When and Why does a Model Fail? A Human-in-the-loop Error Detection Framework for Sentiment Analysis

TL;DR

This paper introduces a human-in-the-loop framework for detecting errors in sentiment analysis models using explainable features, improving error identification before and after deployment.

Contribution

It presents a novel error detection framework combining global and local feature analysis with human assessment, enhancing model reliability.

Findings

High precision in identifying erroneous predictions with limited human intervention

Effective global and local feature contribution analysis for error detection

Improved model assessment prior to deployment

Abstract

Although deep neural networks have been widely employed and proven effective in sentiment analysis tasks, it remains challenging for model developers to assess their models for erroneous predictions that might exist prior to deployment. Once deployed, emergent errors can be hard to identify in prediction run-time and impossible to trace back to their sources. To address such gaps, in this paper we propose an error detection framework for sentiment analysis based on explainable features. We perform global-level feature validation with human-in-the-loop assessment, followed by an integration of global and local-level feature contribution analysis. Experimental results show that, given limited human-in-the-loop intervention, our method is able to identify erroneous model predictions on unseen data with high precision.