Yue Zhang · page 3
Papers
- DART: Leveraging Multi-Agent Disagreement for Tool Recruitment in Multimodal Reasoning (2026 · 1 citation)
Debates between vision-language models (VLMs) often stall because the agents share similar perceptual blind spots and bring no new evidence. - Unveiling the Deficiencies of Pre-trained Text-and-Layout Models in Real-world Visually-rich Document Information Extraction (2026 · 1 citation)
Pre-trained text-and-layout models (PTLMs) score very highly on standard benchmarks for extracting information from visually-rich documents (VrD-IE), but they disappoint in practice. - Breaking Down and Building Up: Mixture of Skill-Based Vision-and-Language Navigation Agents (2026)
Vision-and-Language Navigation (VLN) agents still generalize poorly to unseen buildings and unfamiliar instruction phrasing. - Modular Neural Machine Translation with a Semantic Pivot - Pilot Study Using AMR (2026)
The paper builds a translation pipeline with a semantic pivot, in the spirit of the translation pyramid. Each language gets two modules: one that reads a sentence into English AMR, and one that writes the target sentence from AMR. The training data is a silver AMR corpus made by a strong parser from the United… - Routing with Generated Data: Annotation-Free LLM Skill Estimation and Expert Selection (2026)
LLM routers choose which model in a pool should handle a given query. Most existing routers learn from human-labelled, in-domain data, which is often unavailable when real user traffic is varied and unpredictable. The authors propose Routing with Generated Data (RGD). In this setting, a generator LLM writes queries…
This page in Markdown → https://md.research.elamur.ai/free/en/authors/yue-zhang/page/3/