Research map / Label-efficient and robust learning / Recognition and long-tail
CoCoX: Generating Conceptual and Counterfactual Explanations via Fault-Lines
Arjun Akula, Shuai Wang, Song-Chun Zhu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use CoCoX fault-line explanations instead of no explanation | VGG-16 users · ImageNet 40 classes, non-expert pool | Justified Trust: 21.4 % -> 69.1 % | Improved |
| use CoCoX instead of no explanation | VGG-16 users · ImageNet 40 classes, expert pool | Justified Trust: 28.1 % -> 70.5 % | Improved |
| use CoCoX instead of best baseline (CEM) | non-expert users · ImageNet 40 classes, non-expert pool | Justified Trust: 51.0 % -> 69.1 % | Improved |
| use CoCoX instead of best baseline (CVE) | expert users · ImageNet 40 classes, expert pool | Justified Trust: 64.5 % -> 70.5 % | Improved |
| use attention-map explanations (CAM, Grad-CAM, SmoothGrad) instead of no explanation | non-expert users · ImageNet 40 classes, non-expert pool | Justified Trust: 21.4 % -> 29.2 % (Grad-CAM) | No clear effect |
| use concept-based (TCAV) or counterfactual (CEM, CVE) explanations instead of no explanation | users · ImageNet 40 classes, non-expert pool | Justified Trust: 21.4 % -> 49.7 % (TCAV) | Improved |
| combine concepts with counterfactuals (CoCoX) versus LIME on explanation satisfaction usefulness | non-expert users · ImageNet 40 classes, non-expert pool | Explanation Satisfaction usefulness: 4.2 -> 6.6 | Improved |
| more sessions of familiarization and testing | CoCoX group · non-expert pool over sessions | Justified Trust | Mixed |
Later papers that compared with this method
1 later accepted paper reports CoCoX as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Meaningfully debugging model mistakes using conceptual counterfactual explanations (ICML 2022 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Causal Concept Explanations for Deep Neural Models (NeurIPS 2026 · similarity 0.88)
- Counterfactual Explanations for Misclassified Images: How Human and Machine Explanations Differ (Abstract Reprint) (AAAI 2024 · similarity 0.87)
- Keep the Faith: Faithful Explanations in Convolutional Neural Networks for Case-Based Reasoning (AAAI 2024 · similarity 0.87)
- LINE: LLM-based Iterative Neuron Explanations for Vision Models (NeurIPS 2026 · similarity 0.86)
- Explaining in Diffusion: Explaining a Classifier with Diffusion Semantics (CVPR 2025 · similarity 0.86)
- Comparing the Decision-Making Mechanisms by Transformers and CNNs via Explanation Methods (CVPR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
