Author name cluster

Guanchu Wang

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.

13 papers

2 author rows

AAAI Conference 2026 Conference Paper

Catastrophic Forgetting in Kolmogorov-Arnold Networks

Mohammad Marufur Rahman
Guanchu Wang
Kaixiong Zhou
Minghan Chen
Fan Yang

Catastrophic forgetting is a longstanding challenge in continual learning, where models lose knowledge from earlier tasks when learning new ones. While various mitigation strategies have been proposed for Multi-Layer Perceptrons (MLPs), recent architectural advances like Kolmogorov-Arnold Networks (KANs) have been suggested to offer intrinsic resistance to forgetting by leveraging localized spline-based activations. However, the practical behavior of KANs under continual learning remains unclear, and their limitations are not well understood. To address this, we present a comprehensive study of catastrophic forgetting in KANs and develop a theoretical framework that links forgetting to activation support overlap and intrinsic data dimension. We validate these analyses through systematic experiments on synthetic and vision tasks, measuring forgetting dynamics under varying model configurations and data complexity. Further, we introduce KAN-LoRA, a novel adapter design for parameter-efficient continual fine-tuning of language models, and evaluate its effectiveness in knowledge editing tasks. Our findings reveal that while KANs exhibit promising retention in low-dimensional algorithmic settings, they remain vulnerable to forgetting in high-dimensional domains such as image classification and language modeling. These results advance the understanding of KANs’ strengths and limitations, offering practical insights for continual learning system design.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

Breaking the Frozen Subspace: Importance Sampling for Low-Rank Optimization in LLM Pretraining

Haochen Zhang
Junze Yin
Guanchu Wang
Zirui Liu
Lin Yang
Tianyi Zhang
Anshumali Shrivastava
Vladimir Braverman

Low-rank optimization has emerged as a promising approach to enabling memory-efficient training of large language models (LLMs). Existing low-rank optimization methods typically project gradients onto a low-rank subspace, reducing the memory cost of storing optimizer states. A key challenge in these methods is selecting suitable subspaces to ensure an effective optimization trajectory. Most existing approaches select the dominant subspace to preserve gradient information, as this intuitively provides the best approximation. However, we find that in practice, the dominant subspace stops changing during pretraining, thereby constraining weight updates to similar subspaces. In this paper, we propose importance sampling for low-rank optimization in LLM pretraining with a provable convergence guarantee, which the dominant subspace approach does not have. Empirically, we demonstrate that our method significantly outperforms previous methods in LLM pretraining tasks.

PDF Details

TMLR Journal 2025 Journal Article

Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models

Yang Sui
Yu-Neng Chuang
Guanchu Wang
Jiamu Zhang
Tianyi Zhang
Jiayi Yuan
Hongyi Liu
Andrew Wen

Large Language Models (LLMs) have demonstrated remarkable capabilities in complex tasks. Recent advancements in Large Reasoning Models (LRMs), such as OpenAI o1 and DeepSeek-R1, have further improved performance in System-2 reasoning domains like mathematics and programming by harnessing supervised fine-tuning (SFT) and reinforcement learning (RL) to enhance Chain-of-Thought (CoT) reasoning. However, while longer CoT reasoning sequences improve performance, they also introduce significant computational overhead due to lengthy and redundant outputs, known as the ''overthinking phenomenon''. Efficient Reasoning, which seeks to optimize reasoning length while preserving reasoning capabilities, offers practical benefits such as faster processing times, lower energy consumption, and improved responsiveness, especially valuable for reasoning-intensive applications. Despite its potential, efficient reasoning remains in the early stages of research. In this paper, we provide the first structured survey to systematically investigate and explore the current progress toward achieving efficient reasoning in LLMs. Overall, relying on the inherent mechanism of LLMs, we categorize existing works into several key directions: (1) model-based efficient reasoning, which considers optimizing full-length reasoning models into more concise reasoning models or directly training efficient reasoning models; (2) reasoning output-based efficient reasoning, which aims to dynamically reduce reasoning steps and length during inference; (3) input prompts-based efficient reasoning, which seeks to enhance reasoning efficiency based on input prompt properties such as difficulty or length control. Additionally, we introduce the use of efficient data for training reasoning models, explore the reasoning capabilities of small language models, and discuss evaluation methods and benchmarking.

PDF Details

ICML Conference 2024 Conference Paper

TVE: Learning Meta-attribution for Transferable Vision Explainer

Guanchu Wang
Yu-Neng Chuang
Fan Yang 0023
Mengnan Du
Chia-Yuan Chang 0002
Shaochen (Henry) Zhong
Zirui Liu 0001
Zhaozhuo Xu

Explainable machine learning significantly improves the transparency of deep neural networks. However, existing work is constrained to explaining the behavior of individual model predictions, and lacks the ability to transfer the explanation across various models and tasks. This limitation results in explaining various tasks being time- and resource-consuming. To address this problem, we introduce a Transferable Vision Explainer (TVE) that can effectively explain various vision models in downstream tasks. Specifically, the transferability of TVE is realized through a pre-training process on large-scale datasets towards learning the meta-attribution. This meta-attribution leverages the versatility of generic backbone encoders to comprehensively encode the attribution knowledge for the input instance, which enables TVE to seamlessly transfer to explaining various downstream tasks, without the need for training on task-specific data. Empirical studies involve explaining three different architectures of vision models across three diverse downstream datasets. The experiment results indicate TVE is effective in explaining these tasks without the need for additional training on downstream data.

Details

NeurIPS Conference 2023 Conference Paper

Chasing Fairness Under Distribution Shift: A Model Weight Perturbation Approach

Zhimeng (Stephen) Jiang
Xiaotian Han
Hongye Jin
Guanchu Wang
Rui Chen
Na Zou
Xia Hu

Fairness in machine learning has attracted increasing attention in recent years. The fairness methods improving algorithmic fairness for in-distribution data may not perform well under distribution shifts. In this paper, we first theoretically demonstrate the inherent connection between distribution shift, data perturbation, and model weight perturbation. Subsequently, we analyze the sufficient conditions to guarantee fairness (i. e. , low demographic parity) for the target dataset, including fairness for the source dataset, and low prediction difference between the source and target datasets for each sensitive attribute group. Motivated by these sufficient conditions, we propose robust fairness regularization (RFR) by considering the worst case within the model weight perturbation ball for each sensitive attribute group. We evaluate the effectiveness of our proposed RFR algorithm on synthetic and real distribution shifts across various datasets. Experimental results demonstrate that RFR achieves better fairness-accuracy trade-off performance compared with several baselines. The source code is available at \url{https: //github. com/zhimengj0326/RFR_NeurIPS23}.

PDF Details

ICLR Conference 2023 Conference Paper

CoRTX: Contrastive Framework for Real-time Explanation

Yu-Neng Chuang
Guanchu Wang
Fan Yang 0023
Quan Zhou
Pushkar Tripathi
Xuanting Cai
Xia Hu 0001

Recent advancements in explainable machine learning provide effective and faithful solutions for interpreting model behaviors. However, many explanation methods encounter efficiency issues, which largely limit their deployments in practical scenarios. Real-time explainer (RTX) frameworks have thus been proposed to accelerate the model explanation process by learning an one-feed-forward explainer. Existing RTX frameworks typically build the explainer under the supervised learning paradigm, which requires large amounts of explanation labels as the ground truth. Considering that accurate explanation labels are usually hard to obtain, due to constrained computational resources and limited human efforts, effective explainer training is still challenging in practice. In this work, we propose a COntrastive Real-Time eXplanation (CoRTX) framework to learn the explanation-oriented representation and relieve the intensive dependence of explainer training on explanation labels. Specifically, we design a synthetic strategy to select positive and negative instances for explanation representation learning. Theoretical analysis show that our selection strategy can benefit the contrastive learning process on explanation tasks. Experimental results on three real-world datasets further demonstrate the efficiency and efficacy of our proposed CoRTX framework.

Details

ICML Conference 2023 Conference Paper

DIVISION: Memory Efficient Training via Dual Activation Precision

Guanchu Wang
Zirui Liu 0001
Zhimeng Jiang
Ninghao Liu 0001
Na Zou 0001
Xia Hu 0001

Activation compressed training provides a solution towards reducing the memory cost of training deep neural networks (DNNs). However, state-of-the-art work combines a search of quantization bit-width with the training, which makes the procedure complicated and less transparent. To this end, we propose a simple and effective method to compress DNN training. Our method is motivated by an instructive observation: DNN backward propagation mainly utilizes the low-frequency component (LFC) of the activation maps, while the majority of memory is for caching the high-frequency component (HFC) during the training. This indicates the HFC of activation maps is highly redundant and compressible, which inspires our proposed Dual Activation Precision (DIVISION). During the training, DIVISION preserves a high-precision copy of LFC and compresses the HFC into a light-weight copy with low numerical precision. This can significantly reduce the memory cost while maintaining a competitive model accuracy. Experiment results show DIVISION has better comprehensive performance than state-of-the-art methods, including over 10x compression of activation maps and competitive training throughput, without loss of model accuracy. The source code is available at https: //github. com/guanchuwang/division.

Details

NeurIPS Conference 2023 Conference Paper

Winner-Take-All Column Row Sampling for Memory Efficient Adaptation of Language Model

Zirui Liu
Guanchu Wang
Shaochen (Henry) Zhong
Zhaozhuo Xu
Daochen Zha
Ruixiang (Ryan) Tang
Zhimeng (Stephen) Jiang
Kaixiong Zhou

As the model size grows rapidly, fine-tuning the large pre-trained language model has become increasingly difficult due to its extensive memory usage. Previous works usually focus on reducing the number of trainable parameters in the network. While the model parameters do contribute to memory usage, the primary memory bottleneck during training arises from storing feature maps, also known as activations, as they are crucial for gradient calculation. Notably, machine learning models are typically trained using stochastic gradient descent. We argue that in stochastic optimization, models can handle noisy gradients as long as the gradient estimator is unbiased with reasonable variance. Following this motivation, we propose a new family of unbiased estimators called \sas, for matrix production with reduced variance, which only requires storing the sub-sampled activations for calculating the gradient. Our work provides both theoretical and experimental evidence that, in the context of tuning transformers, our proposed estimators exhibit lower variance compared to existing ones. By replacing the linear operation with our approximated one in transformers, we can achieve up to 2. 7X peak memory reduction with almost no accuracy drop and enables up to $6. 4\times$ larger batch size. Under the same hardware, \sas enables better down-streaming task performance by applying larger models and/or faster training speed with larger batch sizes. The code is available at https: //anonymous. 4open. science/r/WTACRS-A5C5/.

PDF Details

ICML Conference 2022 Conference Paper

Accelerating Shapley Explanation via Contributive Cooperator Selection

Guanchu Wang
Yu-Neng Chuang
Mengnan Du
Fan Yang 0023
Quan Zhou
Pushkar Tripathi
Xuanting Cai
Xia Hu 0001

Even though Shapley value provides an effective explanation for a DNN model prediction, the computation relies on the enumeration of all possible input feature coalitions, which leads to the exponentially growing complexity. To address this problem, we propose a novel method SHEAR to significantly accelerate the Shapley explanation for DNN models, where only a few coalitions of input features are involved in the computation. The selection of the feature coalitions follows our proposed Shapley chain rule to minimize the absolute error from the ground-truth Shapley values, such that the computation can be both efficient and accurate. To demonstrate the effectiveness, we comprehensively evaluate SHEAR across multiple metrics including the absolute error from the ground-truth Shapley value, the faithfulness of the explanations, and running speed. The experimental results indicate SHEAR consistently outperforms state-of-the-art baseline methods across different evaluation metrics, which demonstrates its potentials in real-world applications where the computational resource is limited.

Details

NeurIPS Conference 2021 Conference Paper

Fairness via Representation Neutralization

Mengnan Du
Subhabrata Mukherjee
Guanchu Wang
Ruixiang Tang
Ahmed Awadallah
Xia Hu

Existing bias mitigation methods for DNN models primarily work on learning debiased encoders. This process not only requires a lot of instance-level annotations for sensitive attributes, it also does not guarantee that all fairness sensitive information has been removed from the encoder. To address these limitations, we explore the following research question: Can we reduce the discrimination of DNN models by only debiasing the classification head, even with biased representations as inputs? To this end, we propose a new mitigation technique, namely, Representation Neutralization for Fairness (RNF) that achieves fairness by debiasing only the task-specific classification head of DNN models. To this end, we leverage samples with the same ground-truth label but different sensitive attributes, and use their neutralized representations to train the classification head of the DNN model. The key idea of RNF is to discourage the classification head from capturing spurious correlation between fairness sensitive information in encoder representations with specific class labels. To address low-resource settings with no access to sensitive attribute annotations, we leverage a bias-amplified model to generate proxy annotations for sensitive attributes. Experimental results over several benchmark datasets demonstrate our RNF framework to effectively reduce discrimination of DNN models with minimal degradation in task-specific performance.

PDF Details

NeurIPS Conference 2021 Conference Paper

Revisiting Time Series Outlier Detection: Definitions and Benchmarks

Kwei-Herng Lai
Daochen Zha
Junjie Xu
Yue Zhao
Guanchu Wang
Xia Hu

Time series outlier detection has been extensively studied with many advanced algorithms proposed in the past decade. Despite these efforts, very few studies have investigated how we should benchmark the existing algorithms. In particular, using synthetic datasets for evaluation has become a common practice in the literature, and thus it is crucial to have a general synthetic criterion to benchmark algorithms. This is a non-trivial task because the existing synthetic methods are very different in different applications and the outlier definitions are often ambiguous. To bridge this gap, we propose a behavior-driven taxonomy for time series outliers and categorize outliers into point- and pattern-wise outliers with clear context definitions. Following the new taxonomy, we then present a general synthetic criterion and generate 35 synthetic datasets accordingly. We further identify 4 multivariate real-world datasets from different domains and benchmark 9 algorithms on the synthetic and the real-world datasets. Surprisingly, we observe that some classical algorithms could outperform many recent deep learning approaches. The datasets, pre-processing and synthetic scripts, and the algorithm implementations are made publicly available at https: //github. com/datamllab/tods/tree/benchmark

PDF Details

AAAI Conference 2021 System Paper

TODS: An Automated Time Series Outlier Detection System

Kwei-Herng Lai
Daochen Zha
Guanchu Wang
Junjie Xu
Yue Zhao
Devesh Kumar
Yile Chen
Purav Zumkhawaka

We present TODS, an automated Time Series Outlier Detection System for research and industrial applications. TODS is a highly modular system that supports easy pipeline construction. The basic building block of TODS is primitive, which is an implementation of a function with hyperparameters. TODS currently supports 70 primitives, including data processing, time series processing, feature analysis, detection algorithms, and a reinforcement module. Users can freely construct a pipeline using these primitives and perform endto-end outlier detection with the constructed pipeline. TODS provides a Graphical User Interface (GUI), where users can flexibly design a pipeline with drag-and-drop. Moreover, a data-driven searcher is provided to automatically discover the most suitable pipelines given a dataset. TODS is released under Apache 2. 0 license at https: //github. com/datamllab/tods. A video is available on YouTube1.

PDF Details

IJCAI Conference 2020 Conference Paper

Independent Skill Transfer for Deep Reinforcement Learning

Qiangxing Tian
Guanchu Wang
Jinxin Liu
Donglin Wang
Yachen Kang

Recently, diverse primitive skills have been learned by adopting the entropy as intrinsic reward, which further shows that new practical skills can be produced by combining a variety of primitive skills. This is essentially skill transfer, very useful for learning high-level skills but quite challenging due to the low efficiency of transferring primitive skills. In this paper, we propose a novel efficient skill transfer method, where we learn independent skills and only independent components of skills are transferred instead of the whole set of skills. More concretely, independent components of skills are obtained through independent component analysis (ICA), which always have a smaller amount (or lower dimension) compared with their mixtures. With a lower dimension, independent skill transfer (IST) exhibits a higher efficiency on learning a given task. Extensive experiments including three robotic tasks demonstrate the effectiveness and high efficiency of our proposed IST method in comparison to direct primitive-skill transfer and conventional reinforcement learning.

PDF Details DOI