Toward a Characterization of Loss Functions for Distribution Learning
Authors: Nika Haghtalab, Cameron Musco, Bo Waggoner
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | In this work we study loss functions for learning and evaluating probability distributions over large discrete domains. Unlike classification or regression where a wide variety of loss functions are used, in the distribution learning and density estimation literature, very few losses outside the dominant log loss are applied. We aim to understand this fact, taking an axiomatic approach to the design of loss functions for distributions. We start by proposing a set of desirable criteria that any good loss function should satisfy. Intuitively, these criteria require that the loss function faithfully evaluates a candidate distribution, both in expectation and when estimated on a few samples. |
| Researcher Affiliation | Collaboration | Nika Haghtalab Cornell University nika@cs.cornell.edu Cameron Musco UMass Amherst cmusco@cs.umass.edu Bo Waggoner U. Colorado bwag@colorado.edu Research conducted while at Microsoft Research, New England. Research conducted while at Microsoft Research, New York City. |
| Pseudocode | No | The paper does not contain any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | No | The paper does not provide any statement about releasing source code or a link to a code repository for the described methodology. |
| Open Datasets | No | The paper is theoretical and does not report on experiments using a specific dataset that would require public access information. It references other works for general context but does not provide access details for a dataset used in its own research. |
| Dataset Splits | No | The paper is theoretical and does not report on experiments with data, therefore it does not provide details on training, validation, or test dataset splits. |
| Hardware Specification | No | The paper is theoretical and does not describe experimental procedures, thus no hardware specifications are mentioned. |
| Software Dependencies | No | The paper is theoretical and does not describe experimental procedures that would involve specific software dependencies with version numbers. |
| Experiment Setup | No | The paper is theoretical and focuses on mathematical properties of loss functions, not on specific experimental setups or hyperparameter details. |