Pith. sign in

REVIEW 1 cited by

Implicit biases in multitask and continual learning from a backward error analysis perspective

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 2311.00235 v1 pith:QBK3CHYM submitted 2023-11-01 stat.ML cs.AIcs.LG

Implicit biases in multitask and continual learning from a backward error analysis perspective

classification stat.ML cs.AIcs.LG
keywords learningtermimplicitconflictcontinualmultitaskanalysisbackward
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Using backward error analysis, we compute implicit training biases in multitask and continual learning settings for neural networks trained with stochastic gradient descent. In particular, we derive modified losses that are implicitly minimized during training. They have three terms: the original loss, accounting for convergence, an implicit flatness regularization term proportional to the learning rate, and a last term, the conflict term, which can theoretically be detrimental to both convergence and implicit regularization. In multitask, the conflict term is a well-known quantity, measuring the gradient alignment between the tasks, while in continual learning the conflict term is a new quantity in deep learning optimization, although a basic tool in differential geometry: The Lie bracket between the task gradients.

discussion (0)

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

Forward citations

Cited by 1 Pith paper

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

  1. Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise

    cs.LG 2026-06 unverdicted novelty 7.0

    Fixed-clock optimizer memory turns equal-multiset data shuffle order into an O(η) source of fine-tuning noise, larger than the O(η²) effect in memoryless cases, with a fit-free sizing method derived.