10 papers
Test-time Sparsity for Extreme Fast Action Diffusion
Kangye Ji, Yuan Meng, Jianbo Zhou +3
Action diffusion excels at high-fidelity action generation but incurs heavy computational costs owing to its iterative denoising nature. Despite current technologies showing promis…
Permit: Permission-Aware Representation Intervention for Controlled Generation in Large Language Models
Pengcheng Sun, Lan Zhang, Zhaopeng Zhang +2
Large language models (LLMs) are increasingly deployed in enterprise settings where they handle sensitive documents and user context, raising acute concerns over security and contr…
TS-DP: Reinforcement Speculative Decoding For Temporal Adaptive Diffusion Policy Acceleration
Ye Li, Jiahe Feng, Yuan Meng +6
Diffusion Policy (DP) excels in embodied control but suffers from high inference latency and computational cost due to multiple iterative denoising steps. The temporal complexity o…
SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Acceleration
Ye Li, Yuan Meng, Zewen Sun +7
Vision-Language-Action (VLA) models have attracted increasing attention for their strong control capabilities. However, their high computational cost and low execution frequency hi…
Quantization Meets OOD: Generalizable Quantization-aware Training from a Flatness Perspective
Jiacheng Jiang, Yuan Meng, Chen Tang +4
Current quantization-aware training (QAT) methods primarily focus on enhancing the performance of quantized models on in-distribution (I.D) data, while overlooking the potential pe…
JAQ: Joint Efficient Architecture Design and Low-Bit Quantization with Hardware-Software Co-Exploration
Mingzi Wang, Yuan Meng, Chen Tang +9
The co-design of neural network architectures, quantization precisions, and hardware accelerators offers a promising approach to achieving an optimal balance between performance an…