# Dipankar Srirag

> Ph.D. candidate at UNSW Sydney working on NLP for linguistically diverse nurse-patient communication and clinical decision support in the emergency department, with a background in dialect robustness.

I am a Ph.D. candidate at the [UNSW-NLP group](https://unswnlp.github.io/), where I am fortunate to be advised by [Dr. Aditya Joshi](https://www.unsw.edu.au/staff/aditya-joshi/). My Ph.D. deals with developing language technologies and resources to support linguistically diverse nurse-patient communication and safely support clinical decisions in the emergency department.

I also hold a master's degree from the [University of New South Wales](https://www.unsw.edu.au/), where my thesis dealt with evaluating and improving the robustness of language models to dialects of English. Beyond my research, I mentor undergraduate students through the [Taste of Research](https://www.unsw.edu.au/engineering/student-life/undergraduate-research-opportunities/taste-research) program at UNSW, and manage the [Natural Language Processing course](https://www.unsw.edu.au/course-outlines/course-outline#year=2026&term=Term%201&deliveryMode=In%20Person&deliveryFormat=Standard&teachingPeriod=T1&deliveryLocation=Kensington&courseCode=COMP6713&activityGroupId=1).

- Role: Ph.D. Candidate, UNSW Sydney
- Research interests: Natural Language Processing, Clinical NLP, Emergency Department Triage, Large Language Models, Nurse-Patient Communication, Clinical Decision Support, Dialect Robustness, Varieties of English, Speech and Dialogue Systems
- Website: https://dipankarsrirag.github.io/
- CV: https://dipankarsrirag.github.io/cv.pdf
- Email: d.srirag@unsw.edu.au
- Scholar: https://scholar.google.com/citations?user=8dUrhxwAAAAJ
- GitHub: https://github.com/dipankarsrirag
- Twitter: https://x.com/dipankarsrirag
- Bluesky: https://bsky.app/profile/dipankarsrirag.bsky.social

## Publications

- [TriageSim: A Conversational Emergency Triage Simulation Framework from Structured Electronic Health Records](https://www.isca-archive.org/interspeech_2026/srirag26_interspeech.pdf): Dipankar Srirag, Quoc Dung Nguyen, Aditya Joshi, Padmanesan Narasimhan, Salil Kanhere. Interspeech 2026.
- [LLMs Anchor on Chief Complaint and Fail to Integrate Evidence in Sequential Clinical Triage](https://arxiv.org/pdf/2609.22904): Dipankar Srirag, Haokai Zhao, Ashutosh Kumar, Eleanor Hopper, Michael Dalton, Quoc Dung Nguyen, Aditya Joshi, Salil S. Kanhere, Padmanesan Narasimhan. Preprint, 2026.
- [DiaLLM: An Investigation into the Robustness-Generation Gap in English Dialect Adaptation](https://arxiv.org/abs/2607.07669): Jordan Painter, Dipankar Srirag, Adarsh Kappiyath, Diptesh Kanojia, Aditya Joshi, Lu Yin. To appear at EMNLP 2026 (Main) (preprint).
- [Retell, Reward, Repeat: Reinforcement Learning for Narrative Theory-Informed Story Retelling](https://arxiv.org/pdf/2601.17226): David Y. Liu, Xanthe Muston, Dipankar Srirag, Aditya Joshi, Sebastian Sequoiah-Grayson. INLG 2026.
- [PAREDA: A Multi-Accent Speech Dataset of Natural Language Processing Research Discussions](https://aclanthology.org/2026.speakable-1.2.pdf): Sicheng Jin, Dipankar Srirag, Aditya Joshi. SPEAKABLE @ LREC 2026.
- [From Triage to Discharge: A Survey of NLP Tasks, Methods, and Open Challenges in the Emergency Department](https://arxiv.org/abs/2608.23627): Dipankar Srirag, Aditya Joshi, Padmanesan Narasimhan, Salil Kanhere. To appear at EMNLP 2026 (Main) (preprint).
- [A Taxonomy-Driven Case Study of Australian Web Resources Against Technology-Facilitated Abuse](https://link.springer.com/content/pdf/10.1007/978-981-92-3015-0.pdf): Dipankar Srirag, Xiaolin Cen, Rahat Masood, Aditya Joshi. ACISP 2026.
- [Far Out: Evaluating Language Models on Slang in Australian and Indian English](https://aclanthology.org/2026.vardial-1.2.pdf): Deniz Kaya Dilsiz, Dipankar Srirag, Aditya Joshi. VarDial @ EACL 2026.
- [Nek Minit: Harnessing Pragmatic Metacognitive Prompting for Explainable Sarcasm Detection of Australian and Indian English](https://arxiv.org/pdf/2505.15095): Ishmanbir Singh, Dipankar Srirag, Aditya Joshi. ALTA 2025.
- [BESSTIE: A Benchmark for Sentiment and Sarcasm Classification for Varieties of English](https://aclanthology.org/2025.findings-acl.441.pdf): Dipankar Srirag, Aditya Joshi, Jordan Painter, Diptesh Kanojia. Findings of ACL: ACL 2025.
- [Predicting the Target Word of Game-playing Conversations using a Low-Rank Dialect Adapter for Decoder Models](https://aclanthology.org/2025.naacl-short.2.pdf): Dipankar Srirag, Aditya Joshi, Jacob Eisenstein. NAACL 2025.
- [Evaluating Dialect Robustness of Language Models via Conversation Understanding](https://aclanthology.org/2025.sumeval-2.3.pdf): Dipankar Srirag, Nihar Ranjan Sahoo, Aditya Joshi. SUMEval @ COLING 2025.

## Teaching

- COMP9444 Neural Networks and Deep Learning (UNSW, 26T3): Tutor
- COMP9444 Neural Networks and Deep Learning (UNSW, 26T2): Tutor, student feedback 100%
- COMP9024 Data Structures and Algorithms (UNSW, 26T2): Tutor, student feedback 100%
- COMP6713 Natural Language Processing (UNSW, 26T1): Course Admin
- IEP-48230 Introduction to Engineering Projects (UTS, 26S1): Tutor
- COMP9444 Neural Networks and Deep Learning (UNSW, 25T2): Tutor, student feedback 100%
- COMP9024 Data Structures and Algorithms (UNSW, 25T3): Tutor, student feedback 100%
- COMP9444 Neural Networks and Deep Learning (UNSW, 24T2, 24T3): Tutor, student feedback 97%

## Recent news

- Oct 2026: TriageSim is now published at Interspeech 2026.
- Oct 2026: Two papers accepted to the EMNLP 2026 main conference: our survey From Triage to Discharge, and DiaLLM, which investigates the robustness-generation gap in English dialect adaptation.
- Oct 2026: Retell, Reward, Repeat, on reinforcement learning for narrative theory-informed story retelling, published at INLG 2026.
- Sep 2026: Preprint: LLMs Anchor on Chief Complaint and Fail to Integrate Evidence in Sequential Clinical Triage is on arXiv.
- May 2026: PAREDA, a multi-accent speech dataset of NLP research discussions, published at SPEAKABLE @ LREC 2026.
- Apr 2026: Preprint: From Triage to Discharge: A Survey of NLP Tasks, Methods, and Open Challenges in the Emergency Department is now available.
- Apr 2026: A Taxonomy-Driven Case Study of Australian Web Resources Against Technology-Facilitated Abuse published in the ACISP 2026 proceedings.
- Mar 2026: TriageSim, a conversational emergency triage simulation framework built from electronic health records, released as a Python package with a project site.
- Mar 2026: Far Out, which evaluates LMs on slang in Australian and Indian English, accepted to VarDial @ EACL 2026.
- Nov 2025: Preprint: A Taxonomy-Driven Case Study of Australian Web Resources Against Technology-Facilitated Abuse is on arXiv.

## Pages
- [Research](https://dipankarsrirag.github.io/publications/): full publication list with BibTeX
- [Teaching](https://dipankarsrirag.github.io/courses/)
- [News](https://dipankarsrirag.github.io/news/)