Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Montreuil, Yannis, Yeo, Shu Heng, Carlier, Axel, Ng, Lai Xing, Ooi, Wei Tsang
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2410.15761
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918528392626176
author Montreuil, Yannis
Yeo, Shu Heng
Carlier, Axel
Ng, Lai Xing
Ooi, Wei Tsang
author_facet Montreuil, Yannis
Yeo, Shu Heng
Carlier, Axel
Ng, Lai Xing
Ooi, Wei Tsang
contents Large Language Models excel in generative tasks but exhibit inefficiencies in structured text selection, particularly in extractive question answering. This challenge is magnified in resource-constrained environments, where deploying multiple specialized models for different tasks is impractical. We propose a Learning-to-Defer framework that allocates queries to specialized experts, ensuring high-confidence predictions while optimizing computational efficiency. Our approach integrates a principled allocation strategy with theoretical guarantees on optimal deferral that balances performance and cost. Empirical evaluations on SQuADv1, SQuADv2, and TriviaQA demonstrate that our method enhances answer reliability while significantly reducing computational overhead, making it well-suited for scalable and efficient EQA deployment.
format Preprint
id arxiv_https___arxiv_org_abs_2410_15761
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Optimal Query Allocation in Extractive QA with LLMs: A Learning-to-Defer Framework with Theoretical Guarantees
Montreuil, Yannis
Yeo, Shu Heng
Carlier, Axel
Ng, Lai Xing
Ooi, Wei Tsang
Computation and Language
Machine Learning
Large Language Models excel in generative tasks but exhibit inefficiencies in structured text selection, particularly in extractive question answering. This challenge is magnified in resource-constrained environments, where deploying multiple specialized models for different tasks is impractical. We propose a Learning-to-Defer framework that allocates queries to specialized experts, ensuring high-confidence predictions while optimizing computational efficiency. Our approach integrates a principled allocation strategy with theoretical guarantees on optimal deferral that balances performance and cost. Empirical evaluations on SQuADv1, SQuADv2, and TriviaQA demonstrate that our method enhances answer reliability while significantly reducing computational overhead, making it well-suited for scalable and efficient EQA deployment.
title Optimal Query Allocation in Extractive QA with LLMs: A Learning-to-Defer Framework with Theoretical Guarantees
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2410.15761