12 papers
Combinatorics of degeneracy in vertical systems
Carles Checa, Nidhi Kaihnsa
We characterize the generic consistency of vertical systems in terms of the iterated circuits of the column matroid of the coefficient matrix. More precisely, we prove that such a…
ReproAgent: Contract-Guided Paper-to-Code Reproduction
Xue Hu, Zewei Pan, Zhongyuan Wang +3
Paper-to-code reproduction asks scientific AI agents to turn research papers into executable repositories that preserve the paper's method, protocol and artifacts. This is difficul…
SA-Bench: Evaluating Semantic Alignment in LLM-Based Paper Reproduction
Xue Hu, Zewei Pan, Zeli Su +2
LLM agents can generate paper reproduction code, yet often produce scientifically unfaithful implementations. We define this failure mode as semantic drift, where generated code si…
ExRole: From Team Trajectories to Executable Roles in Multi-Agent Language Models
Zhou Liu, Chaoyang Han, Zewei Pan +2
Roles provide an interpretable interface for organizing language-model agents, yet most multi-agent systems treat them as hand-written prompt labels disconnected from learned behav…
SkillLens: Visual Skill Cards for Retrieval-Augmented GUI Action Prediction and On-Policy Distillation
Zhou Liu, Ligang Huang, Zeli Su +5
Computer-using agents can perceive rich software interfaces, yet their decisions often lack visual procedural memory: they may recognize individual controls without identifying whi…
Source-Grounded Semantic Reinforcement Learning for Low-Resource Target-Language Generation
Zeli Su, Ziyin Zhang, Zewei Pan +8
Low-resource target-language generation is often limited by scarce parallel data, while high-resource source-language monolingual data is abundant but difficult to use with standar…
The Curse of Helpfulness: Inverse Scaling Law in Robustness to Distractor Instructions via DistractionIF
Zeli Su, Zhankai Xu, Tianlei Chen +4
Large Language Models (LLMs) are increasingly deployed in agentic and retrieval-augmented generation (RAG) systems, where they must execute user-specified tasks over externally pro…
FTibSuite: A Comprehensive Resource Suite for Tibetan Vision-Language Modeling
Guixian Xu, Yide Liang, Zeli Su +5
Vision-language models have progressed rapidly, but Tibetan remains a severely underserved low-resource language due to the lack of reproducible training and evaluation infrastruct…
Reinforcement Learning with Semantic Rewards Enables Low-Resource Language Expansion without Alignment Tax
Zeli Su, Ziyin Zhang, Zhou Liu +7
Extending large language models (LLMs) to low-resource languages often incurs an "alignment tax": improvements in the target language come at the cost of catastrophic forgetting in…
SHRP: Specialized Head Routing and Pruning for Efficient Encoder Compression
Zeli Su, Ziyin Zhang, Wenzheng Zhang +3
Transformer encoders are widely deployed in large-scale web services for natural language understanding tasks such as text classification, semantic retrieval, and content ranking.…
CMHG: A Dataset and Benchmark for Headline Generation of Minority Languages in China
Guixian Xu, Zeli Su, Ziyin Zhang +4
Minority languages in China, such as Tibetan, Uyghur, and Traditional Mongolian, face significant challenges due to their unique writing systems, which differ from international st…
Multilingual Encoder Knows more than You Realize: Shared Weights Pretraining for Extremely Low-Resource Languages
Zeli Su, Ziyin Zhang, Guixian Xu +4
While multilingual language models like XLM-R have advanced multilingualism in NLP, they still perform poorly in extremely low-resource languages. This situation is exacerbated by…