Fetching from the wire…
01
02
03
04
05
06
07
08
Research2026-05-11 · source-backed
A new system gives simple queries a fast single-pass while triggering exploration and verification for complex joins. Achieves new SOTA on Spider and BIRD text-to-SQL benchmarks. Addresses a real production problem: uniform compute wastes money on easy queries and underperforms on hard ones.
Each link below shares sources, entities, or timing with this story.
arXiv 2608.05906 keeps a dual-polarity memory of verified corrections and observed dead ends for Text-to-SQL repair: 66.34% to 69.79% on Spider, 47.35% to 48.44% on BIRD. Then the authors say the quiet part: paired analysis supports the Spider gain but is weak on BIRD, MERIT i...
Rubric-Supervised Critic (2603.03800) — Critic model trained on 24 behavioral features achieves +15.9 improvement on SWE-bench via reranking. 83% fewer task attempts with early stopping. Optimal Transport Refusal Ablation (2603.04355) — Achieves up to 11% higher attack success...
Formalizes the "same query, different results" problem via MDPs. Identifies three variance sources: information acquisition, compression, and inference. Achieves 22% stochasticity reduction while maintaining quality. Structured output formatting and ensemble queries are the mo...
PCAS: Policy Compiler for Secure Agentic Systems — The first paper to provide measured enforcement results for agent policy compliance (48% to 93%). Uses dependency graphs and Datalog-derived policy language with a reference monitor intercepting all actions. Three case studies...
arXiv 2607.12227 (Wang et al., incl. Hajishirzi, Tsvetkov, Dasigi) finds two methodological holes in the self-improving-agent literature: methods are never compared against simpler baselines at matched compute budgets, and final performance gets reported on the same public ben...
A July 2 evaluation tested semantic chunking against simple approaches on long structured academic theses using RAGAs, and the sophisticated method didn't win. Performance varied more with document formatting, preprocessing, and query type than with chunking strategy. The auth...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.