Publications and Presentations

You can also find my articles on my Google Scholar profile.

Publications

LASER: An LLM-based ASR Scoring and Evaluation Rubric

Published at EMNLP 2025 (Main conference), 2025

LASER is an LLM-based ASR evaluation metric that aligns closely with human judgments by capturing linguistic nuances across Indian languages better than traditional metrics.

Recommended citation: Amruta Parulekar and Preethi Jyothi. 2025. LASER: An LLM-based ASR Scoring and Evaluation Rubric. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 24773–24782, Suzhou, China. Association for Computational Linguistics.
Download Paper | Download Slides

AMPS: ASR with Multimodal Paraphrase Supervision

Published at NAACL 2025 (Main conference), 2025

Integrated paraphrase supervision in a multimodal pipeline to improve Automatic Speech Recognition for spontaneous and disfluent speech.

Recommended citation: Abhishek Gupta, Amruta Parulekar, Sameep Chattopadhyay, and Preethi Jyothi. 2025. AMPS: ASR with Multimodal Paraphrase Supervision. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers), pages 404–413, Albuquerque, New Mexico. Association for Computational Linguistics.
Download Paper | Download Slides

PathoGen-X: A Cross-Modal Genomic Feature Trans-Align Network for Enhanced Survival Prediction from Histopathology Images

Published at IEEE ISBI 2025 (International Symposium on Biomedical Imaging), 2025

Developed PathoGen-X, a transformer-based framework that translates histopathology image features into the genomic feature space for improved survival prediction without requiring paired genomic data at testing.

Recommended citation: A. Krishna, N. C. Kurian, A. Patil, A. Parulekar, P. J. P and A. Sethi, "Pathogen-X: A Cross-Modal Genomic Feature Trans-Align Network for Enhanced Survival Prediction from Histopathology Images," 2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI), Houston, TX, USA, 2025, pp. 1-4, doi: 10.1109/ISBI60581.2025.10981028.
Download Paper | Download Slides

Parameter-efficient Adaptation of Multilingual Multimodal Models for Low-resource ASR

Published at the 4th Multilingual Representation Learning Workshop, EMNLP 2024, 2024

Combining computationally efficient techniques like speech-based parameter-efficient finetuning and text-only adaptation to improve automatic speech recognition of low resource languages using multimodal multilingual models.

Recommended citation: Abhishek Gupta, Amruta Parulekar, Sameep Chattopadhyay, and Preethi Jyothi. 2024. Parameter-efficient Adaptation of Multilingual Multimodal Models for Low-resource ASR. In Proceedings of the Fourth Workshop on Multilingual Representation Learning (MRL 2024), pages 175–185, Miami, Florida, USA. Association for Computational Linguistics.
Download Paper | Download Slides

Combining Datasets with Different Label Sets for Improved Nucleus Segmentation and Classification

Published and presented in Bioimaging (BIOSTEC), Rome, 2024

A novel loss function and training technique that can be integrated with a multitude of architectures, for consolidating class labels of different nuclei segmentation and classification datasets.

Recommended citation: Parulekar A., Kanwat U., Gupta R., Chippa M., Jacob T., Bameta T., Rane S. and Sethi A. (2024). Combining Datasets with Different Label Sets for Improved Nucleus Segmentation and Classification. In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 1: BIOIMAGING; ISBN 978-989-758-688-0, SciTePress, pages 281-288. DOI: 10.5220/0012380800003657
Download Paper | Download Slides

Towards improving breast cancer detection through multi-modal image generation

Published in the Ultrasonics journal, 2023

Interconversion of CT scans and ultrasounds using wave interference patterns, GANs and fourier domain adaptation.

Recommended citation: Almahfouz Nasser S, Sharma A, Saraf A, Parulekar A, Haria P, Sethi A. Towards improving breast cancer detection through multi-modal image generation. Ultrasonics. 2025 Sep;153:107655. doi: 10.1016/j.ultras.2025.107655. Epub 2025 Apr 15. PMID: 40262439.
Download Paper

Presentations

A Computer Vision Pipeline for Laryngoscopic Image Standardization through Histogram Matching

Poster presented in the 145th American Laryngological Association meet (COSM), Chicago, 2024

A preprocessing pipeline for larynogoscopic videos that includes removal of unusable frames, illumination correction, specularity removal and finally color transfer to a target intensity distribution.

Recommended citation: Parulekar A., Wiercigroch J., Kahrs L. A. and Lin R. J. (2024). A Computer Vision Pipeline for Laryngoscopic Image Standardization through Histogram Matching.
Download Slides

Artificial Intelligence-based Eosinophil Count in Gastrointestinal Tract Biopsy

Poster presented in the American Gatroenterology Association meet (DDW), Chicago and published in the Gastroenterology journal, 2023

An eosinophila detection model conquering severe class imbalance built using UNet architecture.

Recommended citation: Shah H.C., Amarpurkar A.D., Jacob T., Parulekar A.M. and Sethi A. (2023). EP178 ARTIFICIAL INTELLIGENCE BASED EOSINOPHIL COUNT IN GASTROINTESTINAL TRACT BIOPSY. Gastroenterology, 164(6), pp.S-1229.
Download Paper | Download Slides