Find a story

Search Spins

Search titles, summaries, and missing voices across published articles — press releases, announcements, and media coverage.

0 results for “external knowledge”

SPIN Processed News Frame: The Hype

SelfGraphRAG: Bridging the Supervision Gap in Graph-Based RAG with Synthetic QA Generation

SelfGraphRAG is a new research framework that auto-generates synthetic question-answer pairs from knowledge graph structure to train graph-based RAG retrievers without human-labeled data, improving multi-hop QA and classification performance over embedding baselines.

Spin 45% Claim Present in Source AI Risk Moderate
arXiv Computation and Language

Aug 27, 2026

SPIN Processed News Frame: The Hype

Cross-Model Memory Transfer via Target-Side Reader Adaptation

A new research paper proposes 'Engram-style hashed memory' as a reusable external knowledge artifact for large language models, enabling cross-model memory transfer via lightweight target-side reader adaptation rather than full model retraining.

Spin 45% Claim Present in Source AI Risk Moderate
arXiv Computation and Language

Aug 19, 2026

SPIN Processed News Frame: The Halo

Leveraging External Knowledge for Historical Document Restoration via Retrieval-Augmented Large Language Models

Researchers introduced ARI, a retrieval-augmented LLM framework for restoring illegible historical documents—especially named entities—by combining pretrained model knowledge with retrieved external historical context, validated on Korean archival texts.

Spin 55% Claim Present in Source AI Risk Moderate
arXiv Computation and Language

Jul 27, 2026

SPIN Processed News Frame: The Hype

COALA: Robust Contextualized Speech-augmented Language Modeling for ASR via Contrastive Regularizer and Biasing Score Estimation

COALA is a new research framework for improving automatic speech recognition in multi-entity scenarios by introducing contrastive regularization and biasing score estimation to better match audio segments with domain-specific entities.

Spin 45% Claim Present in Source AI Risk Moderate
arXiv Computation and Language

Jul 10, 2026