activity
20242026
collaborators

8 papers

cs.CV2026

Persistent Visual Memory: Sustaining Perception for Deep Generation in LVLMs

Siyuan Huang, Xiaoye Qu, Yafu Li +6

While autoregressive Large Vision-Language Models (LVLMs) demonstrate remarkable proficiency in multimodal tasks, they face a "Visual Signal Dilution" phenomenon, where the accumul…

cs.CV2026

GEMS: Agent-Native Multimodal Generation with Memory and Skills

Zefeng He, Siyuan Huang, Xiaoye Qu +4

Recent multimodal generation models have achieved remarkable progress on general-purpose generation tasks, yet continue to struggle with complex instructions and specialized downst…

cs.CL2026

Pretraining with Token-Level Adaptive Latent Chain-of-Thought

Boyi Zeng, Yiqin Hao, He Li +8

Scaling large language models by increasing parameters and training data is increasingly constrained by limited high-quality corpora and rising communication costs. This work explo…

cs.CL2026

PonderLM: Pretraining Language Models to Ponder in Continuous Space

Boyi Zeng, Shixiang Song, Siyuan Huang +6

Humans ponder before articulating complex sentence elements, enabling deeper cognitive processing through focused effort. In this work, we introduce this pondering process into lan…

cs.CL2025

Training LLMs to be Better Text Embedders through Bidirectional Reconstruction

Chang Su, Dengliang Shi, Siyuan Huang +5

Large language models (LLMs) have increasingly been explored as powerful text embedders. Existing LLM-based text embedding approaches often leverage the embedding of the final toke…

cs.CL2025

Mirror-Consistency: Harnessing Inconsistency in Majority Voting

Siyuan Huang, Zhiyuan Ma, Jintao Du +3

Self-Consistency, a widely-used decoding strategy, significantly boosts the reasoning capabilities of Large Language Models (LLMs). However, it depends on the plurality voting rule…