34 citations · 129 across the 45 of their papers we have counts for
19 papers · 1 filter
Consistency Conditions for Differentiable Surrogate Losses
Drona Khurana, Anish Thilagar, Dhamma Kimpara +1
The statistical consistency of surrogate losses for discrete prediction tasks is often checked via the condition of calibration. However, directly verifying calibration can be ardu…
Structured Prediction with Abstention via the Lovász Hinge
Jessie Finocchiaro, Rafael Frongillo, Enrique Nueve
The Lovász hinge is a convex loss function proposed for binary structured classification, in which k related binary predictions jointly evaluated by a submodular function. Despite…
Hedging and Approximate Truthfulness in Traditional Forecasting Competitions
Mary Monroe, Anish Thilagar, Melody Hsu +1
In forecasting competitions, the traditional mechanism scores the predictions of each contestant against the outcome of each event, and the contestant with the highest total score…
Forecasting Competitions with Correlated Events
Rafael Frongillo, Manuel Lladser, Anish Thilagar +1
Beginning with Witkowski et al. [2022], recent work on forecasting competitions has addressed incentive problems with the common winner-take-all mechanism. Frongillo et al. [2021]…
Proper losses for discrete generative models
Rafael Frongillo, Dhamma Kimpara, Bo Waggoner
We initiate the study of proper losses for evaluating generative models in the discrete setting. Unlike traditional proper losses, we treat both the generative model and the target…
Consistent Polyhedral Surrogates for Top- Classification and Variants
Jessie Finocchiaro, Rafael Frongillo, Emma Goodwill +1
Top- classification is a generalization of multiclass classification used widely in information retrieval, image classification, and other extreme classification settings. Sever…
An Embedding Framework for the Design and Analysis of Consistent Polyhedral Surrogates
Jessie Finocchiaro, Rafael M. Frongillo, Bo Waggoner
We formalize and study the natural approach of designing convex surrogate loss functions via embeddings, for problems such as classification, ranking, or structured prediction. In…
The Structured Abstain Problem and the Lovász Hinge
Jessie Finocchiaro, Rafael Frongillo, Enrique Nueve
The Lovász hinge is a convex surrogate recently proposed for structured binary classification, in which binary predictions are made simultaneously and the error is judged by a…
Surrogate Regret Bounds for Polyhedral Losses
Rafael Frongillo, Bo Waggoner
Surrogate risk minimization is an ubiquitous paradigm in supervised machine learning, wherein a target problem is solved by minimizing a surrogate loss on a dataset. Surrogate regr…
Learning in Matrix Games can be Arbitrarily Complex
Gabriel P. Andrade, Rafael Frongillo, Georgios Piliouras
A growing number of machine learning architectures, such as Generative Adversarial Networks, rely on the design of games which implement a desired functionality via a Nash equilibr…
Unifying Lower Bounds on Prediction Dimension of Consistent Convex Surrogates
Jessie Finocchiaro, Rafael Frongillo, Bo Waggoner
Given a prediction task, understanding when one can and cannot design a consistent convex surrogate loss, particularly a low-dimensional one, is an important and active area of mac…
Efficient Competitions and Online Learning with Strategic Forecasters
Rafael Frongillo, Robert Gomez, Anish Thilagar +1
Winner-take-all competitions in forecasting and machine-learning suffer from distorted incentives. Witkowski et al. 2018 identified this problem and proposed ELF, a truthful mechan…
An Embedding Framework for Consistent Polyhedral Surrogates
Jessie Finocchiaro, Rafael Frongillo, Bo Waggoner
We formalize and study the natural approach of designing convex surrogate loss functions via embeddings, for problems such as classification, ranking, or structured prediction. In…
Multi-Observation Regression
Rafael Frongillo, Nishant A. Mehta, Tom Morgan +1
Recent work introduced loss functions which measure the error of a prediction based on multiple simultaneous observations or outcomes. In this paper, we explore the theoretical and…
Multi-Observation Elicitation
Sebastian Casalaina-Martin, Rafael Frongillo, Tom Morgan +1
We study loss functions that measure the accuracy of a prediction based on multiple data points simultaneously. To our knowledge, such loss functions have not been studied before i…
Elicitation Complexity of Statistical Properties
Rafael Frongillo, Ian A. Kash
A property, or statistical functional, is said to be elicitable if it minimizes expected loss for some loss function. The study of which properties are elicitable sheds light on th…
Generalized Mixability via Entropic Duality
Mark D. Reid, Rafael M. Frongillo, Robert C. Williamson +1
Mixability is a property of a loss which characterizes when fast convergence is possible in the game of prediction with expert advice. We show that a key property of mixability gen…
Generalised Mixability, Constant Regret, and Bayesian Updating
Mark D. Reid, Rafael M. Frongillo, Robert C. Williamson
Mixability of a loss is known to characterise when constant regret bounds are achievable in games of prediction with expert advice through the use of Vovk's aggregating algorithm.…
A Collaborative Mechanism for Crowdsourcing Prediction Problems
Jacob Abernethy, Rafael M. Frongillo
Machine Learning competitions such as the Netflix Prize have proven reasonably successful as a method of "crowdsourcing" prediction tasks. But these competitions have a number of w…