Beir
PulseAugur coverage of Beir — every cluster mentioning Beir across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
New research reframes LLM prefix sharing as a sorting problem
A new paper proposes that optimizing the order of retrieved text pieces in retrieval-augmented generation (RAG) systems can significantly improve Large Language Model (LLM) serving efficiency. The research demonstrates …
-
New QueryRoute benchmark evaluates LLM query reformulation strategies
Researchers have introduced QueryRoute, a new benchmark designed to evaluate query reformulation selection strategies for LLM-based information retrieval. This benchmark addresses the challenge of choosing the optimal q…
-
EviQE method improves LLM query expansion by selecting relevant documents
Researchers have developed EviQE, a novel method for improving Large Language Model (LLM)-based query expansion by focusing on selecting relevant documents for the model to process. This approach separates the evidence …
-
MoE models get smarter pruning, retrieval, and inference efficiency
Researchers are exploring advanced techniques for Mixture-of-Experts (MoE) language models to improve their efficiency and performance. One paper introduces HOPE (Higher-Order Pruning of Experts), a novel pruning object…
-
New Matryoshka Hash Representations improve RAG retrieval efficiency
Researchers have developed Matryoshka Hash Representations (MHR), a novel method for compact semantic retrieval in retrieval-augmented generation (RAG) systems. MHR addresses the challenge of storing large vector indexe…
-
DoPR framework boosts LLM reranking efficiency with compressed document prefixes
Researchers have developed DoPR, a novel framework designed to enhance the efficiency of Large Language Model (LLM) reranking. DoPR addresses the issue of redundant document processing by decoupling offline document pre…
-
MoganColBERT-TR: New Turkish Multi-Vector Retrieval Model Unveiled
Researchers have developed MoganColBERT-TR, a new multi-vector retrieval model specifically designed for the Turkish language. This model builds upon a previously trained ModernBERT encoder and adapts it to the ColBERT …
-
E2Rank unifies text embedding and reranking for efficient search
Researchers have developed E2Rank, a novel framework that unifies text embedding and listwise reranking for more effective and efficient search. This approach extends a single text embedding model to perform both retrie…
-
New arXiv papers explore privacy, efficiency, and LLM integration in dense retrieval
Four new arXiv papers explore advancements in dense retrieval, a key component for large language models in information retrieval tasks. The first paper introduces a privacy-preserving method using learned deep hashing …
-
New method adapts dense retrieval models using agent search traces
Researchers have developed Navigation-Informed Embeddings (NIE), a novel method to adapt dense retrieval models using traces from agentic search workflows. This approach leverages query, retrieval, and stopping traces t…
-
New research reveals fundamental limits in RAG defenses against coordinated poisoning attacks
Researchers have demonstrated a fundamental limitation in current defenses against coordinated poisoning attacks on vector retrieval systems used in retrieval-augmented generation (RAG). These admission-time defenses, w…
-
New DESA method enhances LLM-based retrieval with channel-asymmetric expansion
Researchers have developed DESA (Dense Expansion and Sparse Anchoring), a novel method for improving information retrieval using LLM-based query expansion. Unlike previous approaches that fuse fixed top-L dense and spar…
-
New research explores static pruning and LLM-based query expansion for retrieval systems
Two new research papers explore methods to improve information retrieval systems. The first paper, "Static Pruning Across Sparse Retrieval Regimes," investigates how static pruning techniques can be applied across diffe…
-
New models unify sparse and dense multimodal embeddings, boosting search efficiency
Researchers have introduced UEmbed, a novel decoder-only multimodal embedding model capable of generating both sparse lexical and dense representations within a single causal forward pass. This model aims to unify spars…
-
Hyperbolic geometry retrieval system enables RAG on edge devices
Researchers have developed a novel hybrid retrieval system that leverages hyperbolic geometry for retrieval-augmented generation (RAG) on edge devices. This system projects word embeddings into hyperbolic space, allowin…
-
New CE-QE method enhances lexical retrieval by grounding in semantic evidence
Researchers have developed a new method called Cross-Encoder Query Expansion (CE-QE) to improve information retrieval systems. This technique addresses the limitation of traditional lexical retrieval methods like BM25, …
-
New RAG defense frameworks combat data poisoning attacks · 3 sources tracked
Researchers have developed new defense frameworks to protect Retrieval-Augmented Generation (RAG) systems from data poisoning attacks. RAGuard, presented in two papers, employs a layered approach including adversarial r…
-
PLAID-PRF enhances dense retrieval with centroid-aware pseudo-relevance feedback
Researchers have introduced PLAID-PRF, a novel method for enhancing multi-vector dense retrieval models like ColBERT. This technique leverages centroid-like tokens within the PLAID framework to perform Pseudo-Relevance …
-
KaLM-Reranker-V1: Efficient Document Reranking Model Unveiled
Researchers have introduced KaLM-Reranker-V1, a novel document reranking model designed for efficiency and flexibility in retrieval systems. This model decouples query and passage computation, allowing for faster proces…
-
TileMaxSim kernel boosts GPU retrieval model speed by 220x
Researchers have developed TileMaxSim, a new IO-aware kernel for GPUs designed to significantly accelerate the MaxSim scoring process used in multi-vector retrieval models like ColBERT. Existing implementations are inef…