Reproduce “KDASO: Knowledge-Data Dual-Driven Automated Skill Optimization for Liability Adjudication Tasks” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zi-Jian Cheng, Weiming Wu, Jie Meng, Zhen Peng, Shan Huang, Qun Li, Guobing Wu, Lan-Zhe Guo
KDASO adapts liability adjudication by updating the human-readable skills an LLM uses, without fine-tuning the model. It induces rules from recent cases, refines them against expert reasoning, and routes relevant skills within a context budget. On the Cancel–Appeal ride-hailing task, QwQ-32B with KDASO reaches 73.42% ordinary-liability precision and 87.38% recall, but this does not validate GPS-spoofing scenarios.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.