A

Abstract Sim Query

Developed by biu-nlp
A model that maps abstract sentence descriptions to matching sentences, trained on Wikipedia using a dual-encoder architecture.
Downloads 53
Release Time : 5/13/2023

Model Overview

This model encodes abstract query sentences into vector representations for similarity comparison with vectors generated by a sentence encoder, enabling matching of sentences to query descriptions.

Model Features

Dual-encoder architecture
Employs separate query and sentence encoders in a dual-encoder design, optimizing representations for different text types.
Abstract description matching
Specifically optimized for matching tasks between abstract query descriptions and concrete sentences.
Wikipedia-based training
Trained on Wikipedia data, making it suitable for similarity calculations in encyclopedic texts.

Model Capabilities

Sentence vectorization
Semantic similarity calculation
Abstract query matching

Use Cases

Information retrieval
Corporate relationship queries
Finding sentences describing subsidiary relationships based on abstract descriptions (e.g., 'a company being part of a larger company')
Accurately retrieves sentences describing subsidiary relationships with similarity scores significantly higher than irrelevant sentences
Knowledge base construction
Relationship fact extraction
Extracting sentences matching specific relationship patterns from text
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase