12 papers
Correcting What You Cannot See: Credit Assignment for Perception Distillation in Multimodal Reasoners
Feng Xiong, Leyan Xue, Hongyu Lin
On-policy distillation provides dense supervision for multimodal reasoners, but its trajectory-level reward cannot determine whether a failed answer arose from perception or subseq…
Distill What the Student Can See: Fisher-Projected On-Policy Distillation for Vision-Language Models
Leyan Xue, Feng Xiong, Mingjun Ma +1
On-policy distillation (OPD) samples trajectories from the current student policy and minimizes token-level divergence between student and teacher next-token distributions at prefi…
Visually-Guided Policy Optimization for Multimodal Reasoning
Zengbin Wang, Feng Xiong, Liang Lin +5
Reinforcement learning with verifiable rewards (RLVR) has significantly advanced the reasoning ability of vision-language models (VLMs). However, the inherent text-dominated nature…
Ace-Skill: Bootstrapping Multimodal Agents with Prioritized and Clustered Evolution
Feng Xiong, Zengbin Wang, Yong Wang +5
Self-evolving agents present a promising path toward continual adaptation by distilling task interactions into reusable knowledge artifacts. In practice, this paradigm remains hind…
MMKG-RDS: Reasoning Data Synthesis via Deep Mining of Multimodal Knowledge Graphs
Lun Zhan, Feng Xiong, Huanyong Liu +2
Synthesizing high-quality training data is crucial for enhancing domain models' reasoning abilities. Existing methods face limitations in long-tail knowledge coverage, effectivenes…
Active Zero: Self-Evolving Vision-Language Models through Active Environment Exploration
Jinghan He, Junfeng Fang, Feng Xiong +5
Self-play has enabled large language models to autonomously improve through self-generated challenges. However, existing self-play methods for vision-language models rely on passiv…
AR-MAP: Are Autoregressive Large Language Models Implicit Teachers for Diffusion Large Language Models?
Liang Lin, Feng Xiong, Zengbin Wang +5
Diffusion Large Language Models (DLLMs) have emerged as a powerful alternative to autoregressive models, enabling parallel token generation across multiple positions. However, pref…
Everything in Its Place: Benchmarking Spatial Intelligence of Text-to-Image Models
Zengbin Wang, Xuecai Hu, Yong Wang +3
Text-to-image (T2I) models have achieved remarkable success in generating high-fidelity images, but they often fail in handling complex spatial relationships, e.g., spatial percept…
AdaCuRL: Adaptive Curriculum Reinforcement Learning with Invalid Sample Mitigation and Historical Revisiting
Renda Li, Hailang Huang, Fei Wei +3
Reinforcement learning (RL) has demonstrated considerable potential for enhancing reasoning in large language models (LLMs). However, existing methods suffer from Gradient Starvati…
Position Bias Mitigates Position Bias:Mitigate Position Bias Through Inter-Position Knowledge Distillation
Yifei Wang, Feng Xiong, Yong Wang +3
Positional bias (PB), manifesting as non-uniform sensitivity across different contextual locations, significantly impairs long-context comprehension and processing capabilities. Pr…
HS-STaR: Hierarchical Sampling for Self-Taught Reasoners via Difficulty Estimation and Budget Reallocation
Feng Xiong, Hongling Xu, Yifei Wang +3
Self-taught reasoners (STaRs) enhance the mathematical reasoning abilities of large language models (LLMs) by leveraging self-generated responses for self-training. Recent studies…
Whoever Started the Interference Should End It: Guiding Data-Free Model Merging via Task Vectors
Runxi Cheng, Feng Xiong, Yongxian Wei +2
Model merging seeks to integrate task-specific expert models into a unified architecture while preserving multi-task generalization capabilities, yet parameter interference between…