AttributionScanner: A Visual Analytics System for Model Validation with Metadata-Free Slice Finding

TL;DR

AttributionScanner is a visual analytics system that enables metadata-free data slice finding and model validation for vision models, helping users identify and address model biases and errors efficiently.

Contribution

The paper introduces a novel human-in-the-loop visual analytics system for metadata-free data slice detection and model validation in vision models, with an innovative attribution mosaic visualization.

Findings

Effective identification of model biases and mislabeled data

Improved model performance through targeted regularization

Demonstrated success on benchmark datasets

Abstract

Data slice finding is an emerging technique for validating machine learning (ML) models by identifying and analyzing subgroups in a dataset that exhibit poor performance, often characterized by distinct feature sets or descriptive metadata. However, in the context of validating vision models involving unstructured image data, this approach faces significant challenges, including the laborious and costly requirement for additional metadata and the complex task of interpreting the root causes of underperformance. To address these challenges, we introduce AttributionScanner, an innovative human-in-the-loop Visual Analytics (VA) system, designed for metadata-free data slice finding. Our system identifies interpretable data slices that involve common model behaviors and visualizes these patterns through an Attribution Mosaic design. Our interactive interface provides straightforward guidance for users to detect, interpret, and annotate predominant model issues, such as spurious correlations (model biases) and mislabeled data, with minimal effort. Additionally, it employs a cutting-edge model regularization technique to mitigate the detected issues and enhance the model's performance. The efficacy of AttributionScanner is demonstrated through use cases involving two benchmark datasets, with qualitative and quantitative evaluations showcasing its substantial effectiveness in vision model validation, ultimately leading to more reliable and accurate models.