15 papers
Atomic Task Graph: A Unified Framework for Agentic Planning and Execution
Yue Zhang, Sihan Chen, Ziwen Huang +3
LLM-based agents have shown strong potential for solving complex multi-step tasks, yet existing performance improvements often rely on either scaling to larger backbone models or t…
Sparse ActionGen: Accelerating Diffusion Policy with Real-time Pruning
Kangye Ji, Jianbo Zhou, Yuan Meng +3
Diffusion Policy has dominated action generation due to its strong capabilities for modeling multi-modal action distributions, but its multi-step denoising processes make it imprac…
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…
Block-wise Adaptive Caching for Accelerating Diffusion Policy
Kangye Ji, Yuan Meng, Hanyun Cui +5
Diffusion Policy has demonstrated strong visuomotor modeling capabilities, but its high computational cost renders it impractical for real-time robotic control. Despite huge redund…
VP-VAE: Rethinking Vector Quantization via Adaptive Vector Perturbation
Linwei Zhai, Han Ding, Mingzhi Lin +5
Vector Quantized Variational Autoencoders (VQ-VAEs) are fundamental to modern generative modeling, yet they often suffer from training instability and "codebook collapse" due to th…
WiSparse: Boosting LLM Inference Efficiency with Weight-Aware Mixed Activation Sparsity
Lei Chen, Yuan Meng, Xiaoyu Zhan +2
Large Language Models (LLMs) offer strong capabilities but incur high inference costs due to dense computation and memory access. Training-free activation sparsity is a promising a…