Tianyi Zhang
3 papers · Latest:
Computer Vision
Geometry-Aware State Space Model: A New Paradigm for Whole-Slide Image Representation
BatMIL introduces a geometry-aware state space model with hybrid hyperbolic-Euclidean representations for improved whole-slide image analysis.
2605.05164
Human-Computer InteractionDo LLMs Need to See Everything? A Benchmark and Study of Failures in LLM-driven Smartphone Automation using Screentext vs. Screenshots
This paper introduces DailyDroid, a benchmark for LLM-driven smartphone automation, comparing text-only vs. multimodal inputs and analyzing common failure modes.
2604.17817
Natural Language ProcessingDemystifying OPD: Length Inflation and Stabilization Strategies for Large Language Models
This paper identifies and solves length inflation in on-policy distillation (OPD) for LLMs, improving training stability and performance.
2604.08527
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.