English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Supervised learning can improve the design of state-of-the-art solvers forcombinatorial problems, but labelling large numbers of combinatorial instancesis often impractical due to exponential worst-case complexity. Inspired by therecent success of contrastive pre-training for images, we conduct a scientificstudy of the effect of augmentation design on contrastive pre-training for theBoolean satisfiability problem. While typical graph contrastive pre-traininguses label-agnostic augmentations, our key insight is that many combinatorialproblems have well-studied invariances, which allow for the design oflabel-preserving augmentations. We find that label-preserving augmentations arecritical for the success of contrastive pre-training. We show that ourrepresentations are able to achieve comparable test accuracy tofully-supervised learning while using only 1% of the labels. We alsodemonstrate that our representations are more transferable to larger problemsfrom unseen domains. Our code is available athttps://github.com/h4duan/contrastive-sat.

Augment with Care: Contrastive Learning for Combinatorial Problems