Summary of Discover-then-name: Task-agnostic Concept Bottlenecks Via Automated Concept Discovery, by Sukrut Rao et al.
Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discovery
by Sukrut Rao, Sweta Mahajan, Moritz Böhle, Bernt Schiele
First submitted to arxiv on: 19 Jul 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper proposes a novel Concept Bottleneck Model (CBM) approach, called Discover-then-Name-CBM (DN-CBM), to address the “black-box” problem of deep neural networks. By inverting the typical paradigm, DN-CBM first discovers concepts learnt by the model using sparse autoencoders and then names them for linear probe classification. This method is efficient and agnostic to the downstream task, leveraging concepts already known to the model. The paper evaluates its approach across multiple datasets and CLIP architectures, demonstrating semantically meaningful concepts, interpretable naming, and performant CBMs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper wants to make deep neural networks more understandable by creating a special type of model called Concept Bottleneck Models (CBMs). These models map images into a space that humans can understand, then use these concepts to make predictions. Usually, you have to choose the right concepts before building the CBM, but this new approach is different – it first figures out what concepts the model has learned and then gives them names for easy interpretation. The paper tested its method on many datasets and showed that it works well. |
Keywords
* Artificial intelligence * Classification
