AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Concepts from Representations: Post-hoc Concept Bottleneck Models via Sparse Decomposition of Visual Representations

Shizhan Gong, Xiaofan Zhang, Qi Dou

AAAI 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Concept bottleneck models · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
use PCBM-ReD interpretable bottleneck instead of end-to-end linear probeCLIP ViT-L/14 · 11 classification datasets averagetest accuracy (%): 87.38 -> 86.97No clear effect
PCBM-ReD instead of LaBoCLIP ViT-L/14 · 11 classification datasets averagetest accuracy (%): 85.72 -> 86.97Improved
PCBM-ReD instead of LaBo in few-shotPCBM-ReD · few-shot average over 11 datasets (1-16 shots)test accuracy (%): +5.01% on averageImproved
zero-shot with PCBM-ReD decomposition instead of original CLIP representationCLIP ViT-L/14 · zero-shot average over 11 datasetstest accuracy (%): 69.69 -> 69.73No clear effect
add reconstruction-guided concept selection instead of random or k-means (ablation)PCBM-ReD · CIFAR10/CIFAR100/CUB averagetest accuracyImproved
sparse representation decomposition instead of CLIP-similarity concept scores (ablation)PCBM-ReD · CIFAR10/CIFAR100/CUB averagetest accuracyImproved
data-driven concept creation instead of LLM-generated or WordNet concepts (ablation)PCBM-ReD · CIFAR10/CIFAR100/CUB averagetest accuracyImproved
extract concepts from a mismatched image encoder (ablation)PCBM-ReD · CIFAR10/CIFAR100/CUB averagetest accuracyHurt
increase bottleneck size (ablation)PCBM-ReD · CIFAR10/CIFAR100/CUB averagetest accuracy: saturates near 300 conceptsImproved
add LLM-based concept rating (ablation)PCBM-ReD · human evaluation, 5 datasetsexplanation rating (1-5)Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most 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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy