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

We study compositional generalization, viz., the problem of zero-shotgeneralization to novel compositions of concepts in a domain. Standard neuralnetworks fail to a large extent on compositional learning. We propose TreeStack Memory Units (Tree-SMU) to enable strong compositional generalization.Tree-SMU is a recursive neural network with Stack Memory Units (\SMU s), anovel memory augmented neural network whose memory has a differentiable stackstructure. Each SMU in the tree architecture learns to read from its stack andto write to it by combining the stacks and states of its children throughgating. The stack helps capture long-range dependencies in the problem domain,thereby enabling compositional generalization. Additionally, the stack alsopreserves the ordering of each node's descendants, thereby retaining localityon the tree. We demonstrate strong empirical results on two mathematicalreasoning benchmarks. We use four compositionality tests to assess thegeneralization performance of Tree-SMU and show that it enables accuratecompositional generalization compared to strong baselines such as Transformersand Tree-LSTMs.

Compositional Generalization with Tree Stack Memory Units