Information Retrieval

insight - Information Retrieval

Fusion-in-T5: A Unified Model for Effective Document Ranking by Integrating Multiple Signals

Fusion-in-T5 (FiT5) is a unified document ranking model that integrates text matching, ranking features, and global document information through attention fusion, outperforming complex cascade ranking pipelines.

Generative Relevance Feedback and Convergence of Adaptive Re-Ranking: Improving Passage Retrieval with Large Language Models and Corpus Graphs

Generative relevance feedback and adaptive re-ranking can improve passage retrieval performance, with the most effective approach combining generative pseudo-relevance feedback and adaptive re-ranking over a large corpus graph.

Efficient Inverted Indexes for Approximate Retrieval over Learned Sparse Representations

Seismic, a novel approximate nearest neighbor algorithm, enables efficient and effective retrieval over learned sparse embeddings by leveraging the concentration of importance property in these representations.

KamerRaad: Enhancing Political Information Retrieval and Engagement in Belgium through Hierarchical Summarization and Conversational Interfaces

KamerRaad is an AI tool that leverages large language models and hierarchical summarization to enhance citizens' ability to interactively engage with and understand Belgian parliamentary proceedings.

Counterfactual Queries as Explanations for Search Result Relevance

Providing counterfactual queries as explanations to help users understand and interact with search engine relevance decisions.

Exclusionary Neural Information Retrieval: Challenges and Opportunities

Existing retrieval models struggle to effectively comprehend exclusionary queries, where users explicitly express what they do not want to retrieve. Generative retrieval models exhibit unique advantages in handling such queries compared to sparse and dense retrieval methods.

Simple, Efficient and Effective Approximation of SPLADE for Faster Information Retrieval

A two-step approach to efficiently approximate SPLADE, a learned sparse retrieval model, while maintaining its effectiveness. The first step uses a pruned and reweighted version of SPLADE vectors for fast retrieval, and the second step rescores a sample of documents using the original SPLADE vectors.

Effective Generative Retrieval through Term Set Generation

A novel framework, Term-Set Generation (TSGen), that uses a set of terms as the document identifier (DocID) to address the false pruning problem in existing generative retrieval methods.

Assessing the Impact of Query Generation Strategies on Retrievability Scores

Varying the query generation techniques significantly impacts the computed retrievability scores, posing challenges for result reproducibility across different document collections.

Improving Early-Stage Retrieval with Entity Linking

Expanding the corpus with linked entity names can boost the performance of both sparse and dense retrievers in the early stage of cascaded ranking architectures.

1
2
3
4
5
•••
11

About

Products | Resources

Insights