Nazia Tasnim

Here's me being Pochita at the Newbury Comics!
📮 nimzia [at] bu [dot] edu

Hello there!! I am a CS PhD student at Boston University, advised by Prof. Bryan Plummer on topics related to multimodal learning and parameter-efficient finetuning. My research interest lies at the crosssection of Natural Language Processing and Computer Vision, with a special emphasis on responsible applications and under-resourced domains.

I was fortunate to be advised by Dr. Isaac Johnson and Dr. Martin Gerlach at the Wikimedia Research Team, where I built resources and pipelines for NLP research across the Wikimedia projects.
I have a year of industry experience as a Machine Learning Engineer at Giga Tech Ltd., where I worked on developing the National Syntactic TreeBank for Bangla. I am also a Research Collaborator at Bengali.ai, where I work on building resources for low-resource languages.
I have published in multiple A* conferences and workshops, including Interspeech, NeuRIPS, EMNLP, and CVPR. A full list of my publications can be found here.
I also build different open-source tools and datasets to support the larger community. Some of my projects can be found here.

Outside research and courseworks, I am interested in plants🪴, pottery 🏺 and pop-culture 🖖. Shoot me an email if you want to collaborate on a project, have a question, or just want to say hi~

Imagine a world where AI doesn't just crunch numbers but speaks every language, adapts on a dime, and plays fair with everyone. 

news

Sep 5, 2023	Officially started my PhD at Boston University ✨
Aug 1, 2023	Received Dean’s Fellowship from the Boston University 🤩🤩 !!
Jun 1, 2023	Released a new package!! 🎉 Install from PyPI!

selected publications

RECAST: Reparameterize Compact weight Adaptation for Sequential Tasks,

Under Review, 2024

Bib

Vision-LLMs Can Fool Themselves with Self-Generated Typographic Attacks

NeuRIPS, 2024

Bib PDF

OOD-Speech: A Large Bengali Speech Recognition Dataset for Out-of-Distribution Benchmarking

INTERSPEECH, 2023

Bib PDF

@article{rakib2023oodspeech,
  title = {OOD-Speech: A Large Bengali Speech Recognition Dataset for Out-of-Distribution Benchmarking},
  journal = {INTERSPEECH},
  year = {2023},
  doi = {10.48550/arXiv.2305.09688},
}

Vista: vision transformer enhanced by u-net and image colorfulness frame filtration for automatic retail checkout

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2022

Bib PDF

@article{10.1109/cvprw56347.2022.00359,
  title = {Vista: vision transformer enhanced by u-net and image colorfulness frame filtration for automatic retail checkout},
  journal = {2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)},
  year = {2022},
  doi = {10.1109/cvprw56347.2022.00359},
}

On leveraging data augmentation and ensemble to recognize complex named entities in bangla

Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022), 2022

Bib PDF

@article{10.18653/v1/2022.semeval-1.209,
  title = {On leveraging data augmentation and ensemble to recognize complex named entities in bangla},
  journal = {Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)},
  year = {2022},
  doi = {10.18653/v1/2022.semeval-1.209},
  dimensions = {false},
}

Choice of assemblers has a critical impact on de novo assembly of sars-cov-2 genome and characterizing variants

Briefings in Bioinformatics, 2021

Bib PDF

@article{10.1093/bib/bbab102,
  title = {Choice of assemblers has a critical impact on de novo assembly of sars-cov-2 genome and characterizing variants},
  journal = {Briefings in Bioinformatics},
  year = {2021},
  volume = {22},
  issue = {5},
  doi = {10.1093/bib/bbab102},
}