14 papers
Is Oracle Pruning the True Oracle?
Sicheng Feng, Keda Tao, Huan Wang
Oracle pruning, which selects unimportant weights by minimizing the pruned train loss, has served as the foundation for most neural network pruning methods for over thirty-five yea…
Which Heads Matter for Reasoning? RL-Guided KV Cache Compression
Wenjie Du, Li Jiang, Keda Tao +2
Reasoning large language models exhibit complex reasoning behaviors via extended chain-of-thought generation that are highly fragile to information loss during decoding, creating c…
OmniZip: Audio-Guided Dynamic Token Compression for Fast Omnimodal Large Language Models
Keda Tao, Kele Shao, Bohan Yu +3
Omnimodal large language models (OmniLLMs) have attracted increasing research attention of late towards unified audio-video understanding. However, the high computational cost of p…
TARS: MinMax Token-Adaptive Preference Strategy for Hallucination Reduction in MLLMs
Kejia Zhang, Keda Tao, Zhiming Luo +3
Multimodal large language models (MLLMs) are prone to hallucinations, generating plausible but visually ungrounded outputs, partly because direct preference optimization (DPO) over…
LVOmniBench: Pioneering Long Audio-Video Understanding Evaluation for Omnimodal LLMs
Keda Tao, Yuhua Zheng, Jia Xu +13
Recent advancements in omnimodal large language models (OmniLLMs) have significantly improved the comprehension of audio and video inputs. However, current evaluations primarily fo…
StreamingTOM: Streaming Token Compression for Efficient Video Understanding
Xueyi Chen, Keda Tao, Kele Shao +1
Unlike offline processing, streaming video vision-language models face two fundamental constraints: causality and accumulation. Causality prevents access to future frames that offl…