Papers by Xuewen Zhong

2 papers
CNSL-bench: Benchmarking the Sign Language Understanding Capabilities of MLLMs on Chinese National Sign Language (2026.acl-long)

Copied to clipboard

Challenge: CNSL-bench is the first comprehensive Chinese National Sign Language benchmark . current MLLMs are inferior to human performance, despite advances in multimodal modeling .
Approach: They propose a Chinese National Sign Language benchmark to evaluate multimodal large language models in sign language understanding.
Outcome: The proposed benchmark evaluates 21 open-source and proprietary MLLMs . results show that current models are inferior to human performance .
Selective Contrastive Learning For Gloss Free Sign Language Translation (2026.acl-long)

Copied to clipboard

Challenge: Recent SLT systems adopt CLIP-like Vision-Language pretraining, but the random in-batch contrast provides few, batch-dependent negatives.
Approach: They propose a method to train sign video-text similarity over a time period of 3 months . they use a random in-batch contrast strategy to track negative video- text similarity .
Outcome: The proposed system improves sign language translation by focusing on challenging negatives . the results show that the random in-batch contrast provides few negatives and noisy supervision .

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations