CVPR Findings2026
Learning to Reason: Targeted Knowledge Discovery and Fuzzy Logic Update for Robust Image Recognition
Gurucharan Srinivas, Joshua Niemeijer, Frank Köster
Abstract
Integrating domain knowledge into deep neural networks is a promising way to improve generalization. Existing methods either encode prior knowledge in the loss function or apply post-processing modules, but both depend on iden- Figure 1. The figure shows the impact of integrating selfdiscovered implicit concepts via our Differentiable Knowledge Unit. Shown: input image, ground truth, initial probabilities p(y k |z), refined probabilities p∆(y k |ẑ), and visualizations of concept activations to provide insight into the learned semantics.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on3
- Swin Transformer V2: Scaling Up Capacity and ResolutionZe Liu, Han Hu, Yutong Lin, Zhuliang Yao et al.CVPR 2022 · 2,138 citations
- Concept Bottleneck ModelsPang Wei Koh, Thao Nguyen, Yew Siang Tang, Stephen Mussmann et al.ICML 2020 · 1,233 citations
- LogicSeg: Parsing Visual Semantics with Neural Logic Learning and ReasoningLiulei Li, Wenguan Wang, Yang YiICCV 2023 · 52 citations
Related papers
- Incorporating probabilistic domain knowledge into deep multiple instance learningGhadi S. Al Hajj, Aliaksandr Hubin, Chakravarthi Kanduri, Milena Pavlovic et al.ICML 2024 · 3 citations
- Towards Knowledge-augmented Bayesian Deep Learning For Computer VisionWang Ma, Hanjing Wang, Yufei Zhang, Darsha Udayanga et al.CVPR 2026
- GLUECons: A Generic Benchmark for Learning under ConstraintsHossein Rajaby Faghihi, Aliakbar Nafar, Chen Zheng, Roshanak Mirzaee et al.AAAI 2023 · 18 citations
- Implicit Differentiable Outlier Detection Enable Robust Deep Multimodal AnalysisZhu Wang, Sourav Medya, Sathya N. RaviNeurIPS 2023 · 4 citations
- Knowledge Consistency between Neural Networks and BeyondRuofan Liang, Tianlin Li, Longfei Li, Jing Wang et al.ICLR 2020 · 30 citations
