My research is in artificial intelligence, focusing on the abilities and limitations of language models, and creating models for all kinds of human languages and language varieties. I am particularly interested in applications of formal language theory to both areas.
Teaching
Recent and selected publications
Akriti Dhasmana, Aarohi Srivastava, and David Chiang.
Donor
Rank: donor language selection for low-resource cross-lingual speech recognition.
2026.
arXiv:2608.11441.
PDF
BibTeX
@misc{dhasmana-etal-2026,
author = "Dhasmana, Akriti and Srivastava, Aarohi and Chiang, David",
title = "{D}onor{R}ank: Donor Language Selection for Low-Resource Cross-Lingual Speech Recognition",
year = "2026",
note = "arXiv:2608.11441",
url = "https://arxiv.org/abs/2608.11441"
}
Yijia Jessica Zhu and David Chiang.
Exact-solution volume and length generalization in transformers.
2026.
arXiv:2610.07676.
PDF
BibTeX
@misc{zhu-chiang-2026,
author = "Zhu, Yijia Jessica and Chiang, David",
title = "Exact-Solution Volume and Length Generalization in Transformers",
year = "2026",
note = "arXiv:2610.07676",
url = "https://arxiv.org/abs/2610.07676"
}
Gavin Dooley, Andy Yang, Yijia Jessica Zhu, David Chiang, Peter Cholak, and Anand Pillay.
VC dimension and expressivity of real-valued transformers.
2026.
arXiv:2609.31999.
PDF
BibTeX
@misc{dooley-etal-2026,
author = "Dooley, Gavin and Yang, Andy and Zhu, Yijia Jessica and Chiang, David and Cholak, Peter and Pillay, Anand",
title = "{VC} Dimension and Expressivity of Real-Valued Transformers",
year = "2026",
note = "arXiv:2609.31999",
url = "https://arxiv.org/abs/2609.31999"
}
Chihiro Taguchi, Éric
Le Ferrand, Hirosi Nakagawa, Hitomi Ono, Kanji Kato, Emily Prud'hommeaux, and David Chiang.
Pretrained self-supervised speech models can recognize unseen consonants.
In
Interspeech, 1453–1458. 2026.
doi:10.21437/Interspeech.2026-2848.
DOI
BibTeX
@inproceedings{taguchi26_interspeech,
author = "Taguchi, Chihiro and {Le Ferrand}, Éric and Nakagawa, Hirosi and Ono, Hitomi and Kato, Kanji and Prud'hommeaux, Emily and Chiang, David",
title = "Pretrained self-supervised speech models can recognize unseen consonants",
year = "2026",
booktitle = "Interspeech",
pages = "1453--1458",
doi = "10.21437/Interspeech.2026-2848",
issn = "2958-1796"
}
Aarohi Srivastava and David Chiang.
Different perturbations, different mechanisms: understanding continued pre-training for zero-shot dialect robustness.
2026.
arXiv:2608.05510.
PDF
BibTeX
@misc{srivastava-chiang-2026,
author = "Srivastava, Aarohi and Chiang, David",
title = "Different Perturbations, Different Mechanisms: Understanding Continued Pre-training for Zero-Shot Dialect Robustness",
year = "2026",
note = "arXiv:2608.05510",
url = "https://arxiv.org/abs/2608.05510"
}
Andy Yang, Pascal Bergstr
äßer, Georg Zetzsche, David Chiang, and Anthony W. Lin.
Length generalization bounds for transformers.
In
Proc. ICML. 2026.
PDF
BibTeX
@inproceedings{yang-etal-2026-length,
author = {Yang, Andy and Bergstr{\"a\ss}er, Pascal and Zetzsche, Georg and Chiang, David and Lin, Anthony W.},
title = "Length Generalization Bounds for Transformers",
booktitle = "Proc. ICML",
year = "2026",
url = "https://proceedings.mlr.press/v306/yang26c.html"
}
Chihiro Taguchi, Yukinori Takubo, and David Chiang.
Automatic speech recognition for documenting endangered languages: case study of
Ikema
Miyakoan.
In
Proc. LREC, 3547–3555. 2026.
BibTeX
@inproceedings{taguchi-etal-2026-automatic,
author = "Taguchi, Chihiro and Takubo, Yukinori and Chiang, David",
title = "Automatic Speech Recognition for Documenting Endangered Languages: Case Study of {I}kema {M}iyakoan",
booktitle = "Proc. LREC",
year = "2026",
pages = "3547--3555"
}
Stephen Bothwell, Kaitlin Stephan, Hildegund M
üller, and David Chiang.
From pagin
ā to webpage: on developing and documenting a digitized
Latin collection.
Journal of Open Humanities Data, 2026.
doi:10.5334/johd.397.
DOI
BibTeX
@article{bothwell-etal-2026-pagina,
author = {Bothwell, Stephen and Stephan, Kaitlin and M{\"u}ller, Hildegund and Chiang, David},
title = "From Pagin{\=a} to Webpage: On Developing and Documenting a Digitized {L}atin Collection",
journal = "Journal of Open Humanities Data",
year = "2026",
volume = "11",
article = "61",
doi = "10.5334/johd.397"
}
Akriti Dhasmana, Aarohi Srivastava, and David Chiang.
Dialect matters: cross-lingual
ASR transfer for low-resource
Indic language varieties.
In
Proc. Workshop on NLP for Similar Languages, Varieties and Dialects. 2026.
PDF
BibTeX
@inproceedings{dhasmana-etal-2026-dialect,
author = "Dhasmana, Akriti and Srivastava, Aarohi and Chiang, David",
title = "Dialect Matters: Cross-Lingual {ASR} Transfer for Low-Resource {I}ndic Language Varieties",
url = "https://aclanthology.org/2026.vardial-1.12/",
booktitle = "Proc. Workshop on NLP for Similar Languages, Varieties and Dialects",
year = "2026"
}
Andy Yang, Anej Svete, Jiaoda Li, Anthony Widjaja Lin, Jonathan Rawski, Ryan Cotterell, and David Chiang.
Probability distributions computed by autoregressive transformers.
In
Proc. ICLR. 2026.
PDF
BibTeX
@inproceedings{yang-etal-2026-probability,
author = "Yang, Andy and Svete, Anej and Li, Jiaoda and Lin, Anthony Widjaja and Rawski, Jonathan and Cotterell, Ryan and Chiang, David",
title = "Probability Distributions Computed by Autoregressive Transformers",
year = "2026",
booktitle = "Proc. ICLR",
url = "https://openreview.net/forum?id=gZIcyx1tQY"
}
Andy Yang, Christopher Watson, Anton Xue, Satwik Bhattamishra, Jose Llarena, William Merrill, Emile Dos Santos Ferreira, Anej Svete, and David Chiang.
The transformer cookbook.
Transactions on Machine Learning Research, January 2026.
PDF
BibTeX
@article{yang-etal-2025-cookbook,
author = "Yang, Andy and Watson, Christopher and Xue, Anton and Bhattamishra, Satwik and Llarena, Jose and Merrill, William and Dos Santos Ferreira, Emile and Svete, Anej and Chiang, David",
title = "The Transformer Cookbook",
journal = "Transactions on Machine Learning Research",
month = "January",
year = "2026",
url = "https://openreview.net/forum?id=sPshCSvDrX"
}
Katsumi Ibaraki and David Chiang.
Frustratingly easy data augmentation for low-resource
ASR.
2025.
arXiv:2509.15373.
PDF
BibTeX
@misc{ibaraki-chiang-2025-frustratingly,
author = "Ibaraki, Katsumi and Chiang, David",
note = "{arXiv}:2509.15373",
year = "2025",
title = "Frustratingly Easy Data Augmentation for Low-Resource {ASR}",
url = "https://arxiv.org/abs/2509.15373"
}
Chihiro Taguchi, Seng Mai, Keita Kurabe, Yusuke Sakai, Georgina Agyei, Soudabeh Eslami, and David Chiang.
Languages still left behind: toward a better multilingual machine translation benchmark.
In
Proc. EMNLP, 20142–20154. 2025.
doi:10.18653/v1/2025.emnlp-main.1018.
PDF
BibTeX
@inproceedings{taguchi-etal-2025-languages,
author = "Taguchi, Chihiro and Mai, Seng and Kurabe, Keita and Sakai, Yusuke and Agyei, Georgina and Eslami, Soudabeh and Chiang, David",
title = "Languages Still Left Behind: Toward a Better Multilingual Machine Translation Benchmark",
year = "2025",
booktitle = "Proc. EMNLP",
url = "https://aclanthology.org/2025.emnlp-main.1018/",
doi = "10.18653/v1/2025.emnlp-main.1018",
pages = "20142--20154"
}
Andy Yang, Micha
ël Cadilhac, and David Chiang.
Knee-deep in
C-RASP: a transformer depth hierarchy.
In
Proc. NeurIPS 38. 2025.
doi:10.52202/085713-1249.
DOI
BibTeX
@inproceedings{yang+:2025,
author = {Yang, Andy and Cadilhac, Micha{\"e}l and Chiang, David},
title = "Knee-Deep in {C-RASP}: A Transformer Depth Hierarchy",
booktitle = "Proc. NeurIPS 38",
proc_url = "https://papers.nips.cc/paper\_files/paper/2025/hash/354a6e7be4f1a35127746b8147a31bfe-Abstract-Conference.html",
doi = "10.52202/085713-1249",
year = "2025"
}
Aarohi Srivastava and David Chiang.
We're calling an intervention: exploring fundamental hurdles in adapting language models to nonstandard text.
In
Proc. Workshop on Noisy and User-Generated Text. 2025.
Best Paper Award.
PDF
BibTeX
@inproceedings{srivastava-chiang-2025,
author = "Srivastava, Aarohi and Chiang, David",
title = "We're Calling an Intervention: Exploring Fundamental Hurdles in Adapting Language Models to Nonstandard Text",
booktitle = "Proc. Workshop on Noisy and User-Generated Text",
year = "2025",
url = "https://arxiv.org/abs/2404.07304"
}
Chihiro Taguchi and David Chiang.
Language complexity and speech recognition accuracy: orthographic complexity hurts, phonological complexity doesn't.
In
Proc. ACL. 2024.
Outstanding Paper Award and Senior Area Chair Award.
PDF
BibTeX
@inproceedings{taguchi-chiang-2024-complexity,
author = "Taguchi, Chihiro and Chiang, David",
title = "Language Complexity and Speech Recognition Accuracy: Orthographic Complexity Hurts, Phonological Complexity Doesn't",
year = "2024",
url = "https://aclanthology.org/2024.acl-long.827/",
booktitle = "Proc. ACL"
}
Fahim Faisal, Orevaoghene Ahia, Aarohi Srivastava, Kabir Ahuja, David Chiang, Yulia Tsvetkov, and Antonios Anastasopoulos.
DIALECTBENCH: a
NLP benchmark for dialects, varieties, and closely-related languages.
In
Proc. ACL. 2024.
Social Impact Award.
PDF
BibTeX
@inproceedings{faisal+:2024,
author = "Faisal, Fahim and Ahia, Orevaoghene and Srivastava, Aarohi and Ahuja, Kabir and Chiang, David and Tsvetkov, Yulia and Anastasopoulos, Antonios",
title = "{DIALECTBENCH}: A {NLP} Benchmark for Dialects, Varieties, and Closely-Related Languages",
year = "2024",
booktitle = "Proc. ACL",
url = "https://aclanthology.org/2024.acl-long.777/"
}
full list