Compaction beats full transcript in memory
Compaction scores 71.4 percent vs 64.6 percent with 34 percent fewer tokens on LoCoMo.
Stories per day · 30 days
Relatedknowledge-graph2channel-estimation1competitive-programming1conversational-ai1csi-prediction1data-integration1dermatology1edge-ai1embedding-vectors1hallucination1knowledge-poisoning1llm-evaluation1medical-ai1memory-compaction1multi-hop-reasoning1multimodal1offline-inference1offline-rag1
17 stories, newest first · one card per story
Stories tagged #retrieval-augmented-generation
Retrieval-Augmented Generation Improves LLM Performance on Competitive Programming
Embedding-based retrieval with similarity search enhances LLMs in solving algorithmic problems from CodeContests dataset.
SkinMind uses graph probing for dermatology retrieval-augmented generation
Authors report average gains of 8.70 and 8.89 points with GPT-5.4-mini and Gemini-2.5-Flash on six VQA benchmarks.
KG-RAG4SM retrieves knowledge-graph subgraphs for schema matching
Authors report precision gains of 39.79% on MIMIC and 69.20% on Synthea versus named prior methods, without retraining.
arXiv measures text-vector integrity gap in offline RAG
Verifier recomputes vectors within cosine tolerance to close the integrity gap, exposing up to 5.5% of queries to attackers who control vectors.
Researchers identify text-vector integrity gap in offline RAG
Offline RAG knowledge packs can be hijacked with up to 5.5% of queries exposed due to text-vector mismatch.
RAG-CSI retrieves channel history to improve CSI prediction
On a multi-scene dataset, the method cut normalized mean squared error 15.42% and improved all five metrics by more than 10%.
Signed knowledge packs miss poisoned offline RAG updates on Pi 5
A validly signed poisoned pack zeroed 30 previously correct answers, and 14 returned the injected answer; hash checks missed six of nine attacks.
SARL trains search models to track missing retrieved information
Authors report a 24.3% relative gain over the strongest baseline, with no extra assessment at test time.
A Security Meta-Model for Retrieval-Augmented Generation Systems
A RAG security meta-model catalogs attacks, weaknesses, risks, and CIA impact drawn from 43 papers.
World Embedding Benchmark
World Embedding Benchmark offers 8,000 simulation cases with physical annotations to test how video embeddings encode physics.
World Embedding Benchmark
World Embedding Benchmark pairs 8,000 physics-annotated simulated videos with three eval tasks, finding omnimodal embeddings weak at physical alignment.
What Would RAG Look Like If Miranda Priestly Were the User?
Akamai blog asks how retrieval-augmented generation would serve a demanding Miranda Priestly-style user.
InceptionRAG: Stealthy Poisoning Attack Against Retrieval-Augmented Generation
InceptionRAG fragments malicious payloads into dormant passages that trigger LLMs to self-deduce misinformation via multi-hop reasoning, bypassing existing RAG poisoning defenses.
CiteShade: Citation Laundering in Multi-Source Retrieval-Augmented Generation and Its Counterfactual Defense
CiteShade attack makes RAG models cite trusted sources for attacker-chosen wrong answers, raising wrong-answer rate from 0.01 to 0.68.
ReCite: Agentic Reasoning for Faithful Citation
ReCite is an agentic citation framework using claim-level reasoning and verification, outperforming large generative models in strict citation accuracy.
ReMoMask-2: Latent Retrieval-Augmented Masked Motion Generation
ReMoMask-2 rebuilds retrieval in the generator's latent space for text-to-motion generation, achieving lowest FID on KIT-ML and SnapMoGen.