Modern machine learning demands a large amount of training data. Weak supervision is a promising approach to meet this demand. It aggregates multiple labeling functions (LFs)–noisy, user-provided labeling heuristics—to rapidly and cheaply curate probabilistic labels for large-scale unlabeled data. However, standard assumptions in weak supervision—such as user-specified class balance, similar accuracy of an LF in classifying different classes, and full knowledge of LF dependency at inference time—might be undesirable in practice. In response, we present Firebolt, a new weak supervision framework that seeks to operate under weaker assumptions. In particular, Firebolt learns the class balance and class-specific accuracy of LFs jointly from unlabeled data. It carries out inference in an efficient and interpretable manner. We analyze the parameter estimation error of Firebolt and characterize its impact on downstream model performance. Furthermore, we show that on five publicly available datasets, Firebolt outperforms a state-of-the-art weak supervision method by up to 5.8 points in AUC. We also provide a case study in the production setting of a tech company, where a Firebolt-supervised model outperforms the existing weakly-supervised production model by 1.3 points in AUC and speedup label model training and inference from one hour to three minutes.

Curation of large fully supervised datasets has become one of the major roadblocks for machine learning. Weak supervision provides an alternative to supervised learning by training with cheap, noisy, and possibly correlated labeling functions from varying sources. The key challenge in weakly supervised learning is combining the different weak supervision signals while navigating misleading correlations in their errors. In this paper, we propose a simple data-free approach for combining weak supervision signals by defining a constrained space for the possible labels of the weak signals and training with a random labeling within this constrained space. Our method is efficient and stable, converging after a few iterations of gradient descent. We prove theoretical conditions under which the worst-case error of the randomized label decreases with the rank of the linear constraints. We show experimentally that our method outperforms other weak supervision methods on various text- and image-classification tasks.

We consider the task of training classifiers without fully labeled data. We propose a weakly supervised method---adversarial label learning---that trains classifiers to perform well when noisy and possibly correlated labels are provided. Our framework allows users to provide different weak labels and multiple constraints on these labels. Our model then attempts to learn parameters for the data by solving a zero-sum game for the binary problems and a non-zero sum game optimization for multi-class problems. The game is between an adversary that chooses labels for the data and a model that minimizes the error made by the adversarial labels. The weak supervision constrains what labels the adversary can choose. The method therefore minimizes an upper bound of the classifier's error rate using projected primal-dual subgradient descent. Minimizing this bound protects against bias and dependencies in the weak supervision. We first show the performance of our framework on binary classification tasks then we extend our algorithm to show its performance on multiclass datasets. Our experiments show that our method can train without labels and outperforms other approaches for weakly supervised learning.

Social media plays a major role during and after major natural disasters (e.g., hurricanes, large-scale fires, etc.), as people “on the ground” post useful information on what is actually happening. Given the large amounts of posts, a major challenge is identifying the information that is useful and actionable. Emergency responders are largely interested in finding out what events are taking place so they can properly plan and deploy resources. In this paper we address the problem of automatically identifying important sub-events (within a large-scale emergency “event”, such as a hurricane). In particular, we present a novel, unsupervised learning framework to detect sub-events in Tweets for retrospective crisis analysis. We first extract noun-verb pairs and phrases from raw tweets as sub-event candidates. Then, we learn a semantic embedding of extracted noun-verb pairs and phrases, and rank them against a crisis-specific ontology. We filter out noisy and irrelevant information then cluster the noun-verb pairs and phrases so that the top-ranked ones describe the most important sub-events. Through quantitative experiments on two large crisis data sets (Hurricane Harvey and the 2015 Nepal Earthquake), we demonstrate the effectiveness of our approach over the state-of-the-art. Our qualitative evaluation shows better performance compared to our baseline.

We consider the task of training classifiers without labels. We propose a weakly supervised method—adversarial label learning—that trains classifiers to perform well against an adversary that chooses labels for training data. The weak supervision constrains what labels the adversary can choose. The method therefore minimizes an upper bound of the classifier’s error rate using projected primal-dual subgradient descent. Minimizing this bound protects against bias and dependencies in the weak supervision. Experiments on real datasets show that our method can train without labels and outperforms other approaches for weakly supervised learning.