paper-with-me

Papers

Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages

2021-10-16 · ACL 2022 5 · C. M. Downey, Shannon Drizin, Levon Haroutunian, Shivin Thukral

We show that unsupervised sequence-segmentation performance can be transferred to extremely low-resource languages by pre-training a Masked Segmental Language Model (Downey et al., 2021) multilingually. Further, we show that this transfer can be achieved by training over a collection of low-resource languages that are typologically similar (but phylogenetically unrelated) to the target language. In our experiments, we transfer from a collection of 10 Indigenous American languages (AmericasNLP, Mager et al., 2021) to K'iche', a Mayan language. We compare our multilingual model to a monolingual (from-scratch) baseline, as well as a model pre-trained on Quechua only. We show that the multilingual pre-trained approach yields consistent segmentation quality across target dataset sizes, exceeding the monolingual baseline in 6/10 experimental settings. Our model yields especially strong results at small target sizes, including a zero-shot performance of 20.6 F1. These results have promising implications for low-resource NLP pipelines involving human-like linguistic units, such as the sparse transcription framework proposed by Bird (2020).

📄 PDF Abstract BibTeX arXiv:2110.08415

Code (1)

cmdowney88/xlslm 공식 구현 pytorch

Tasks

Language ModelingLanguage ModellingSegmentation

Methods 이 논문이 사용한 방법론

American 설명 없음

Similar Papers 제목 키워드 기반

Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages

2021-11-16 · ACL ARR November 2021 11 · Anonymous

We show that unsupervised sequence-segmentation performance can be transferred to extremely low-resource languages by pre-training a Masked Segmental Language Model (Downey et al., 2021) multilingually. Further, we show …

Language ModelingLanguage Modelling

Morfessor-enriched features and multilingual training for canonical morphological segmentation

2022-07-01 · NAACL (SIGMORPHON) 2022 7 · Aku Rouhe, Stig-Arne Grönroos, Sami Virpioja, Mathias Creutz 외

In our submission to the SIGMORPHON 2022 Shared Task on Morpheme Segmentation, we study whether an unsupervised morphological segmentation method, Morfessor, can help in a supervised setting. Previous research has shown …

Morpheme SegmentaitonSentence

Bilingual alignment transfers to multilingual alignment for unsupervised parallel text mining

2021-04-15 · ACL 2022 5 · Chih-chan Tien, Shane Steinert-Threlkeld

This work presents methods for learning cross-lingual sentence representations using paired or unpaired bilingual texts. We hypothesize that the cross-lingual alignment strategy is transferable, and therefore a model tra…

Language ModelingLanguage ModellingRetrievalSentence+1

TAEC: Unsupervised Action Segmentation with Temporal-Aware Embedding and Clustering

2023-03-09 · Wei Lin, Anna Kukleva, Horst Possegger, Hilde Kuehne 외

Temporal action segmentation in untrimmed videos has gained increased attention recently. However, annotating action classes and frame-wise boundaries is extremely time consuming and cost intensive, especially on large-s…

Action SegmentationClusteringSegmentationTemporal Action Segmentation+1

MultiLinguahah : A New Unsupervised Multilingual Acoustic Laughter Segmentation Method

2026-05-07 · Sofia Callejas, Nahuel Gomez, Catherine Pelachaud, Brian Ravenet 외 arxiv

Laughter is a social non-vocalization that is universal across cultures and languages, and is crucial for human communication, including social bonding and communication signaling. However, detecting laughter in audio is…

Anomaly Detection