Dan Garrette
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Dan Garrette has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Ekaterina Shutova
PCS ILLC ILLC ILLC University of Amsterdam University of Amsterdam University of Amsterdam University of Amsterdam
4 shared publications - Rochelle Choenni
University of Amsterdam University of Edinburgh
4 shared publications - Noah Constant
Google research
3 shared publications - Iulia Turc
Google research
2 shared publications Jan A. Botha
Google research
2 shared publications- Jason Riesa
Google research
2 shared publications
Latest Publications
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 287-301
Examining Modularity in Multilingual LMs via Language-Specialized Subnetworks
Rochelle Choenni, Ekaterina Shutova, Dan Garrette. Findings of the Association for Computational Linguistics: NAACL 2024. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 287-301
Open AccessExamining Modularity in Multilingual LMs via Language-Specialized Subnetworks
Rochelle Choenni, Ekaterina Shutova, Dan Garrette. Findings of the Association for Computational Linguistics: NAACL 2024. 2024.
- Open Access
Examining Modularity in Multilingual LMs via Language-Specialized Subnetworks
Recent work has proposed explicitly inducing language-wise modularity in multilingual LMs via sparse fine-tuning (SFT) on per-language subnetworks as a means of better guiding cross-lingual sharing. In this work, we investigate (1) the degree to which language-wise modularity naturally arises within models with no special modularity interventions, and (2) how cross-lingual sharing and interference differ between such models and those with explicit SFT-guided subnetwork modularity. To quantify language specialization and cross-lingual interaction, we use a Training Data Attribution method that estimates the degree to which a model's predictions are influenced by in-language or cross-language training examples. Our results show that language-specialized subnetworks do naturally arise, and that SFT, rather than always increasing modularity, can decrease language specialization of subnetworks in favor of more cross-lingual sharing.
Open AccessExamining Modularity in Multilingual LMs via Language-Specialized Subnetworks
Recent work has proposed explicitly inducing language-wise modularity in multilingual LMs via sparse fine-tuning (SFT) on per-language subnetworks as a means of better guiding cross-lingual sharing. In this work, we investigate (1) the degree to which language-wise modularity naturally arises within models with no special modularity interventions, and (2) how cross-lingual sharing and interference differ between such models and those with explicit SFT-guided subnetwork modularity. To quantify language specialization and cross-lingual interaction, we use a Training Data Attribution method that estimates the degree to which a model's predictions are influenced by in-language or cross-language training examples. Our results show that language-specialized subnetworks do naturally arise, and that SFT, rather than always increasing modularity, can decrease language specialization of subnetworks in favor of more cross-lingual sharing.
- Open Access
FRMT: A Benchmark for Few-Shot Region-Aware Machine Translation
We present FRMT, a new dataset and evaluation benchmark for Few-shot Region-aware Machine Translation, a type of style-targeted translation. The dataset consists of professional translations from English into two regional variants each of Portuguese and Mandarin Chinese. Source documents are selected to enable detailed analysis of phenomena of interest, including lexically distinct terms and distractor terms. We explore automatic evaluation metrics for FRMT and validate their correlation with expert human evaluation across both region-matched and mismatched rating scenarios. Finally, we present a number of baseline models for this task, and offer guidelines for how researchers can train, evaluate, and compare their own models. Our dataset and evaluation code are publicly available: https://bit.ly/frmt-task
Open AccessFRMT: A Benchmark for Few-Shot Region-Aware Machine Translation
We present FRMT, a new dataset and evaluation benchmark for Few-shot Region-aware Machine Translation, a type of style-targeted translation. The dataset consists of professional translations from English into two regional variants each of Portuguese and Mandarin Chinese. Source documents are selected to enable detailed analysis of phenomena of interest, including lexically distinct terms and distractor terms. We explore automatic evaluation metrics for FRMT and validate their correlation with expert human evaluation across both region-matched and mismatched rating scenarios. Finally, we present a number of baseline models for this task, and offer guidelines for how researchers can train, evaluate, and compare their own models. Our dataset and evaluation code are publicly available: https://bit.ly/frmt-task
- Open Access
- 10 August 2023
- Published by MIT Press in Computational Linguistics
- Vol. 49 (3) , 1-29
Cross-Lingual Transfer with Language-Specific Subnetworks for Low-Resource Dependency Parsing
Large multilingual language models typically share their parameters across all languages, which enables cross-lingual task transfer, but learning can also be hindered when training updates from different languages are in conflict. In this article, we propose novel methods for using language-specific subnetworks, which control cross-lingual parameter sharing, to reduce conflicts and increase positive transfer during fine-tuning. We introduce dynamic subnetworks, which are jointly updated with the model, and we combine our methods with meta-learning, an established, but complementary, technique for improving cross-lingual transfer. Finally, we provide extensive analyses of how each of our methods affects the models.
- 10 August 2023
- Published by MIT Press in Computational Linguistics
- Vol. 49 (3) , 1-29
Open AccessCross-Lingual Transfer with Language-Specific Subnetworks for Low-Resource Dependency Parsing
Large multilingual language models typically share their parameters across all languages, which enables cross-lingual task transfer, but learning can also be hindered when training updates from different languages are in conflict. In this article, we propose novel methods for using language-specific subnetworks, which control cross-lingual parameter sharing, to reduce conflicts and increase positive transfer during fine-tuning. We introduce dynamic subnetworks, which are jointly updated with the model, and we combine our methods with meta-learning, an established, but complementary, technique for improving cross-lingual transfer. Finally, we provide extensive analyses of how each of our methods affects the models.
- Top Cited PapersOpen Access
- 29 June 2023
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 11, 671-685
FRMT: A Benchmark for Few-Shot Region-Aware Machine Translation
We present FRMT, a new dataset and evaluation benchmark for Few-shot Region-aware Machine Translation, a type of style-targeted translation. The dataset consists of professional translations from English into two regional variants each of Portuguese and Mandarin Chinese. Source documents are selected to enable detailed analysis of phenomena of interest, including lexically distinct terms and distractor terms. We explore automatic evaluation metrics for FRMT and validate their correlation with expert human evaluation across both region-matched and mismatched rating scenarios. Finally, we present a number of baseline models for this task, and offer guidelines for how researchers can train, evaluate, and compare their own models. Our dataset and evaluation code are publicly available: https://bit.ly/frmt-task.
- 29 June 2023
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 11, 671-685
Top Cited PapersOpen AccessFRMT: A Benchmark for Few-Shot Region-Aware Machine Translation
We present FRMT, a new dataset and evaluation benchmark for Few-shot Region-aware Machine Translation, a type of style-targeted translation. The dataset consists of professional translations from English into two regional variants each of Portuguese and Mandarin Chinese. Source documents are selected to enable detailed analysis of phenomena of interest, including lexically distinct terms and distractor terms. We explore automatic evaluation metrics for FRMT and validate their correlation with expert human evaluation across both region-matched and mismatched rating scenarios. Finally, we present a number of baseline models for this task, and offer guidelines for how researchers can train, evaluate, and compare their own models. Our dataset and evaluation code are publicly available: https://bit.ly/frmt-task.
- Open Access
Character-Aware Models Improve Visual Text Rendering
Current image generation models struggle to reliably produce well-formed visual text. In this paper, we investigate a key contributing factor: popular text-to-image models lack character-level input features, making it much harder to predict a word's visual makeup as a series of glyphs. To quantify this effect, we conduct a series of experiments comparing character-aware vs. character-blind text encoders. In the text-only domain, we find that character-aware models provide large gains on a novel spelling task (WikiSpell). Applying our learnings to the visual domain, we train a suite of image generation models, and show that character-aware variants outperform their character-blind counterparts across a range of novel text rendering tasks (our DrawText benchmark). Our models set a much higher state-of-the-art on visual spelling, with 30+ point accuracy gains over competitors on rare words, despite training on far fewer examples.
Open AccessCharacter-Aware Models Improve Visual Text Rendering
Current image generation models struggle to reliably produce well-formed visual text. In this paper, we investigate a key contributing factor: popular text-to-image models lack character-level input features, making it much harder to predict a word's visual makeup as a series of glyphs. To quantify this effect, we conduct a series of experiments comparing character-aware vs. character-blind text encoders. In the text-only domain, we find that character-aware models provide large gains on a novel spelling task (WikiSpell). Applying our learnings to the visual domain, we train a suite of image generation models, and show that character-aware variants outperform their character-blind counterparts across a range of novel text rendering tasks (our DrawText benchmark). Our models set a much higher state-of-the-art on visual spelling, with 30+ point accuracy gains over competitors on rare words, despite training on far fewer examples.
- Top Cited PapersOpen Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 13244-13257
How do languages influence each other? Studying cross-lingual data sharing during LM fine-tuning
Rochelle Choenni, Dan Garrette, Ekaterina Shutova. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 13244-13257
Top Cited PapersOpen AccessHow do languages influence each other? Studying cross-lingual data sharing during LM fine-tuning
Rochelle Choenni, Dan Garrette, Ekaterina Shutova. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.
- Open Access
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 250
Fine-tuning mSLAM for the SIGMORPHON 2022 Shared Task on Grapheme-to-Phoneme Conversion
Dan Garrette. Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology. 2023.
- 1 January 2023
- Published by Association for Computational Linguistics (ACL)
- p. 250
Open AccessFine-tuning mSLAM for the SIGMORPHON 2022 Shared Task on Grapheme-to-Phoneme Conversion
Dan Garrette. Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology. 2023.
- Open Access
CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation
Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly-used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model's ability to adapt. In this paper, we present CANINE, a neural encoder that operates directly on character sequences, without explicit tokenization or vocabulary, and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, CANINE combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. CANINE outperforms a comparable mBERT model by 2.8 F1 on TyDi QA, a challenging multilingual benchmark, despite having 28% fewer model parameters.
Open AccessCANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation
Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly-used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model's ability to adapt. In this paper, we present CANINE, a neural encoder that operates directly on character sequences, without explicit tokenization or vocabulary, and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, CANINE combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. CANINE outperforms a comparable mBERT model by 2.8 F1 on TyDi QA, a challenging multilingual benchmark, despite having 28% fewer model parameters.
- Top Cited PapersOpen Access
- 31 January 2022
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 10, 73-91
Canine: Pre-training an Efficient Tokenization-Free Encoder for Language Representation
Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model’s ability to adapt. In this paper, we present Canine, a neural encoder that operates directly on character sequences—without explicit tokenization or vocabulary—and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, Canine combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. Canine outperforms a comparable mBert model by 5.7 F1 on TyDi QA, a challenging multilingual benchmark, despite having fewer model parameters.
- 31 January 2022
- Published by MIT Press in Transactions of the Association for Computational Linguistics
- Vol. 10, 73-91
Top Cited PapersOpen AccessCanine: Pre-training an Efficient Tokenization-Free Encoder for Language Representation
Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model’s ability to adapt. In this paper, we present Canine, a neural encoder that operates directly on character sequences—without explicit tokenization or vocabulary—and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, Canine combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. Canine outperforms a comparable mBert model by 5.7 F1 on TyDi QA, a challenging multilingual benchmark, despite having fewer model parameters.