Tag
Presents Guixu, a valuation-driven data discovery system for autonomous AI agents that uses proxy-label propagation and knapsack optimization for task-aware data valuation, with on-chain attestation for verifiable procurement.
Google Data Cloud's frontier AI team discusses a new approach to evaluating AI agents using information theory to create a meta-benchmark called Discovery Bench that measures how vague a query can be before an agent fails, providing a more nuanced map of agent capabilities than simple pass/fail exams.
CODA-BENCH is a new benchmark for evaluating code agents on data-intensive tasks, bridging the gap between code-centric and data-centric evaluations. It includes over 1,000 tasks from 31 communities, with realistic data scale and noise, revealing that even top agents achieve only 61.1% success rate.
OpenMetadata is a fast-growing open-source unified metadata platform offering data discovery, observability, and governance with 84+ connectors and no-code data quality tools.