The reviewed record of science sign in
Pith

arxiv: 2106.09488 · v1 · pith:VOFTGECP · submitted 2021-06-11 · eess.AS · cs.CL· cs.LG· cs.SD

Scaling Laws for Acoustic Models

Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:VOFTGECPrecord.jsonopen to challenge →

classification eess.AS cs.CLcs.LGcs.SD
keywords modelsizelawsscalingtrainingmodelslossacoustic
0
0 comments X
read the original abstract

There is a recent trend in machine learning to increase model quality by growing models to sizes previously thought to be unreasonable. Recent work has shown that autoregressive generative models with cross-entropy objective functions exhibit smooth power-law relationships, or scaling laws, that predict model quality from model size, training set size, and the available compute budget. These scaling laws allow one to choose nearly optimal hyper-parameters given constraints on available training data, model parameter count, or training computation budget. In this paper, we demonstrate that acoustic models trained with an auto-predictive coding loss behave as if they are subject to similar scaling laws. We extend previous work to jointly predict loss due to model size, to training set size, and to the inherent "irreducible loss" of the task. We find that the scaling laws accurately match model performance over two orders of magnitude in both model size and training set size, and make predictions about the limits of model performance.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 2 Pith papers

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

  1. Scaling Laws and Interpretability of Learning from Repeated Data

    cs.LG 2022-05 accept novelty 6.0

    Repeating 0.1% of training data 100 times degrades an 800M parameter model's performance to that of a 400M model by damaging copying mechanisms and induction heads associated with generalization.

  2. The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models

    cs.LG 2022-01 conditional novelty 6.0

    More capable RL agents exploit reward misspecifications more often, with phase transitions in behavior, and anomaly detectors can identify misaligned policies.