Label Augmentation with Reinforced Labeling for Weak Supervision
Not on the map. Why?
Abstract (arXiv)
Weak supervision (WS) is an alternative to the traditional supervised learning to address the need for ground truth. Data programming is a practical WS approach that allows programmatic labeling data samples using labeling functions (LFs) instead of hand-labeling each data point. However, the existing approach fails to fully exploit the domain knowledge encoded into LFs, especially when the LFs' coverage is low. This is due to the common data programming pipeline that neglects to utilize data features during the generative process. This paper proposes a new approach called reinforced labeling (RL). Given an unlabeled dataset and a set of LFs, RL augments the LFs' outputs to cases not covered by LFs based on similarities among samples. Thus, RL can lead to higher labeling coverage for training an end classifier. The experiments on several domains (classification of YouTube comments, wine quality, and weather prediction) result in considerable gains. The new approach produces significant performance improvement, leading up to +21 points in accuracy and +61 points in F1 scores compared to the state-of-the-art data programming approach.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 32 of its 46 references that are in the map. First posted on arXiv 13 Apr 2022.
Builds on
32 of 46 references.
References · 32
- 2017Snorkel: Rapid Training Data Creation with Weak Supervisioncited by 372
- 2016Data Programming: Creating Large Training Sets, Quicklycited by 246
- 2018Training Complex Models with Multi-Task Weak Supervisioncited by 106
- 2019Implicit Semantic Data Augmentation for Deep Networkscited by 106
- 2020Fast and Three-rious: Speeding Up Weak Supervision with Triplet Methodscited by 98
- 2017Who Said What: Modeling Individual Labelers Improves Classificationcited by 77
- 2017Neural Ranking Models with Weak Supervisioncited by 61
- 2020Learning from Rules Generalizing Labeled Exemplarscited by 59
- 2018Training Classifiers with Natural Language Explanationscited by 58
- 2018Snorkel DryBell: A Case Study in Deploying Weak Supervision at Industrial Scalecited by 52
- 2020Interactive Weak Supervision: Learning Useful Heuristics for Data Labelingcited by 52
- 2019Augmenting Neural Networks with First-order Logiccited by 51
20 more
- 2017Learning the Structure of Generative Models without Labeled Datacited by 50
- 2017A Bayesian Data Augmentation Approach for Learning Deep Modelscited by 50
- 2017Learning From Noisy Singly-labeled Datacited by 50
- 2021Self-Training with Weak Supervisioncited by 49
- 2020Denoising Multi-Source Weak Supervision for Neural Text Classificationcited by 42
- 2019Learning Dependency Structures for Weak Supervision Modelscited by 32
- 2017Estimating Accuracy from Unlabeled Data: A Probabilistic Logic Approachcited by 27
- 2019NERO: A Neural Rule Grounding Framework for Label-Efficient Relation Extractioncited by 24
- 2019Slice-based Learning: A Programming Model for Residual Learning in Critical Data Slicescited by 23
- 2017Fidelity-Weighted Learningcited by 19
- 2019GOGGLES: Automatic Image Labeling with Affinity Codingcited by 19
- 2017Inferring Generative Model Structure with Static Analysiscited by 18
- 2018Adversarial Label Learningcited by 9
- 2021Knodle: Modular Weakly Supervised Learning with PyTorchcited by 7
- 2018Bootstrapping Conversational Agents With Weak Supervisioncited by 6
- 2018Adversarial Data Programming: Using GANs to Relax the Bottleneck of Curated Labeled Datacited by 4
- 2020Train and You'll Miss It: Interactive Model Iteration with Weak Supervision and Pre-Trained Embeddingscited by 4
- 2020Learning Structured Representations of Entity Names using Active Learning and Weak Supervisioncited by 4
- 2016Socratic Learning: Augmenting Generative Models to Incorporate Latent Subsets in Training Datacited by 3
- 2020Generative Adversarial Data Programmingcited by 2
Built on by
Citing papers per year: 2023: 1.
1 paper.