lab Google — The Keyword (AI) · 1d ago · releases

UN launches System Data Commons to unify global statistics

from English

The UN system is launching the UN System Data Commons, an open-source platform built on Google's Data Commons that unites global statistics into a single interconnected resource known as an AI-ready knowledge graph. Supported by Google.org, the project aims to make critical data universally accessible and help track global progress in real time.

The platform automatically integrates metrics, timelines, and geographic boundaries from siloed datasets into one environment.
Users can explore data through natural language search or filter by location and themes like health and education.
AI assistant capabilities allow agents to autonomously fetch authoritative figures using the Model Context Protocol (MCP).
The UN system aims to include 80% of its statistical datasets in the platform by 2027.

The initiative helps analysts spend less time formatting spreadsheets and more time uncovering key trends, while allowing anyone from researchers to journalists to access trusted, official facts.

Importance 1/3 Trust 3/3 Google — The Keyword (AI) Google DeepMind AI agents Retrieval & RAG

Read original

arxiv arXiv cs.CL · 94d ago · 6 views

ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

ProvenanceGuard introduces a source-aware verifier for MCP-based LLM agents that detects cross-source conflation by routing claims to specific evidence sources and comparing stated attribution with actual source ownership. It achieves block F1 of 0.802 and source accuracy of 0.858 on 260 source-eligible claims, outperforming source-blind baselines, and detects all injected attribution swaps in 50 clinical probes.

arxiv arXiv cs.AI · 94d ago · 4 views

ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

arxiv arXiv cs.CL · 94d ago · 4 views

MODE-RAG: Evaluating and Reducing Hallucinations in M-RAG

MODE-RAG proposes a multi-agent system using Variational Free Energy to dynamically gate interventions and reduce cross-modal hallucinations in retrieval-augmented generation. It integrates Monte Carlo Tree Search and logit perturbations to address causal fabrications and sycophancy, with dedicated agents ensuring factual verification and formatting stability. Evaluated via ModeVent, a subset of MultiVent, the system significantly improves robustness against logical fabrications.

media Hugging Face Forums · 3h ago

jbsalles shares SelMem, an experimental Rust project for LLM selective memory

The author introduces SelMem, an early-stage experimental project written in Rust that explores whether LLM agents can develop persistent behavioral divergence through a reconstructive memory system. This system allows agents to forget, distort, and consolidate experiences over time, potentially leading to different internal narratives even with identical histories.

media Hugging Face Forums · 3d ago · 7 views

Prototype separates agent resource discovery from delivery to keep costs low

An early, AI-assisted prototype for agents to discover and use static Markdown methods and templates has been shared for design feedback. The system decouples discovery from delivery by returning short search candidates rather than full documents.