Pith. sign in

REVIEW 1 cited by

Improving Regression Performance with Distributional Losses

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1806.04613 v1 pith:6AWQWGKG submitted 2018-06-12 stat.ML cs.LG

classification stat.MLcs.LG
keywords regressionevidencelearninglossperformancealternativeconvertingdistributional
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

There is growing evidence that converting targets to soft targets in supervised learning can provide considerable gains in performance. Much of this work has considered classification, converting hard zero-one values to soft labels---such as by adding label noise, incorporating label ambiguity or using distillation. In parallel, there is some evidence from a regression setting in reinforcement learning that learning distributions can improve performance. In this work, we investigate the reasons for this improvement, in a regression setting. We introduce a novel distributional regression loss, and similarly find it significantly improves prediction accuracy. We investigate several common hypotheses, around reducing overfitting and improved representations. We instead find evidence for an alternative hypothesis: this loss is easier to optimize, with better behaved gradients, resulting in improved generalization. We provide theoretical support for this alternative hypothesis, by characterizing the norm of the gradients of this loss.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Aftab: A Comprehensive Benchmark of CNN Encoders and Advanced Value Functions in Parallelized Q-Networks

    cs.LG 2026-08 conditional novelty 5.0 of 10

    A composite CNN and multi-head value architecture, Aftab, improves buffer-free PQN on Atari-57 from 2.692 to 6.479 IQM human-normalized score, but its Procgen generalization advantage is fragile.

Pith tools