CMU and Microsoft's AdaPRA uses adaptive rubrics to cut data science agent errors
CarnegieMellonU · hf · 2026-10-29
AdaPRA, from CMU and Microsoft, targets process errors in data science agents: insufficient exploration, unsound validation, coding bugs, and poor budget use (re-running the same EDA and baselines every attempt). It introduces a persistent coding agent harness that accumulates knowledge across attempts on the same task, plus a process reward agent guided by adaptive rubrics at inference time to check things like submission integrity — reducing errors and improving leaderboard performance.
More from coding & agent
- DeepMind opens Antigravity Science Program, pairing Gemini agents with AlphaFold for research — DeryaTR_ · 2026-10-29
- Refactoring one SKILL.md into a Kanban agent graph: same output, 3x tokens, 4x time — granawkins · 2026-10-29
- 10-person Spawn spends 30x payroll on AI tokens, ships 77,000 changes a month — mattshumer_ · 2026-10-29
- Grok Bot locked out by Amazon simply bought the same item elsewhere, cheaper — jamesdouma · 2026-10-29
- Dev goes vibe coding C++: JUCE Acid-303 plugin parameter code as proof — kylegawley · 2026-10-29
- Open-source Jev undercuts Clay: $0.038 per verified lead vs $167/mo plans — HeyAmit_ · 2026-10-29