt a causal perspective to analyze finetuning process, through which we highlight the critical importance of explicitly modeling multiple scales and reveal the shortcomings of naive approaches. Focusing on encoder-based TSFMs, we propose Multiscale finetuning (MSFT), a simple yet general framework that explicitly integrates multi-scale modeling into the finetuning process. Experimental results on three different backbones (Moirai, Moment and Units) demonstrate that TSFMs finetuned with MSFT not only outperform naive and typical parameter efficient finetuning methods but also surpass state-of-the-art deep learning methods. Codes are available at https: //github. com/zqiao11/MSFT.

PDF Details

NeurIPS Conference 2025 Conference Paper

ShapeX: Shapelet-Driven Post Hoc Explanations for Time Series Classification Models

  • Bosong Huang
  • Ming Jin
  • Yuxuan Liang
  • Johan Barthelemy
  • Debo Cheng
  • Qingsong Wen
  • Chenghao Liu
  • Shirui Pan

Explaining time series classification models is crucial, particularly in high-stakes applications such as healthcare and finance, where transparency and trust play a critical role. Although numerous time series classification methods have identified key subsequences, known as shapelets, as core features for achieving state-of-the-art performance and validating their pivotal role in classification outcomes, existing post-hoc time series explanation (PHTSE) methods primarily focus on timestep-level feature attribution. These explanation methods overlook the fundamental prior that classification outcomes are predominantly driven by key shapelets. To bridge this gap, we present ShapeX, an innovative framework that segments time series into meaningful shapelet-driven segments and employs Shapley values to assess their saliency. At the core of ShapeX lies the Shapelet Describe-and-Detect (SDD) framework, which effectively learns a diverse set of shapelets essential for classification. We further demonstrate that ShapeX produces explanations which reveal causal relationships instead of just correlations, owing to the atomicity properties of shapelets. Experimental results on both synthetic and real-world datasets demonstrate that ShapeX outperforms existing methods in identifying the most relevant subsequences, enhancing both the precision and causal fidelity of time series explanations.

PDF Details

TMLR Journal 2025 Journal Article

UniTST: Effectively Modeling Inter-Series and Intra-Series Dependencies for Multivariate Time Series Forecasting

  • Juncheng Liu
  • Chenghao Liu
  • Gerald Woo
  • Yiwei Wang
  • Bryan Hooi
  • Caiming Xiong
  • Doyen Sahoo

Transformer-based models have emerged as powerful tools for multivariate time series forecasting (MTSF). However, existing Transformer models often fall short of capturing both intricate dependencies across variate and temporal dimensions in MTS data. Some recent models are proposed to separately capture variate and temporal dependencies through either two sequential or parallel attention mechanisms. However, these methods cannot directly and explicitly learn the intricate inter-series and intra-series dependencies. In this work, we first demonstrate that these dependencies are very important as they usually exist in real-world data. To directly model these dependencies, we propose a transformer-based model UniTST containing a unified attention mechanism on the flattened patch tokens. Additionally, we add a dispatcher module which reduces the complexity and makes the model feasible for a potentially large number of variates. Although our proposed model employs a simple architecture, it offers compelling performance as shown in our extensive experiments on several datasets for time series forecasting.

PDF Details

ICML Conference 2025 Conference Paper

VisionTS: Visual Masked Autoencoders Are Free-Lunch Zero-Shot Time Series Forecasters

  • Mouxiang Chen
  • Lefei Shen
  • Zhuo Li 0014
  • Xiaoyun Joy Wang
  • Jianling Sun
  • Chenghao Liu

Foundation models have emerged as a promising approach in time series forecasting (TSF). Existing approaches either repurpose large language models (LLMs) or build large-scale time series datasets to develop TSF foundation models for universal forecasting. However, these methods face challenges due to the severe cross-domain gap or in-domain heterogeneity. This paper explores a new road to building a TSF foundation model from rich, high-quality natural images. Our key insight is that a visual masked autoencoder, pre-trained on the ImageNet dataset, can naturally be a numeric series forecaster. By reformulating TSF as an image reconstruction task, we bridge the gap between image pre-training and TSF downstream tasks. Surprisingly, without further adaptation in the time series domain, the proposed VisionTS could achieve better zero-shot forecast performance than existing TSF foundation models. With fine-tuning for one epoch, VisionTS could further improve the forecasting and achieve state-of-the-art performance in most cases. Extensive experiments reveal intrinsic similarities between images and real-world time series, suggesting that visual models may offer a "free lunch" for TSF and highlight the potential for future cross-modality research. Our code is available in the https: //github. com/Keytoyze/VisionTS.

Details

ICML Conference 2024 Conference Paper

Characterizing ResNet's Universal Approximation Capability

  • Chenghao Liu
  • Enming Liang
  • Minghua Chen 0001

Since its debut in 2016, ResNet has become arguably the most favorable architecture in deep neural network (DNN) design. It effectively addresses the gradient vanishing/exploding issue in DNN training, allowing engineers to fully unleash DNN’s potential in tackling challenging problems in various domains. Despite its practical success, an essential theoretical question remains largely open: how well/best can ResNet approximate functions? In this paper, we answer this question for several important function classes, including polynomials and smooth functions. In particular, we show that ResNet with constant width can approximate Lipschitz continuous function with a Lipschitz constant $\mu$ using $\mathcal{O}(c(d)(\varepsilon/\mu)^{-d/2})$ tunable weights, where $c(d)$ is a constant depending on the input dimension $d$ and $\epsilon>0$ is the target approximation error. Further, we extend such a result to Lebesgue-integrable functions with the upper bound characterized by the modulus of continuity. These results indicate a factor of $d$ reduction in the number of tunable weights compared with the classical results for ReLU networks. Our results are also order-optimal in $\varepsilon$, thus achieving optimal approximation rate, as they match a generalized lower bound derived in this paper. This work adds to the theoretical justifications for ResNet’s stellar practical performance.

Details

YNIMG Journal 2024 Journal Article

Deep learning enables accurate brain tissue microstructure analysis based on clinically feasible diffusion magnetic resonance imaging

  • Yuxing Li
  • Zhizheng Zhuo
  • Chenghao Liu
  • Yunyun Duan
  • Yulu Shi
  • Tingting Wang
  • Runzhi Li
  • Yanli Wang

Diffusion magnetic resonance imaging (dMRI) allows non-invasive assessment of brain tissue microstructure. Current model-based tissue microstructure reconstruction techniques require a large number of diffusion gradients, which is not clinically feasible due to imaging time constraints, and this has limited the use of tissue microstructure information in clinical settings. Recently, approaches based on deep learning (DL) have achieved promising tissue microstructure reconstruction results using clinically feasible dMRI. However, it remains unclear whether the subtle tissue changes associated with disease or age are properly preserved with DL approaches and whether DL reconstruction results can benefit clinical applications. Here, we provide the first evidence that DL approaches to tissue microstructure reconstruction yield reliable brain tissue microstructure analysis based on clinically feasible dMRI scans. Specifically, we reconstructed tissue microstructure from four different brain dMRI datasets with only 12 diffusion gradients, a clinically feasible protocol, and the neurite orientation dispersion and density imaging (NODDI) and spherical mean technique (SMT) models were considered. With these results we show that disease-related and age-dependent alterations of brain tissue were accurately identified. These findings demonstrate that DL tissue microstructure reconstruction can accurately quantify microstructural alterations in the brain based on clinically feasible dMRI.

Details DOI

ICML Conference 2024 Conference Paper

Identifiability Matters: Revealing the Hidden Recoverable Condition in Unbiased Learning to Rank

  • Mouxiang Chen
  • Chenghao Liu
  • Zemin Liu
  • Zhuo Li 0014
  • Jianling Sun

Unbiased Learning to Rank (ULTR) aims to train unbiased ranking models from biased click logs, by explicitly modeling a generation process for user behavior and fitting click data based on examination hypothesis. Previous research found empirically that the true latent relevance is mostly recoverable through click fitting. However, we demonstrate that this is not always achievable, resulting in a significant reduction in ranking performance. This research investigates the conditions under which relevance can be recovered from click data in the first principle. We initially characterize a ranking model as identifiable if it can recover the true relevance up to a scaling transformation, a criterion sufficient for the pairwise ranking objective. Subsequently, we investigate an equivalent condition for identifiability, articulated as a graph connectivity test problem: the recovery of relevance is feasible if and only if the identifiability graph (IG), derived from the underlying structure of the dataset, is connected. The presence of a disconnected IG may lead to degenerate cases and suboptimal ranking performance. To tackle this challenge, we introduce two methods, namely node intervention and node merging, designed to modify the dataset and restore the connectivity of the IG. Empirical results derived from a simulated dataset and two real-world LTR benchmark datasets not only validate our proposed theory, but also demonstrate the effectiveness of our methods in alleviating data bias when the relevance model is unidentifiable.

Details

ICML Conference 2024 Conference Paper

ReLU Network with Width d+O(1) Can Achieve Optimal Approximation Rate

  • Chenghao Liu
  • Minghua Chen 0001

The prevalent employment of narrow neural networks, characterized by their minimal parameter count per layer, has led to a surge in research exploring their potential as universal function approximators. A notable result in this field states that networks with just a width of $d+1$ can approximate any continuous function for input dimension $d$ arbitrarily well. However, the optimal approximation rate for these narrowest networks, i. e. , the optimal relation between the count of tunable parameters and the approximation error, remained unclear. In this paper, we address this gap by proving that ReLU networks with width $d+1$ can achieve the optimal approximation rate for continuous functions over the domain $[0, 1]^d$ under $L^p$ norm for $p\in[1, \infty)$. We further show that for the uniform norm, a width of $d+11$ is sufficient. We also extend the results to narrow feed-forward networks with various activations, confirming their capability to approximate at the optimal rate. This work adds to the understanding of universal approximation of narrow networks.

Details

NeurIPS Conference 2024 Conference Paper

Time-FFM: Towards LM-Empowered Federated Foundation Model for Time Series Forecasting

  • Qingxiang Liu
  • Xu Liu
  • Chenghao Liu
  • Qingsong Wen
  • Yuxuan Liang

Unlike natural language processing and computer vision, the development of Foundation Models (FMs) for time series forecasting is blocked due to data scarcity. While recent efforts are focused on building such FMs by unlocking the potential of language models (LMs) for time series analysis, dedicated parameters for various downstream forecasting tasks need training, which hinders the common knowledge sharing across domains. Moreover, data owners may hesitate to share the access to local data due to privacy concerns and copyright protection, which makes it impossible to simply construct a FM on cross-domain training instances. To address these issues, we propose Time-FFM, a Federated Foundation Model for Time series forecasting by leveraging pretrained LMs. Specifically, we begin by transforming time series into the modality of text tokens. To bootstrap LMs for time series reasoning, we propose a prompt adaption module to determine domain-customized prompts dynamically instead of artificially. Given the data heterogeneity across domains, we design a personalized federated training strategy by learning global encoders and local prediction heads. Our comprehensive experiments indicate that Time-FFM outperforms state-of-the-arts and promises effective few-shot and zero-shot forecaster. The code is available at https: //github. com/CityMind-Lab/NeurIPS24-Time-FFM/tree/main.

PDF Details DOI

ICML Conference 2024 Conference Paper

Unified Training of Universal Time Series Forecasting Transformers

  • Gerald Woo
  • Chenghao Liu
  • Akshat Kumar
  • Caiming Xiong
  • Silvio Savarese
  • Doyen Sahoo

Deep learning for time series forecasting has traditionally operated within a one-model-per-dataset framework, limiting its potential to leverage the game-changing impact of large pre-trained models. The concept of universal forecasting, emerging from pre-training on a vast collection of time series datasets, envisions a single Large Time Series Model capable of addressing diverse downstream forecasting tasks. However, constructing such a model poses unique challenges specific to time series data: (i) cross-frequency learning, (ii) accommodating an arbitrary number of variates for multivariate time series, and (iii) addressing the varying distributional properties inherent in large-scale data. To address these challenges, we present novel enhancements to the conventional time series Transformer architecture, resulting in our proposed M asked Enc o der-based Un i ve r s a l T i me Series Forecasting Transformer ( Moirai ). Trained on our newly introduced Large-scale Open Time Series Archive (LOTSA) featuring over 27B observations across nine domains, Moirai achieves competitive or superior performance as a zero-shot forecaster when compared to full-shot models. Code, data, and model weights can be found at https: //github. com/SalesforceAIResearch/uni2ts.

Details

YNIMG Journal 2023 Journal Article

CarveMix: A simple data augmentation method for brain lesion segmentation

  • Xinru Zhang
  • Chenghao Liu
  • Ni Ou
  • Xiangzhu Zeng
  • Zhizheng Zhuo
  • Yunyun Duan
  • Xiaoliang Xiong
  • Yizhou Yu

Brain lesion segmentation provides a valuable tool for clinical diagnosis and research, and convolutional neural networks (CNNs) have achieved unprecedented success in the segmentation task. Data augmentation is a widely used strategy to improve the training of CNNs. In particular, data augmentation approaches that mix pairs of annotated training images have been developed. These methods are easy to implement and have achieved promising results in various image processing tasks. However, existing data augmentation approaches based on image mixing are not designed for brain lesions and may not perform well for brain lesion segmentation. Thus, the design of this type of simple data augmentation method for brain lesion segmentation is still an open problem. In this work, we propose a simple yet effective data augmentation approach, dubbed as CarveMix, for CNN-based brain lesion segmentation. Like other mixing-based methods, CarveMix stochastically combines two existing annotated images (annotated for brain lesions only) to obtain new labeled samples. To make our method more suitable for brain lesion segmentation, CarveMix is lesion-aware, where the image combination is performed with a focus on the lesions and preserves the lesion information. Specifically, from one annotated image we carve a region of interest (ROI) according to the lesion location and geometry with a variable ROI size. The carved ROI then replaces the corresponding voxels in a second annotated image to synthesize new labeled images for network training, and additional harmonization steps are applied for heterogeneous data where the two annotated images can originate from different sources. Besides, we further propose to model the mass effect that is unique to whole brain tumor segmentation during image mixing. To evaluate the proposed method, experiments were performed on multiple publicly available or private datasets, and the results show that our method improves the accuracy of brain lesion segmentation. The code of the proposed method is available at https://github.com/ZhangxinruBIT/CarveMix.git.

Details DOI

IJCAI Conference 2023 Conference Paper

FedET: A Communication-Efficient Federated Class-Incremental Learning Framework Based on Enhanced Transformer

  • Chenghao Liu
  • Xiaoyang Qu
  • Jianzong Wang
  • Jing Xiao

Federated Learning (FL) has been widely concerned for it enables decentralized learning while ensuring data privacy. However, most existing methods unrealistically assume that the classes encountered by local clients are fixed over time. After learning new classes, this impractical assumption will make the model's catastrophic forgetting of old classes significantly severe. Moreover, due to the limitation of communication cost, it is challenging to use large-scale models in FL, which will affect the prediction accuracy. To address these challenges, we propose a novel framework, Federated Enhanced Transformer (FedET), which simultaneously achieves high accuracy and low communication cost. Specifically, FedET uses Enhancer, a tiny module, to absorb and communicate new knowledge, and applies pre-trained Transformers combined with different Enhancers to ensure high precision on various tasks. To address local forgetting caused by new classes of new tasks and global forgetting brought by non-i. i. d class imbalance across different local clients, we proposed an Enhancer distillation method to modify the imbalance between old and new knowledge and repair the non-i. i. d. problem. Experimental results demonstrate that FedET's average accuracy on a representative benchmark dataset is 14. 1% higher than the state-of-the-art method, while FedET saves 90% of the communication cost compared to the previous method.

PDF Details DOI

ICML Conference 2023 Conference Paper

Learning Deep Time-index Models for Time Series Forecasting

  • Gerald Woo
  • Chenghao Liu
  • Doyen Sahoo
  • Akshat Kumar
  • Steven C. H. Hoi

Deep learning has been actively applied to time series forecasting, leading to a deluge of new methods, belonging to the class of historical-value models. Yet, despite the attractive properties of time-index models, such as being able to model the continuous nature of underlying time series dynamics, little attention has been given to them. Indeed, while naive deep time-index models are far more expressive than the manually predefined function representations of classical time-index models, they are inadequate for forecasting, being unable to generalize to unseen time steps due to the lack of inductive bias. In this paper, we propose DeepTime, a meta-optimization framework to learn deep time-index models which overcome these limitations, yielding an efficient and accurate forecasting model. Extensive experiments on real world datasets in the long sequence time-series forecasting setting demonstrate that our approach achieves competitive results with state-of-the-art methods, and is highly efficient. Code is available at https: //github. com/salesforce/DeepTime.

Details

ICLR Conference 2023 Conference Paper

Learning Fast and Slow for Online Time Series Forecasting

  • Quang Pham
  • Chenghao Liu
  • Doyen Sahoo
  • Steven C. H. Hoi

Despite the recent success of deep learning for time series forecasting, these methods are not scalable for many real-world applications where data arrives sequentially. Training deep neural forecasters on the fly is notoriously challenging because of their limited ability to adapt to non-stationary environments and remember old knowledge. We argue that the fast adaptation capability of deep neural networks is critical and successful solutions require handling changes to both new and recurring patterns effectively. In this work, inspired by the Complementary Learning Systems (CLS) theory, we propose Fast and Slow learning Network (FSNet) as a novel framework to address the challenges of online forecasting. Particularly, FSNet improves the slowly-learned backbone by dynamically balancing fast adaptation to recent changes and retrieving similar old knowledge. FSNet achieves this mechanism via an interaction between two novel complementary components: (i) a per-layer adapter to support fast learning from individual layers, and (ii) an associative memory to support remembering, updating, and recalling repeating events. Extensive experiments on real and synthetic datasets validate FSNet's efficacy and robustness to both new and recurring patterns. Our code is publicly available at: \url{https://github.com/salesforce/fsnet/}.

Details

JMLR Journal 2023 Journal Article

Merlion: End-to-End Machine Learning for Time Series

  • Aadyot Bhatnagar
  • Paul Kassianik
  • Chenghao Liu
  • Tian Lan
  • Wenzhuo Yang
  • Rowan Cassius
  • Doyen Sahoo
  • Devansh Arpit

We introduce Merlion, an open-source machine learning library for time series. It features a unified interface for many commonly used models and datasets for forecasting and anomaly detection on both univariate and multivariate time series, along with standard pre/post-processing layers. It has several modules to improve ease-of-use, including a no-code visual dashboard, anomaly score calibration to improve interpetability, AutoML for hyperparameter tuning and model selection, and model ensembling. Merlion also provides an evaluation framework that simulates the live deployment of a model in production, and a distributed computing backend to run time series models at industrial scale. This library aims to provide engineers and researchers a one-stop solution to rapidly develop models for their specific time series needs and benchmark them across multiple datasets. [abs] [ pdf ][ bib ] [ code ] &copy JMLR 2023. ( edit, beta )

PDF Details

ICLR Conference 2022 Conference Paper

Continual Normalization: Rethinking Batch Normalization for Online Continual Learning

  • Quang Pham
  • Chenghao Liu
  • Steven C. H. Hoi

Existing continual learning methods use Batch Normalization (BN) to facilitate training and improve generalization across tasks. However, the non-i.i.d and non-stationary nature of continual learning data, especially in the online setting, amplify the discrepancy between training and testing in BN and hinder the performance of older tasks. In this work, we study the cross-task normalization effect of BN in online continual learning where BN normalizes the testing data using moments biased towards the current task, resulting in higher catastrophic forgetting. This limitation motivates us to propose a simple yet effective method that we call Continual Normalization (CN) to facilitate training similar to BN while mitigating its negative effect. Extensive experiments on different continual learning algorithms and online scenarios show that CN is a direct replacement for BN and can provide substantial performance improvements. Our implementation will be made publicly available upon acceptance.

Details

ICLR Conference 2022 Conference Paper

CoST: Contrastive Learning of Disentangled Seasonal-Trend Representations for Time Series Forecasting

  • Gerald Woo
  • Chenghao Liu
  • Doyen Sahoo
  • Akshat Kumar
  • Steven C. H. Hoi

Deep learning has been actively studied for time series forecasting, and the mainstream paradigm is based on the end-to-end training of neural network architectures, ranging from classical LSTM/RNNs to more recent TCNs and Transformers. Motivated by the recent success of representation learning in computer vision and natural language processing, we argue that a more promising paradigm for time series forecasting, is to first learn disentangled feature representations, followed by a simple regression fine-tuning step -- we justify such a paradigm from a causal perspective. Following this principle, we propose a new time series representation learning framework for long sequence time series forecasting named CoST, which applies contrastive learning methods to learn disentangled seasonal-trend representations. CoST comprises both time domain and frequency domain contrastive losses to learn discriminative trend and seasonal representations, respectively. Extensive experiments on real-world datasets show that CoST consistently outperforms the state-of-the-art methods by a considerable margin, achieving a 21.3% improvement in MSE on multivariate benchmarks. It is also robust to various choices of backbone encoders, as well as downstream regressors. Code is available at https://github.com/salesforce/CoST.

Details

NeurIPS Conference 2022 Conference Paper

LBD: Decouple Relevance and Observation for Individual-Level Unbiased Learning to Rank

  • Mouxiang Chen
  • Chenghao Liu
  • Zemin Liu
  • Jianling Sun

Using Unbiased Learning to Rank (ULTR) to train the ranking model with biased click logs has attracted increased research interest. The key idea is to explicitly model the user's observation behavior when building the ranker with a large number of click logs. Considering the simplicity, recent efforts are mainly based on the position bias hypothesis, in which the observation only depends on the position. However, this hypothesis does not hold in many scenarios due to the neglect of the distinct characteristics of individuals in the same position. On the other hand, directly modeling observation bias for each individual is quite challenging, since the effects of each individual's features on relevance and observation are entangled. It is difficult to ravel out this coupled effect and thus obtain a correct relevance model from click data. To address this issue, we first present the concept of coupling effect for individual-level ULTR. Then, we develop the novel Lipschitz and Bernoulli Decoupling (LBD) model to decouple the effects on relevance and observation at the individual level. We prove theoretically that our proposed method could recover the correct relevance order for the ranking objective. Empirical results on two LTR benchmark datasets show that the proposed model outperforms the state-of-the-art baselines and verify its effectiveness in debiasing data.

PDF Details

ICLR Conference 2021 Conference Paper

Contextual Transformation Networks for Online Continual Learning

  • Quang Pham
  • Chenghao Liu
  • Doyen Sahoo
  • Steven C. H. Hoi

Continual learning methods with fixed architectures rely on a single network to learn models that can perform well on all tasks. As a result, they often only accommodate common features of those tasks but neglect each task's specific features. On the other hand, dynamic architecture methods can have a separate network for each task, but they are too expensive to train and not scalable in practice, especially in online settings. To address this problem, we propose a novel online continual learning method named ``Contextual Transformation Networks” (CTN) to efficiently model the \emph{task-specific features} while enjoying neglectable complexity overhead compared to other fixed architecture methods. Moreover, inspired by the Complementary Learning Systems (CLS) theory, we propose a novel dual memory design and an objective to train CTN that can address both catastrophic forgetting and knowledge transfer simultaneously. Our extensive experiments show that CTN is competitive with a large scale dynamic architecture network and consistently outperforms other fixed architecture methods under the same standard backbone. Our implementation can be found at \url{https://github.com/phquang/Contextual-Transformation-Network}.

Details

NeurIPS Conference 2021 Conference Paper

DualNet: Continual Learning, Fast and Slow

  • Quang Pham
  • Chenghao Liu
  • Steven Hoi

According to Complementary Learning Systems (CLS) theory~\cite{mcclelland1995there} in neuroscience, humans do effective \emph{continual learning} through two complementary systems: a fast learning system centered on the hippocampus for rapid learning of the specifics and individual experiences, and a slow learning system located in the neocortex for the gradual acquisition of structured knowledge about the environment. Motivated by this theory, we propose a novel continual learning framework named ``DualNet", which comprises a fast learning system for supervised learning of pattern-separated representation from specific tasks and a slow learning system for unsupervised representation learning of task-agnostic general representation via a Self-Supervised Learning (SSL) technique. The two fast and slow learning systems are complementary and work seamlessly in a holistic continual learning framework. Our extensive experiments on two challenging continual learning benchmarks of CORE50 and miniImageNet show that DualNet outperforms state-of-the-art continual learning methods by a large margin. We further conduct ablation studies of different SSL objectives to validate DualNet's efficacy, robustness, and scalability. Code is publicly available at \url{https: //github. com/phquang/DualNet}.

PDF Details

IJCAI Conference 2021 Conference Paper

Node-wise Localization of Graph Neural Networks

  • Zemin Liu
  • Yuan Fang
  • Chenghao Liu
  • Steven C. H. Hoi

Graph neural networks (GNNs) emerge as a powerful family of representation learning models on graphs. To derive node representations, they utilize a global model that recursively aggregates information from the neighboring nodes. However, different nodes reside at different parts of the graph in different local contexts, making their distributions vary across the graph. Ideally, how a node receives its neighborhood information should be a function of its local context, to diverge from the global GNN model shared by all nodes. To utilize node locality without overfitting, we propose a node-wise localization of GNNs by accounting for both global and local aspects of the graph. Globally, all nodes on the graph depend on an underlying global GNN to encode the general patterns across the graph; locally, each node is localized into a unique model as a function of the global model and its local context. Finally, we conduct extensive experiments on four benchmark graphs, and consistently obtain promising performance surpassing the state-of-the-art GNNs.

PDF Details DOI

AAAI Conference 2021 Conference Paper

Relative and Absolute Location Embedding for Few-Shot Node Classification on Graph

  • Zemin Liu
  • Yuan Fang
  • Chenghao Liu
  • Steven C.H. Hoi

Node classification is an important problem on graphs. While recent advances in graph neural networks achieve promising performance, they require abundant labeled nodes for training. However, in many practical scenarios, there often exist novel classes in which only one or a few labeled nodes are available as supervision, known as few-shot node classification. Although meta-learning has been widely used in vision and language domains to address few-shot learning, its adoption on graphs has been limited. In particular, graph nodes in a few-shot task are not independent and relate to each other. To deal with this, we propose a novel model called Relative and Absolute Location Embedding (RALE) hinged on the concept of hub nodes. Specifically, RALE captures the task-level dependency by assigning each node a relative location within a task, as well as the graph-level dependency by assigning each node an absolute location on the graph to further align different tasks toward learning a transferable prior. Finally, extensive experiments on three public datasets demonstrate the state-of-the-art performance of RALE.

PDF Details

AAAI Conference 2019 Conference Paper

Discrete Social Recommendation

  • Chenghao Liu
  • Xin Wang
  • Tao Lu
  • Wenwu Zhu
  • Jianling Sun
  • Steven Hoi

Social recommendation, which aims at improving the performance of traditional recommender systems by considering social information, has attracted broad range of interests. As one of the most widely used methods, matrix factorization typically uses continuous vectors to represent user/item latent features. However, the large volume of user/item latent features results in expensive storage and computation cost, particularly on terminal user devices where the computation resource to operate model is very limited. Thus when taking extra social information into account, precisely extracting K most relevant items for a given user from massive candidates tends to consume even more time and memory, which imposes formidable challenges for efficient and accurate recommendations. A promising way is to simply binarize the latent features (obtained in the training phase) and then compute the relevance score through Hamming distance. However, such a two-stage hashing based learning procedure is not capable of preserving the original data geometry in the real-value space and may result in a severe quantization loss. To address these issues, this work proposes a novel discrete social recommendation (DSR) method which learns binary codes in a unified framework for users and items, considering social information. We further put the balanced and uncorrelated constraints on the objective to ensure the learned binary codes can be informative yet compact, and finally develop an efficient optimization algorithm to estimate the model parameters. Extensive experiments on three real-world datasets demonstrate that DSR runs nearly 5 times faster and consumes only with 1/37 of its real-value competitor’s memory usage at the cost of almost no loss in accuracy.

PDF Details

AAAI Conference 2018 Conference Paper

Unified Locally Linear Classifiers With Diversity-Promoting Anchor Points

  • Chenghao Liu
  • Teng Zhang
  • Peilin Zhao
  • Jianling Sun
  • Steven Hoi

Locally Linear Support Vector Machine (LLSVM) has been actively used in classification tasks due to its capability of classifying nonlinear patterns. However, existing LLSVM suffers from two drawbacks: (1) a particular and appropriate regularization for LLSVM has not yet been addressed; (2) it usually adopts a three-stage learning scheme composed of learning anchor points by clustering, learning local coding coordinates by a predefined coding scheme, and finally learning for training classifiers. We argue that this decoupled approaches oversimplifies the original optimization problem, resulting in a large deviation due to the disparate purpose of each step. To address the first issue, we propose a novel diversified regularization which could capture infrequent patterns and reduce the model size without sacrificing the representation power. Based on this regularization, we develop a joint optimization algorithm among anchor points, local coding coordinates and classifiers to simultaneously minimize the overall classification risk, which is termed as Diversified and Uni- fied Locally Linear Support Vector Machine(DU-LLSVM for short). To the best of our knowledge, DU-LLSVM is the first principled method that directly learns sparse local coding and can be easily generalized to other supervised learning models. Extensive experiments showed that DU-LLSVM consistently surpassed several state-of-the-art methods with a predefined local coding scheme (e. g. LLSVM) or a supervised anchor point learning(e. g. SAPL-LLSVM).

PDF Details

IJCAI Conference 2017 Conference Paper

Locally Linear Factorization Machines

  • Chenghao Liu
  • Teng Zhang
  • Peilin Zhao
  • Jun Zhou
  • Jianling Sun

Factorization Machines (FMs) are a widely used method for efficiently using high-order feature interactions in classification and regression tasks. Unfortunately, despite increasing interests in FMs, existing work only considers high order information of the input features which limits their capacities in non-linear problems and fails to capture the underlying structures of more complex data. In this work, we present a novel Locally Linear Factorization Machines (LLFM) which overcomes this limitation by exploring local coding technique. Unlike existing local coding classifiers that involve a phase of unsupervised anchor point learning and predefined local coding scheme which is suboptimal as the class label information is not exploited in discovering the encoding and thus can result in a suboptimal encoding for prediction, we formulate a joint optimization over the anchor points, local coding coordinates and FMs variables to minimize classification or regression risk. Empirically, we demonstrate that our approach achieves much better predictive accuracy than other competitive methods which employ LLFM with unsupervised anchor point learning and predefined local coding scheme.

PDF Details

AAAI Conference 2016 Conference Paper

Online ARIMA Algorithms for Time Series Prediction

  • Chenghao Liu
  • Steven C.H. Hoi
  • Peilin Zhao
  • Jianling Sun

Autoregressive integrated moving average (ARIMA) is one of the most popular linear models for time series forecasting due to its nice statistical properties and great flexibility. However, its parameters are estimated in a batch manner and its noise terms are often assumed to be strictly bounded, which restricts its applications and makes it inefficient for handling large-scale real data. In this paper, we propose online learning algorithms for estimating ARIMA models under relaxed assumptions on the noise terms, which is suitable to a wider range of applications and enjoys high computational efficiency. The idea of our ARIMA method is to reformulate the ARIMA model into a task of full information online optimization (without random noise terms). As a consequence, we can online estimation of the parameters in an efficient and scalable way. Furthermore, we analyze regret bounds of the proposed algorithms, which guarantee that our online ARIMA model is provably as good as the best ARIMA model in hindsight. Finally, our encouraging experimental results further validate the effectiveness and robustness of our method.

PDF Details
v2026.09.13