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

Inferring programs which generate 2D and 3D shapes is important for reverseengineering, editing, and more. Training models to perform this task iscomplicated because paired (shape, program) data is not readily available formany domains, making exact supervised learning infeasible. However, it ispossible to get paired data by compromising the accuracy of either the assignedprogram labels or the shape distribution. Wake-sleep methods use samples from agenerative model of shape programs to approximate the distribution of realshapes. In self-training, shapes are passed through a recognition model, whichpredicts programs that are treated as "pseudo-labels" for those shapes. Relatedto these approaches, we introduce a novel self-training variant unique toprogram inference, where program pseudo-labels are paired with their executedoutput shapes, avoiding label mismatch at the cost of an approximate shapedistribution. We propose to group these regimes under a single conceptualframework, where training is performed with maximum likelihood updates sourcedfrom either Pseudo-Labels or an Approximate Distribution (PLAD). We evaluatethese techniques on multiple 2D and 3D shape program inference domains.Compared with policy gradient reinforcement learning, we show that PLADtechniques infer more accurate shape programs and converge significantlyfaster. Finally, we propose to combine updates from different PLAD methodswithin the training of a single model, and find that this approach outperformsany individual technique.

PLAD: Learning to Infer Shape Programs with Pseudo-Labels and Approximate Distributions