Retrieval-Augmented Generation

45 summarised stories about Retrieval-Augmented Generation, each linking back to the original source. Browse all topics →

+ Follow this topic

Key players

The New Stack Amazon Bedrock Amazon Web Services MarkTechPost OpenAI GPT-4.1

Tuesday, 18 August 2026

Improve contract search accuracy with auto-generated filters in Amazon Bedrock

Amazon Web Services 1 month ago 16

Amazon Web Services describes AIDA, an AI-powered system for searching large contract repositories using retrieval-augmented generation with metadata filtering on Amazon Bedrock Knowledge Bases. The system uses implicit metadata pre-filtering before semantic search, followed by explicit application-layer constraints, to narrow results and reduce retrieval noise in legal documents. Enterprises can now query complex contracts in natural language with improved accuracy, though AWS emphasizes that AI-generated interpretations should still be reviewed by legal professionals.

Read the full story

Enterprise AI Retrieval-Augmented Generation

Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers

Hugging Face 1 month ago 42

Sentence Transformers v6.0 introduces MultiVectorEncoder, a new model type for late-interaction retrieval that keeps one vector per token instead of compressing text into a single vector, enabling stronger retrieval at the cost of larger indexes. The MaxSim operator scores queries by finding each query token's best match across all document tokens and summing those similarities, preserving both semantic understanding and exact-match capability without the lossy compression of dense embeddings. This approach trades increased storage requirements—roughly 42x more vectors per document—for improved retrieval quality on multi-requirement queries, rare entities, and out-of-domain data, with compressed indexes remaining comparable to dense embedding storage in practice.

Read the full story

Retrieval-Augmented Generation Embedding Models

Every AI story that matters,
in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.