mlsanityData to 7 Oct 2026

The SaTML '24 CNN Interpretability Competition: New Innovations for Concept-Level Interpretability

arXiv 2404.02949 ↗First posted 3 Apr 2024cs.LG cs.AIcited by 3

Not on the map. Why?

Abstract (arXiv)

Interpretability techniques are valuable for helping humans understand and oversee AI systems. The SaTML 2024 CNN Interpretability Competition solicited novel methods for studying convolutional neural networks (CNNs) at the ImageNet scale. The objective of the competition was to help human crowd-workers identify trojans in CNNs. This report showcases the methods and results of four featured competition entries. It remains challenging to help humans reliably diagnose trojans via interpretability tools. However, the competition's entries have contributed new techniques and set a new record on the benchmark from Casper et al., 2023.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 15 of its 20 references that are in the map. First posted on arXiv 3 Apr 2024.

Builds on

15 of 20 references.

References · 15

  1. 2013Intriguing properties of neural networkscited by 5,114
  2. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  3. 2017Towards A Rigorous Science of Interpretable Machine Learningcited by 1,400
  4. 2017Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)cited by 1,309
  5. 2017Explanation in Artificial Intelligence: Insights from the Social Sciencescited by 1,201
  6. 2017Adversarial Patchcited by 514
  7. 2023Interpreting CLIP's Image Representation via Text-Based Decompositioncited by 189
  8. 2020Compositional Explanations of Neuronscited by 172
  9. 2022Natural Language Descriptions of Deep Visual Featurescited by 136
  10. 2022Toward Transparent AI: A Survey on Interpreting the Inner Structures of Deep Neural Networkscited by 126
  11. 2023Text-To-Concept (and Back) via Cross-Model Alignmentcited by 63
  12. 2021Robust Feature-Level Adversaries are Interpretability Toolscited by 27
3 more
  1. 2022Diagnostics for Deep Neural Networks with Automated Copy/Paste Attackscited by 6
  2. 2024Zoom-shot: Fast and Efficient Unsupervised Zero-Shot Transfer of CLIP to Vision Encoders with Multimodal Losscited by 3
  3. 2023Prototype Generation: Robust Feature Visualisation for Data Independent Interpretabilitycited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 2; 2025: 0; 2026: 1.

3 papers.

  1. 2026Interpretability Can Be Actionablecited by 5
  2. 2024TracrBench: Generating Interpretability Testbeds with Large Language Modelscited by 4
  3. 2024TextCAVs: Debugging vision models using textcited by 1