Huan Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Huan Wang has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
Lingcheng Kong
2 shared publicationsYing Li
Westlake University
2 shared publicationsChang Liu
AWS AI Lab, Amazon
1 shared publicationChengfei Lv
Alibaba Group
1 shared publicationDong Xu
School of Computing and Data Science The University of Hong Kong
1 shared publication- Guanzhong Tian
Zhejiang University
1 shared publication
Latest Publications
- 5 October 2026
- Published by Springer Nature
- p. 567-584
EVAR: Edge Visual Autoregressive Models via Principled Pruning
- 5 October 2026
- Published by Springer Nature
- p. 567-584
EVAR: Edge Visual Autoregressive Models via Principled Pruning
- 15 September 2026
- Published by Springer Nature
- p. 362-378
LISA: Locality-Informed Speculative Decoding for Accelerating Autoregressive Image Generation
- 15 September 2026
- Published by Springer Nature
- p. 362-378
LISA: Locality-Informed Speculative Decoding for Accelerating Autoregressive Image Generation
- 12 September 2026
- Published by Springer Nature
- p. 580-598
TARS: MinMax Token-Adaptive Preference Strategy for Hallucination Reduction in MLLMs
- 12 September 2026
- Published by Springer Nature
- p. 580-598
TARS: MinMax Token-Adaptive Preference Strategy for Hallucination Reduction in MLLMs
- Open Access
DICE: Diffusion Large Language Models Excel at Generating CUDA Kernels
Diffusion large language models (dLLMs) have emerged as a compelling alternative to autoregressive (AR) LLMs, owing to their capacity for parallel token generation. This paradigm is particularly well-suited for code generation, where holistic structural planning and non-sequential refinement are critical. Despite this potential, tailoring dLLMs for CUDA kernel generation remains challenging, obstructed not only by the high specialization but also by the severe lack of high-quality training data. To address these challenges, we construct CuKe, an augmented supervised fine-tuning dataset optimized for high-performance CUDA kernels. On top of it, we propose a bi-phase curated reinforcement learning (BiC-RL) framework consisting of a CUDA kernel infilling stage and an end-to-end CUDA kernel generation stage. Leveraging this training framework, we introduce DICE, a series of diffusion large language models designed for CUDA kernel generation, spanning three parameter scales, 1.7B, 4B, and 8B. Extensive experiments on KernelBench demonstrate that DICE significantly outperforms both autoregressive and diffusion LLMs of comparable scale, establishing a new state-of-the-art for CUDA kernel generation.
Open AccessDICE: Diffusion Large Language Models Excel at Generating CUDA Kernels
Diffusion large language models (dLLMs) have emerged as a compelling alternative to autoregressive (AR) LLMs, owing to their capacity for parallel token generation. This paradigm is particularly well-suited for code generation, where holistic structural planning and non-sequential refinement are critical. Despite this potential, tailoring dLLMs for CUDA kernel generation remains challenging, obstructed not only by the high specialization but also by the severe lack of high-quality training data. To address these challenges, we construct CuKe, an augmented supervised fine-tuning dataset optimized for high-performance CUDA kernels. On top of it, we propose a bi-phase curated reinforcement learning (BiC-RL) framework consisting of a CUDA kernel infilling stage and an end-to-end CUDA kernel generation stage. Leveraging this training framework, we introduce DICE, a series of diffusion large language models designed for CUDA kernel generation, spanning three parameter scales, 1.7B, 4B, and 8B. Extensive experiments on KernelBench demonstrate that DICE significantly outperforms both autoregressive and diffusion LLMs of comparable scale, establishing a new state-of-the-art for CUDA kernel generation.
- Open Access
ConCuR: Conciseness Makes State-of-the-Art Kernel Generation
GPU kernel generation by LLMs has recently experienced rapid development, leveraging test-time scaling and reinforcement learning techniques. However, a key challenge for kernel generation is the scarcity of high-quality data, as most high-quality kernels are proprietary and not open-source. This challenge prevents us from leveraging supervised fine-tuning to align LLMs to the kernel generation task. To address this challenge, we develop a pipeline that generates and curates high-quality CUDA kernels with reasoning traces, motivated by a critical observation that concise yet informative reasoning traces result in robust generation of high-performance kernels. Using this pipeline, we construct our dataset ConCuR and introduce our model KernelCoder, which is the first model trained on a curated dataset consisting of PyTorch, reasoning, and CUDA kernel pairs, to our knowledge. In the KernelBench setup, our model achieves significant improvements over the existing top-performing model, QwQ-32B, and outperforms all open-source models fine-tuned for kernel generation, as well as frontier models such as DeepSeek-V3.1-Think and Claude-4-sonnet. Finally, we show that the average reasoning length can serve as a metric to assess the difficulty of kernel generation tasks. The observations, metrics, and our data collection and curation pipeline can help obtain better data in the kernel generation task in the future.
Open AccessConCuR: Conciseness Makes State-of-the-Art Kernel Generation
GPU kernel generation by LLMs has recently experienced rapid development, leveraging test-time scaling and reinforcement learning techniques. However, a key challenge for kernel generation is the scarcity of high-quality data, as most high-quality kernels are proprietary and not open-source. This challenge prevents us from leveraging supervised fine-tuning to align LLMs to the kernel generation task. To address this challenge, we develop a pipeline that generates and curates high-quality CUDA kernels with reasoning traces, motivated by a critical observation that concise yet informative reasoning traces result in robust generation of high-performance kernels. Using this pipeline, we construct our dataset ConCuR and introduce our model KernelCoder, which is the first model trained on a curated dataset consisting of PyTorch, reasoning, and CUDA kernel pairs, to our knowledge. In the KernelBench setup, our model achieves significant improvements over the existing top-performing model, QwQ-32B, and outperforms all open-source models fine-tuned for kernel generation, as well as frontier models such as DeepSeek-V3.1-Think and Claude-4-sonnet. Finally, we show that the average reasoning length can serve as a metric to assess the difficulty of kernel generation tasks. The observations, metrics, and our data collection and curation pipeline can help obtain better data in the kernel generation task in the future.
- Open Access
On-Device Diffusion Transformer Policy for Efficient Robot Manipulation
Diffusion Policies have significantly advanced robotic manipulation tasks via imitation learning, but their application on resource-constrained mobile platforms remains challenging due to computational inefficiency and extensive memory footprint. In this paper, we propose LightDP, a novel framework specifically designed to accelerate Diffusion Policies for real-time deployment on mobile devices. LightDP addresses the computational bottleneck through two core strategies: network compression of the denoising modules and reduction of the required sampling steps. We first conduct an extensive computational analysis on existing Diffusion Policy architectures, identifying the denoising network as the primary contributor to latency. To overcome performance degradation typically associated with conventional pruning methods, we introduce a unified pruning and retraining pipeline, optimizing the model's post-pruning recoverability explicitly. Furthermore, we combine pruning techniques with consistency distillation to effectively reduce sampling steps while maintaining action prediction accuracy. Experimental evaluations on the standard datasets, \ie, PushT, Robomimic, CALVIN, and LIBERO, demonstrate that LightDP achieves real-time action prediction on mobile devices with competitive performance, marking an important step toward practical deployment of diffusion-based policies in resource-limited environments. Extensive real-world experiments also show the proposed LightDP can achieve performance comparable to state-of-the-art Diffusion Policies.
Open AccessOn-Device Diffusion Transformer Policy for Efficient Robot Manipulation
Diffusion Policies have significantly advanced robotic manipulation tasks via imitation learning, but their application on resource-constrained mobile platforms remains challenging due to computational inefficiency and extensive memory footprint. In this paper, we propose LightDP, a novel framework specifically designed to accelerate Diffusion Policies for real-time deployment on mobile devices. LightDP addresses the computational bottleneck through two core strategies: network compression of the denoising modules and reduction of the required sampling steps. We first conduct an extensive computational analysis on existing Diffusion Policy architectures, identifying the denoising network as the primary contributor to latency. To overcome performance degradation typically associated with conventional pruning methods, we introduce a unified pruning and retraining pipeline, optimizing the model's post-pruning recoverability explicitly. Furthermore, we combine pruning techniques with consistency distillation to effectively reduce sampling steps while maintaining action prediction accuracy. Experimental evaluations on the standard datasets, \ie, PushT, Robomimic, CALVIN, and LIBERO, demonstrate that LightDP achieves real-time action prediction on mobile devices with competitive performance, marking an important step toward practical deployment of diffusion-based policies in resource-limited environments. Extensive real-world experiments also show the proposed LightDP can achieve performance comparable to state-of-the-art Diffusion Policies.
- Open Access
SparseSSM: Efficient Selective Structured State Space Models Can Be Pruned in One-Shot
State-space language models such as Mamba match Transformer quality while permitting linear complexity inference, yet still comprise billions of parameters that hinder deployment. Existing one-shot pruning methods are tailored to attention blocks and fail to account for the time-shared and discretized state-transition matrix at the heart of the selective state-space module (SSM). In this paper, we introduce SparseSSM, the first training-free pruning framework that extends the classic optimal brain surgeon (OBS) framework to state space architectures. Our layer-wise algorithm (i) derives an approximate second-order saliency score that aggregates Hessian-trace information across time steps, (ii) incorporates a component sensitivity analysis to guide feed-forward network (FFN) pruning, which also sheds light on where redundancy resides in mamba architecture, (iii) can be easily extended to semi-structured and structured sparsity. Empirically, we prune 50% of SSM weights without fine-tuning and observe no zero-shot accuracy loss, achieving the current state-of-the-art pruning algorithm for Mamba-based LLMs.
Open AccessSparseSSM: Efficient Selective Structured State Space Models Can Be Pruned in One-Shot
State-space language models such as Mamba match Transformer quality while permitting linear complexity inference, yet still comprise billions of parameters that hinder deployment. Existing one-shot pruning methods are tailored to attention blocks and fail to account for the time-shared and discretized state-transition matrix at the heart of the selective state-space module (SSM). In this paper, we introduce SparseSSM, the first training-free pruning framework that extends the classic optimal brain surgeon (OBS) framework to state space architectures. Our layer-wise algorithm (i) derives an approximate second-order saliency score that aggregates Hessian-trace information across time steps, (ii) incorporates a component sensitivity analysis to guide feed-forward network (FFN) pruning, which also sheds light on where redundancy resides in mamba architecture, (iii) can be easily extended to semi-structured and structured sparsity. Empirically, we prune 50% of SSM weights without fine-tuning and observe no zero-shot accuracy loss, achieving the current state-of-the-art pruning algorithm for Mamba-based LLMs.
- 29 November 2024
- Published by Springer Nature
- p. 112-129
A Simple Low-Bit Quantization Framework for Video Snapshot Compressive Imaging
- 29 November 2024
- Published by Springer Nature
- p. 112-129
A Simple Low-Bit Quantization Framework for Video Snapshot Compressive Imaging