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

Compositional generalization is a troubling blind spot for neural languagemodels. Recent efforts have presented techniques for improving a model'sability to encode novel combinations of known inputs, but less work has focusedon generating novel combinations of known outputs. Here we focus on this latter"decode-side" form of generalization in the context of gSCAN, a syntheticbenchmark for compositional generalization in grounded language understanding.We present Recursive Decoding (RD), a novel procedure for training and usingseq2seq models, targeted towards decode-side generalization. Rather thangenerating an entire output sequence in one pass, models are trained to predictone token at a time. Inputs (i.e., the external gSCAN environment) are thenincrementally updated based on predicted tokens, and re-encoded for the nextdecoder time step. RD thus decomposes a complex, out-of-distribution sequencegeneration task into a series of incremental predictions that each resemblewhat the model has already seen during training. RD yields dramatic improvementon two previously neglected generalization tasks in gSCAN. We provide analysesto elucidate these gains over failure of a baseline, and then discussimplications for generalization in naturalistic grounded languageunderstanding, and seq2seq more generally.

Recursive Decoding: A Situated Cognition Approach to Compositional Generation in Grounded Language Understanding