Text Retrieval Conference
PulseAugur coverage of Text Retrieval Conference — every cluster mentioning Text Retrieval Conference across labs, papers, and developer communities, ranked by signal.
-
New Study Audits Open Jev-Style Models for Retrieval Independence
A new arXiv paper investigates the request independence of open Jev-style models, specifically examining how changes to questions, evidence, and candidate eligibility affect retrieval outcomes. The study found that requ…
-
New SAFE-QE method tackles LLM hallucinated terms in query expansion
Researchers have developed a new method called SAFE-QE to address the issue of hallucinated terms generated by large language models (LLMs) during query expansion. This training-free, retriever-agnostic approach identif…
-
JevForest introduces new approach for budgeted feature acquisition
Researchers have developed JevForest, a novel feature acquisition policy designed for prediction tasks with limited observation budgets. This policy aggregates proposals from bootstrapped trees, weighting them by global…
-
LLM self-improvement loops suffer from 'winner's curse' due to noisy selection
A new arXiv paper explores the phenomenon of "winner's curse" in self-improving Large Language Models (LLMs). The study, which uses Qwen models to rewrite their own instructions, found that most proposed changes after t…
-
New SanSi model introduces 'System 1.5 thinking' for looped decision-making
Researchers have introduced SanSi, a novel looped typed decision model designed for "System 1.5 thinking," which aims to bridge the gap between rapid, single-pass System 1 thinking and full text generation. SanSi allows…
-
TREC 2025 Product Search Track details new recommendation dataset
The TREC 2025 Product Search and Recommendation Track has been detailed in a new overview paper, building upon previous iterations from 2023 and 2024. This track focuses on improving e-commerce search and recommendation…
-
New ERGO method optimizes text classification by learning from errors
Researchers have developed a new method called Error-Guided Optimization (ERGO) for text classification tasks. ERGO iteratively diagnoses classification failures on batches of data and generates targeted decision rules …
-
Small-population ES fine-tuning for LLMs shows promise with reward adjustments
A new research paper explores the effectiveness of Evolutionary Strategies (ES) for fine-tuning large language models, particularly when using binary rewards. The study found that the perceived need for large population…
-
New Geometric Approach Enhances Sparse Representation Classification Stability
Researchers have developed a new approach to Sparse Representation Classification (SRC) by focusing on the geometry of learned representations to ensure stable residual inference. Their work separates training from infe…
-
DoGMaTiQ pipeline automates QA nugget generation for report evaluation
Researchers have developed DoGMaTiQ, a new pipeline designed to automatically generate question-and-answer (QA) nuggets for evaluating long-form reports, particularly those generated by retrieval-augmented generation (R…
-
Lightweight Retrieval-Augmented Generation and Large Language Model-Based Modeling for Scalable Patient-Trial Matching
Researchers have developed new benchmarks and frameworks to evaluate and improve the performance of large language models (LLMs) in clinical settings. PhysicianBench offers a comprehensive evaluation for LLM agents on r…