BM25
PulseAugur coverage of BM25 — every cluster mentioning BM25 across labs, papers, and developer communities, ranked by signal.
- used by Bangla 90%
- authored by arXiv 70%
- used by DagsHub 70%
- used by Gotit.pub 70%
- used by reciprocal rank fusion 70%
- uses reciprocal rank fusion 70%
- uses LLM 70%
- uses BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation 70%
- used by Beir 70%
- uses Faiss 70%
- used by LongMemEval-S 70%
- uses Hierarchical Navigable Small-World graphs 70%
24 day(s) with sentiment data
-
RAG enables LLM paper Q&A tools to cite sources accurately
Retrieval-augmented generation (RAG) is a technique that allows LLM-based tools to answer questions by first retrieving relevant passages from a large corpus and then generating an answer based on those passages. This a…
-
AI engineers tackle LLM hallucinations with layered controls
An AI engineer's approach to mitigating Large Language Model (LLM) hallucinations involves a multi-layered strategy rather than attempting complete elimination. Key techniques include grounding responses in retrieved do…
-
Dhara retrieval engine enhances Indian Knowledge Systems search
Researchers have developed Dhara, a novel retrieval engine designed to efficiently search Indian Knowledge Systems (IKS). This system integrates hybrid search, entity notes, fuzzy verse retrieval, and neural reranking t…
-
Agent memory optimized for efficiency, moving beyond simple embeddings
This article details the development and measurement of an agent's operational memory system, focusing on efficiency and curated context rather than sheer size. The author describes a system that uses a hybrid approach,…
-
Data lake enables hybrid product search, cutting catalog copies
Researchers have developed a method to enable hybrid product search directly from data lake tables, eliminating the need for separate copies of e-commerce catalogs. This approach leverages Apache Pinot, an OLAP engine, …
-
RAG explained: Why enterprise Q&A needs more than just large language models
Directly feeding internal company documents to large language models for Q&A is ineffective due to issues like hallucination, context window limitations, and lack of accountability. The solution, retrieval-augmented gen…
-
Local code search tool 'ctx' combines BM25 and dense vectors
The article details the inner workings of `ctx`, a TypeScript command-line tool designed for local code search. It explains how `ctx` combines BM25 lexical search with dense vector embeddings, fused using reciprocal ran…
-
AI interview question probes multi-layered approach to reducing LLM hallucinations
An AI engineer's interview question focuses on reducing LLM hallucinations by treating it as a system-level problem. Key strategies include grounding the model with retrieval-augmented generation (RAG), optimizing retri…
-
AI models struggle with scientific claim verification despite advances
Two recent papers explore the challenge of verifying scientific claims against their source material, particularly focusing on multimodal large language models (LLMs). The NTCIR-19 SciClaimEval task involved 22 teams us…
-
New CompSkill method enhances LLM agent skill retrieval using execution experience
Researchers have introduced CompSkill, a novel method for improving skill retrieval in large language model (LLM) agents. Unlike traditional methods that rely solely on skill descriptions, CompSkill incorporates executi…
-
RAG query rewriting: Recall metric is a trap, HyDE fails due to filler
A technical blog post challenges conventional wisdom in retrieval-augmented generation (RAG) pipelines, specifically regarding query rewriting and the metric of recall. The author demonstrates through experiments that s…
-
RAG pipeline safety fails under adversarial testing
A recent evaluation of a retrieval-augmented generation (RAG) pipeline revealed significant vulnerabilities in its safety mechanisms. The system, designed with hybrid retrieval and a relevance threshold, failed to preve…
-
RAG systems focus on retrieval improvements and reliability testing · 8 sources tracked
Researchers are exploring advanced techniques to improve Retrieval-Augmented Generation (RAG) systems, focusing on enhancing the retrieval component which is often the bottleneck. Apple Inc. has developed a method using…
-
New Dhārā Bench benchmark targets retrieval for Indian Knowledge Systems
Researchers have introduced Dhārā Bench, a new retrieval benchmark designed to evaluate information retrieval systems on Indian Knowledge Systems (IKS). The benchmark includes 325 questions across 12 categories and a da…
-
New SearchInject System Mitigates LLM Prompt Injection Attacks
Researchers have developed SearchInject, a novel method to mitigate indirect prompt injection attacks in search-augmented large language models. This system controls the admission of retrieved content by combining score…
-
New Queryable Gaussian View Index System Enhances Scene Asset Retrieval
Researchers have introduced the Queryable Gaussian View Index (QGVI), a novel system designed for querying existing Gaussian scene assets and their associated calibrated reference data. QGVI offers two main routes: QGVI…
-
Qdrant 1.20 integrates BM25 keyword ranking into vector database index
Qdrant has released version 1.20, introducing significant enhancements to its open-source vector database. This update integrates BM25 keyword ranking directly into the full-text index, eliminating the need for separate…
-
AI research papers face rising cross-topic name collisions
A new study has identified a significant increase in cross-topic name collisions within AI research papers, making it harder to identify specific works. The research analyzed over 564,000 AI-related arXiv records and fo…
-
Study highlights growing name collision issue in AI paper retrieval
A new study published on arXiv investigates the growing problem of cross-topic name collisions in scientific paper retrieval, particularly within AI-related research. The paper, titled "PRISM" Is Not All You Need, found…
-
Heading-aware chunking doubles RAG retrieval accuracy
A study on retrieval-augmented generation (RAG) chunking strategies revealed that organizing text by headings significantly improves retrieval accuracy. Heading-aware chunking, which splits documents at section breaks a…