13 papers
The Shape of Time: Video-Token Contrast for Temporal Understanding in VideoLMs
Yumeng Shi, Quanyu Long, Yin Wu +1
Seeing frames in order does not mean representing time. Modern VideoLMs receive ordered video streams, yet their main supervision acts on generated text rather than video-token rep…
Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond
Meng Chu, Xuan Billy Zhang, Kevin Qinghong Lin +47
As AI systems move from generating text to accomplishing goals through sustained interaction, the ability to model environment dynamics becomes a central bottleneck. Agents that ma…
Self-Verification Dilemma: Experience-Driven Suppression of Overused Checking in LLM Reasoning
Quanyu Long, Kai Jie Jiang, Jianda Chen +3
Large Reasoning Models (LRMs) achieve strong performance by generating long reasoning traces with reflection. Through a large-scale empirical analysis, we find that a substantial f…
MemSkill: Learning and Evolving Memory Skills for Self-Evolving Agents
Haozhen Zhang, Quanyu Long, Jianzhu Bao +4
Most Large Language Model (LLM) agent memory systems rely on a small set of static, hand-designed operations for extracting memory. These fixed procedures hard-code human priors ab…
From Competition to Synergy: Unlocking Reinforcement Learning for Subject-Driven Image Generation
Ziwei Huang, Ying Shu, Hao Fang +5
Subject-driven image generation models face a fundamental trade-off between identity preservation (fidelity) and prompt adherence (editability). While online reinforcement learning…
Causality Matters: How Temporal Information Emerges in Video Language Models
Yumeng Shi, Quanyu Long, Yin Wu +1
Video language models (VideoLMs) have made significant progress in multimodal understanding. However, temporal understanding, which involves identifying event order, duration, and…
LLMC+: Benchmarking Vision-Language Model Compression with a Plug-and-play Toolkit
Chengtao Lv, Bilang Zhang, Yang Yong +7
Large Vision-Language Models (VLMs) exhibit impressive multi-modal capabilities but suffer from prohibitive computational and memory demands, due to their long visual token sequenc…
Coordinating Search-Informed Reasoning and Reasoning-Guided Search in Claim Verification
Qisheng Hu, Quanyu Long, Wenya Wang
Multi-hop claim verification is inherently challenging, requiring multi-step reasoning to construct verification chains while iteratively searching for information to uncover hidde…
Reinforcing Compositional Retrieval: Retrieving Step-by-Step for Composing Informative Contexts
Quanyu Long, Jianda Chen, Zhengyuan Liu +3
Large Language Models (LLMs) have demonstrated remarkable capabilities across numerous tasks, yet they often rely on external context to handle complex tasks. While retrieval-augme…
Static or Dynamic: Towards Query-Adaptive Token Selection for Video Question Answering
Yumeng Shi, Quanyu Long, Wenya Wang
Video question answering benefits from the rich information in videos, enabling various applications. However, the large volume of tokens generated from long videos presents challe…
BOOST: Bootstrapping Strategy-Driven Reasoning Programs for Program-Guided Fact-Checking
Qisheng Hu, Quanyu Long, Wenya Wang
Large language model pipelines have improved automated fact-checking for complex claims, yet many approaches rely on few-shot in-context learning with demonstrations that require s…
Visual-RAG: Benchmarking Text-to-Image Retrieval Augmented Generation for Visual Knowledge Intensive Queries
Yin Wu, Quanyu Long, Jing Li +2
Retrieval-augmented generation (RAG) is a paradigm that augments large language models (LLMs) with external knowledge to tackle knowledge-intensive question answering. While severa…
Decomposition Dilemmas: Does Claim Decomposition Boost or Burden Fact-Checking Performance?
Qisheng Hu, Quanyu Long, Wenya Wang
Fact-checking pipelines increasingly adopt the Decompose-Then-Verify paradigm, where texts are broken down into smaller claims for individual verification and subsequently combined…