News

Amrita Singh has been selected as a DAAD AInet Fellow. Congratulations, Amrita! 🎉

Five papers published at EMNLP 2026. In the main conference: a benchmark framework for specialist domains by Bao Gia Doan, DiaLLM by Jordan Painter, and From Triage to Discharge by Dipankar Srirag. In Findings: IndicKLAR by Debajyoti Mazumder and SAPS by Sharath Naganna.

Our workshop on Community-Based and Socially Aware Approaches to Natural Language Processing, has been accepted at COLING 2026. The co-organisers are Lia Draetta, Chiara Ferrando, Adriana S. Pagano and Viviana Patti from University of Turin (Lia, Chiara and Viviana) and Federal University of Minas Gerais (Adriana).

Two papers published at INLG 2026: Retell, Reward, Repeat by David Liu, on reinforcement learning for narrative theory-informed story retelling, and In Vino Veritas and Vulnerabilities by Anudeex Shetty, on LLM safety under “drunk” language.

Our tutorial, Emergency Department: A Real-World Edge Case for Clinical NLP, has been accepted at AJCAI 2026. It runs on 27 November 2026.

Our research on getting AI “drunk” to expose new security risks in chatbots was featured on ABC News, news.com.au and the UNSW Newsroom . Find out more about the paper here.

Co-organising the ALTA 2026 Shared Task on sentiment and sarcasm classification for Australian and British English, built on our BESSTIE dataset.

ResPL, on reasoning-based pseudo-labels for sarcasm detection in low-resource Arabic dialects by Ali Almutairi, accepted to ICNLSP 2026.

Two papers accepted at Interspeech 2026: stuttering detection in paediatric speech and TriageSim.

Suhan Karaca, co-supervised by Amrita Singh and Aditya Joshi, published a paper at Customizable NLP workshop @ ACL 2026

Amrita’s system paper on Arabic medical text classification accepted to AbjadNLP @ EACL 2026.

Far Out, evaluating language models on slang in Australian and Indian English, accepted to VarDial @ EACL 2026.

PAREDA, a multi-accent speech dataset of NLP research discussions by Sicheng (Stefan) Jin, accepted to the SPEAKABLE workshop @ LREC 2026.

TriageSim, a conversational emergency triage simulation framework from electronic health records, released as a Python package.

Charles’s paper on evaluating LLMs for masked sentence prediction accepted to IJCNLP-AACL 2025.

Nek Minit received the Best Paper Honorable Mention at ALTA 2025, co-organised by Aditya.

Amrita’s survey on classification tasks for legal contracts accepted to Artificial Intelligence Review.

BESSTIE, a benchmark for sentiment and sarcasm across varieties of English, accepted to Findings of ACL 2025. Dataset on HuggingFace.

RACCOON, retrieval-augmented geocoding for news articles, accepted to TheWebConf 2025.

Dipankar presented our work on dialect adaptation for game-playing conversations at NAACL 2025 in Albuquerque.

Aditya’s survey on NLP for dialects of a language published in ACM Computing Surveys.

Dipankar’s paper on evaluating dialect robustness of language models accepted to SUMEval @ COLING 2025.

Aditya presented a tutorial on NLP for lower-resource scenarios at COLING 2025 in Abu Dhabi.

Fai Lu’s paper on multilingual LGBTQIA+ hate speech detection accepted to ALTA 2024.

Aditya’s paper on balancing classical and deep learning in NLP pedagogy accepted to TeachingNLP @ ACL 2024.

BAMBINO-LM, bilingual human-inspired continual pre-training of BabyLM, accepted to CMCL @ ACL 2024.

Duke’s Spectraformer, a unified random feature framework for transformers, accepted to ACM TIST.

Duke’s paper on transformer-based ensemble for AI-generated text detection accepted to ALTA 2023.

Aditya’s paper on evaluating LLMs using an Indian language LGBTQ+ lexicon accepted to AI & Ethics.