A Unified Approach to Count-Based Weakly-Supervised Learning

2023/11/22 by Shukla, Vinay, Zeng, Zhe, Ahmed, Kareem +1 · 3 citations
#Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · doi:10.48550/arxiv.2311.13718

Abstract

High-quality labels are often very scarce, whereas unlabeled data with inferred weak labels occurs more naturally. In many cases, these weak labels dictate the frequency of each respective class over a set of instances. In this paper, we develop a unified approach to learning from such weakly-labeled data, which we call count-based weakly-supervised learning. At the heart of our approach is the ability to compute the probability of exactly k out of n outputs being set to true. This computation is differentiable, exact, and efficient. Building upon the previous computation, we derive a count loss penalizing the model for deviations in its distribution from an arithmetic constraint defined over label counts. We evaluate our approach on three common weakly-supervised learning paradigms and observe that our proposed approach achieves state-of-the-art or highly competitive results across all three of the paradigms.

Cited by

• Learning Robust Diffusion Models from Imprecise Supervision
• Accessible, Realistic, and Fair Evaluation of Positive-Unlabeled Learning Algorithms
• Exploiting the Potential Supervision Information of Clean Samples in Partial Label Learning

Related

• Bandit Label Inference for Weakly Supervised Learning
• Data Consistency for Weakly Supervised Learning
• Easy Learning from Label Proportions
• Reliable Weakly Supervised Learning: Maximize Gain and Maintain Safeness
• Label Propagation with Weak Supervision