Pith. sign in

REVIEW

Retrieval Enhanced Model for Commonsense Generation

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 2105.11174 v1 pith:3QEXGUC7 submitted 2021-05-24 cs.CL cs.AI

Retrieval Enhanced Model for Commonsense Generation

classification cs.CL cs.AI
keywords commonsensegenerationsentencefine-tuningretrievalabilityachievesapproach
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Commonsense generation is a challenging task of generating a plausible sentence describing an everyday scenario using provided concepts. Its requirement of reasoning over commonsense knowledge and compositional generalization ability even puzzles strong pre-trained language generation models. We propose a novel framework using retrieval methods to enhance both the pre-training and fine-tuning for commonsense generation. We retrieve prototype sentence candidates by concept matching and use them as auxiliary input. For fine-tuning, we further boost its performance with a trainable sentence retriever. We demonstrate experimentally on the large-scale CommonGen benchmark that our approach achieves new state-of-the-art results.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.