Inducing Systematicity in Transformers by Attending to Structurally Quantized Embeddings

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3AYBRRMRQA" target="_blank" >RIV/00216208:11320/25:YBRRMRQA - isvavai.cz</a>

Result on the web

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85204459093&partnerID=40&md5=60c1c0c5d1df214472490765afd3fade" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85204459093&partnerID=40&md5=60c1c0c5d1df214472490765afd3fade</a>

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

Inducing Systematicity in Transformers by Attending to Structurally Quantized Embeddings

Original language description

Transformers generalize to novel compositions of structures and entities after being trained on a complex dataset, but easily overfit on datasets of insufficient complexity. We observe that when the training set is sufficiently complex, the model encodes structurally equivalent sentences using a systematic attention pattern. Inspired by this observation, we propose SQ-Transformer (Structurally Quantized) that explicitly encourages systematicity in the embeddings and attention layers even with low-complexity data. At the embedding level, we introduce Structure-oriented Vector Quantization (SoVQ) to cluster word embeddings into several classes of structurally equivalent entities. At the attention level, we devise the Systematic Attention Layer (SAL) and an alternative, Systematically Regularized Layer (SRL) that operate on the quantized word embeddings so that sentences of the same structure are encoded with invariant or similar attention patterns. Empirically, we show SQ-Transformer achieves stronger compositional generalization than the vanilla Transformer on multiple low-complexity semantic parsing and machine translation datasets. In our analysis, we show SoVQ indeed learns a syntactically clustered embedding space, and SAL/SRL induces generalizable attention patterns, altogether leading to improved systematicity. © 2024 Association for Computational Linguistics.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

—

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

Proc. Annu. Meet. Assoc. Comput Linguist.

ISBN

979-889176094-3

ISSN

0736-587X

e-ISSN

—

Number of pages

24

Pages from-to

8360-8383

Publisher name

Association for Computational Linguistics (ACL)

Place of publication

—

Event location

Bangkok

Event date

Jan 1, 2025

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—