Signals 4 · free daily AI digest

Type Diversity Enables Transformers to Generalise Compositionally

Paper recorded by Signals 4 on 2026-09-11 in cs.CL. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-11 on arXiv · recorded by Signals 4 on 2026-09-14

Category: cs.CL · 自然语言处理 · first seen 2026-09-14

Abstract

Compositional generalisation has been divided into lexical and structural generalisation. Previous work has found that structural generalisation is harder than lexical for Transformers. We propose that this difference is not inherent to Transformers, but due to the high diversity of lexical types and low diversity of structural types in the specific datasets of these previous works. By type divers

Read on arXiv →

#50 most recent of 186 cs.CL papers we have recorded · ↑ newer: RESKILL: Explicit Failure Attribution and Structured Repair for Intera · ↓ older: SAS: Simple Attention Sparsification via End-to-End Optimization of Co
Cite this page: Type Diversity Enables Transformers to Generalise Compositionally: the #50 most recent of 186 cs.CL papers we have recorded (as of 2026-09-11). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/type-diversity-enables-transformers-to-generalise-compositionally.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.CL papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions