InterVLS: Interactive Model Understanding and Improvement with Vision-Language Surrogates
Not on the map. Why?
Abstract (arXiv)
Deep learning models are widely used in critical applications, highlighting the need for pre-deployment model understanding and improvement. Visual concept-based methods, while increasingly used for this purpose, face challenges: (1) most concepts lack interpretability, (2) existing methods require model knowledge, often unavailable at run time. Additionally, (3) there lacks a no-code method for post-understanding model improvement. Addressing these, we present InterVLS. The system facilitates model understanding by discovering text-aligned concepts, measuring their influence with model-agnostic linear surrogates. Employing visual analytics, InterVLS offers concept-based explanations and performance insights. It enables users to adjust concept influences to update a model, facilitating no-code model improvement. We evaluate InterVLS in a user study, illustrating its functionality with two scenarios. Results indicates that InterVLS is effective to help users identify influential concepts to a model, gain insights and adjust concept influence to improve the model. We conclude with a discussion based on our study results.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 23 of its 50 references that are in the map. First posted on arXiv 6 Nov 2023.
Builds on
23 of 50 references.
References · 23
- 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
- 2022Training language models to follow instructions with human feedbackcited by 15,712
- 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,690
- 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
- 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
- 2015Learning Deep Features for Discriminative Localizationcited by 2,430
- 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,348
- 2017Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)cited by 1,309
- 2017Network Dissection: Quantifying Interpretability of Deep Visual Representationscited by 703
- 2017Interpretable Explanations of Black Boxes by Meaningful Perturbationcited by 611
- 2017The (Un)reliability of saliency methodscited by 331
11 more
- 2018Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontierscited by 117
- 2019FairVis: Visual Analytics for Discovering Intersectional Bias in Machine Learningcited by 93
- 2020Interpretable and Accurate Fine-grained Recognition via Region Groupingcited by 75
- 2019Summit: Scaling Deep Learning Interpretability by Visualizing Activation and Attribution Summarizationscited by 72
- 2018Safely Entering the Deep: A Review of Verification and Validation for Machine Learning and a Challenge Elicitation in the Automotive Industrycited by 38
- 2019Role of Spatial Context in Adversarial Robustness for Object Detectioncited by 35
- 2023Charting the Sociotechnical Gap in Explainable AI: A Framework to Address the Gap in XAIcited by 33
- 2020VATLD: A Visual Analytics System to Assess, Understand and Improve Traffic Light Detectioncited by 28
- 2022ConceptExplainer: Interactive Explanation for Deep Neural Networks from a Concept Perspectivecited by 18
- 2021Human-in-the-loop Extraction of Interpretable Concepts in Deep Learning Modelscited by 17
- 2022Self-supervised Semantic Segmentation Grounded in Visual Conceptscited by 6
Built on by
Citing papers per year: 2024: 1.
1 paper.