SEAL : Interactive Tool for Systematic Error Analysis and Labeling

TL;DR

SEAL is an interactive tool designed to identify and semantically characterize systematic errors in NLP models, especially on tail data or rare groups, using language and image models for better understanding.

Contribution

The paper introduces SEAL, a novel interactive toolkit that combines error slice identification with semantic labeling using language and image models for NLP error analysis.

Findings

Effective identification of high-error data slices.

Semantic labeling enhances understanding of failure modes.

Visual feature generation aids in characterizing error groups.

Abstract

With the advent of Transformers, large language models (LLMs) have saturated well-known NLP benchmarks and leaderboards with high aggregate performance. However, many times these models systematically fail on tail data or rare groups not obvious in aggregate evaluation. Identifying such problematic data groups is even more challenging when there are no explicit labels (e.g., ethnicity, gender, etc.) and further compounded for NLP datasets due to the lack of visual features to characterize failure modes (e.g., Asian males, animals indoors, waterbirds on land, etc.). This paper introduces an interactive Systematic Error Analysis and Labeling (\seal) tool that uses a two-step approach to first identify high error slices of data and then, in the second step, introduce methods to give human-understandable semantics to those underperforming slices. We explore a variety of methods for coming up with coherent semantics for the error groups using language models for semantic labeling and a text-to-image model for generating visual features. SEAL toolkit and demo screencast is available at https://huggingface.co/spaces/nazneen/seal.