Semantic Clustering based Deduction Learning for Image Recognition and Classification

TL;DR

This paper introduces a semantic clustering based deduction learning method that enhances image recognition by enabling models to infer class relations and improve robustness through semantic priors and clustering in the feature space.

Contribution

It proposes a novel semantic clustering approach that mimics human reasoning, guiding deep models to learn class relations and improve robustness with semantic priors.

Findings

Outperforms state-of-the-art classifiers on benchmark datasets

Demonstrates improved robustness with noisy labels

Shows effective high-level semantic clustering in feature space

Abstract

The paper proposes a semantic clustering based deduction learning by mimicking the learning and thinking process of human brains. Human beings can make judgments based on experience and cognition, and as a result, no one would recognize an unknown animal as a car. Inspired by this observation, we propose to train deep learning models using the clustering prior that can guide the models to learn with the ability of semantic deducing and summarizing from classification attributes, such as a cat belonging to animals while a car pertaining to vehicles. %Specifically, if an image is labeled as a cat, then the model is trained to learn that "this image is totally not any random class that is the outlier of animal". The proposed approach realizes the high-level clustering in the semantic space, enabling the model to deduce the relations among various classes during the learning process. In addition, the paper introduces a semantic prior based random search for the opposite labels to ensure the smooth distribution of the clustering and the robustness of the classifiers. The proposed approach is supported theoretically and empirically through extensive experiments. We compare the performance across state-of-the-art classifiers on popular benchmarks, and the generalization ability is verified by adding noisy labeling to the datasets. Experimental results demonstrate the superiority of the proposed approach.