arXiv cs.LG
· Papers
StructAgent: Harness Long-horizon Digital Agents with Unified Causal Structure
arXiv:2607.11388v1 Announce Type: cross Abstract: Recent advances in large language models (LLMs) and vision-language models (VLMs) have enabled increasingly capable digital agents for computer use. However, real-world tasks are often long-horizon and involve evolving contexts containing accumulated observations, inter