• Publications
  • Influence
Foreshadowing the Benefits of Incidental Supervision
TLDR
A unified PAC-Bayesian Informativeness measure (PABI) is proposed, characterizing the reduction in uncertainty that incidental supervision signals provide and demonstrating PABI's use in quantifying various types of incidental signals. Expand
Bayes Consistency vs. H-Consistency: The Interplay between Surrogate Loss Functions and the Scoring Function Class
TLDR
It is found that while some calibrated surrogates can indeed fail to provide H -consistency when minimized over a naturallooking but naïvely chosen scoring function class F , the situation can potentially be remedied by minimizing them over a more carefully chosen class of scoring functions F . Expand
Convex Calibrated Surrogates for the Multi-Label F-Measure
TLDR
This paper designs convex surrogate losses that have the property that minimizing the surrogate loss yields a Bayes optimal multi-label classifier for the F-measure, and provides a quantitative regret transfer bound for these surrogates. Expand
Ordinal Embedding with a Latent Factor Model
Constructing low-dimensional embeddings based on ordinal measurements has been a subject of significant recent interest, motivated in part by machine learning applications using human input in aExpand
Learning from Noisy Labels with No Change to the Training Process
TLDR
A quantitative regret transfer bound is provided, which bounds the target regret on the true distribution in terms of the CPE regrets on the noisy distribution, and suggests that the sample complexity of learning under CCN increases as the noise matrix approaches singularity. Expand