Pith. sign in

REVIEW 2 cited by

Understanding self-supervised Learning Dynamics without Contrastive Pairs

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 2102.06810 v4 pith:L3UTNVO4 submitted 2021-02-12 cs.LG cs.AIcs.CV

classification cs.LGcs.AIcs.CV
keywords pairspredictorlearninglinearnon-contrastivesimplewithoutcollapse
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

While contrastive approaches of self-supervised learning (SSL) learn representations by minimizing the distance between two augmented views of the same data point (positive pairs) and maximizing views from different data points (negative pairs), recent \emph{non-contrastive} SSL (e.g., BYOL and SimSiam) show remarkable performance {\it without} negative pairs, with an extra learnable predictor and a stop-gradient operation. A fundamental question arises: why do these methods not collapse into trivial representations? We answer this question via a simple theoretical study and propose a novel approach, DirectPred, that \emph{directly} sets the linear predictor based on the statistics of its inputs, without gradient training. On ImageNet, it performs comparably with more complex two-layer non-linear predictors that employ BatchNorm and outperforms a linear predictor by $2.5\%$ in 300-epoch training (and $5\%$ in 60-epoch). DirectPred is motivated by our theoretical study of the nonlinear learning dynamics of non-contrastive SSL in simple linear networks. Our study yields conceptual insights into how non-contrastive SSL methods learn, how they avoid representational collapse, and how multiple factors, like predictor networks, stop-gradients, exponential moving averages, and weight decay all come into play. Our simple theory recapitulates the results of real-world ablation studies in both STL-10 and ImageNet. Code is released https://github.com/facebookresearch/luckmatters/tree/master/ssl.

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Depth-Regularized JEPA World Models Learn More Transferable Representations from Real Outdoor Robot Data

    cs.CV 2026-07 conditional novelty 5.0 of 10

    Adding a training-only depth-alignment loss to a compact LeWM JEPA world model improves downstream representation quality and out-of-domain transfer on real agricultural robot data.

  2. Contrastive Self-Supervised Network Intrusion Detection using Augmented Negative Pairs

    cs.LG 2025-09 conditional novelty 5.0 of 10

    CLAN clusters genuine benign network flows while repelling augmented copies, then classifies new flows by distance to the cluster centroid; on Lycos2017 it reports the highest mean AUROC among compared SSL and anomaly...

Pith tools