AI-NLP-ML Group, IIT Patna

AI-NLP-ML Group, IIT Patna The group is dedicated to exploring the frontiers of Artificial Intelligence, Machine Learning and Natural Language Processing under the able guidance of Prof.

The Artificial Intelligence-Natural Language Processing-Machine Learning (AI-NLP-ML) group at Department of Computer Science and Engineering, IIT Patna has started its official journey in June 2015. Pushpak Bhattacharyya. The group also consists of other two faculty members, Dr. Asif Ekbal and Dr. Sriparna Saha, and around 30 members including research scholars, research engineers, lexicographers,

B.Tech & M.Tech students. Several industry sponsored projects are currently being undertaken. Elsevier, the renowned scientific literature publishing company has set up the Elsevier Centre of Excellence for Natural Language Processing to conduct research and development in some of the novel areas of AI, NLP, and ML. Another company, ezDI has set up the Sushrut-eZDI Research Lab on Health Informatics, which is dedicated to developing products for healthcare and contribute to research having significant outcomes.

We are delighted to announce that IIT Patna is organizing IEEE BuildSEC’25(https://cgi.cse.unsw.edu.au/~buildsec/)📝 Impo...
31/10/2025

We are delighted to announce that IIT Patna is organizing IEEE BuildSEC’25(https://cgi.cse.unsw.edu.au/~buildsec/)

📝 Important Dates
▪️ Main Conference: 1-3 December 2025
▪️ Workshop & Tutorials: 1 December 2025
📍 Venue: IIT Patna, Bihar, India

Key Highlights:
• The conference goes beyond technical discussions to foster broader dialogue.
• It provides a unique platform to collectively address today’s most critical cybersecurity and privacy challenges.
• It explores how cutting-edge technologies can be leveraged to enhance security and resilience.
• It emphasizes the importance of user studies in safeguarding citizens in cyberspace.

22/10/2025

From our group, we recently wrote three surveys, focusing on the LLM: Reasoning strategies in LLM; LLM for Low-resource Machine Translation and LLM beyond Sentence Level MT.

1. Dibyanayan Bandyopadhyay, Soham Bhattacharjee, Asif Ekbal (2025). Thinking Machines: A Survey of LLM based Reasoning Strategies, Link: https://arxiv.org/abs/2503.10814

2. Baban Gain, Dibyanayan Bandyopadhyay and Asif Ekbal (2025). Bridging the Linguistic Divide: A Survey on Leveraging Large Language Models for Machine Translation, Link: https://arxiv.org/abs/2504.01919v3

3. Ramakrishna Appicharla, Baban Gain, Santanu Pal, Asif Ekbal (2025). Beyond the Sentence: A Survey on Context-Aware Machine Translation with Large Language Models, Link: https://arxiv.org/abs/2506.07583

Dibyanayan Bandyopadhyay, Soham Bhattacharjee, Baban Gain, Dr. Santanu Pal, Asif Ekbal

We deeply mourn the passing of our beloved mentor and guide, Prof. Pushpak Bhattacharyya.His vision, kindness, and pione...
11/10/2025

We deeply mourn the passing of our beloved mentor and guide, Prof. Pushpak Bhattacharyya.
His vision, kindness, and pioneering work in NLP have shaped countless lives.

An alumnus of AI-NLP-ML Lab, Dr. Tirthankar Ghosal, shares a heartfelt tribute:

Prof. Pushpak Bhattacharyya তুমি রবে নীরবে হৃদয়ে মম ... (You will remain silently in my heart ... always!) - Rabindranath Tagore Dear Sir, I will be indebted to you for life. Even as I write these words, I cannot reconcile myself to the finality of them. You ...

October 5, 2025, will forever remain a dark day for us. We lost not only a leader but the very foundation upon which thi...
08/10/2025

October 5, 2025, will forever remain a dark day for us. We lost not only a leader but the very foundation upon which this lab was built — Prof. Pushpak Bhattacharyya

Since the inception of the AI-NLP-ML Lab in June 2015, Prof. Pushpak’s vision has been the central force behind its establishment. With fellow torchbearers Dr. Asif Ekbal and Dr. Sriparna Saha - nurtured under his guidance and vision-he stood as the lab’s strongest pillar, shaping it with his principles, unmatched expertise, and foresight. What began as a dream a decade ago now stands as a lab of national and international repute, a testament to his unwavering guidance that continues to inspire the lab today.

To speak of his contributions feels almost inadequate. His teaching proficiency was simply unmatched — a gift that touched every student fortunate enough to sit in his classroom. His research vision was crystal clear, uncompromising in integrity, and always ahead of its time. At a moment when academia often gets lost in the race of publishing for numbers, he reminded us what it means to do research the right way. His contribution to the NLP community made him the President of ACL, the highest international governing body of computational linguistics and NLP.

But what we will miss most goes beyond the technical. By watching him, we absorbed values that will forever stay with us — his work ethic, his integrity, his preparedness, his courage to stand firm in the toughest situations, and his never back down spirit. These were not lessons he taught explicitly; they were lessons we absorbed by watching him live them every single day.

He believed that a truly educated individual is one who is not only knowledgeable but also healthy in mind and body, capable of contributing meaningfully to society. All-round development of a student involves nurturing every aspect of their personality — intellectual, physical, emotional, social, and moral. He used to motivate us to take care of physical health, and do regular exercise and sports in building stamina, discipline, and teamwork. We remember how he himself used to participate with us in cricket, and badminton.

Today, we are shaken by his untimely departure. The sense of loss is immeasurable. Yet in this grief, we also feel deeply blessed that over the years, knowingly or unknowingly, he helped us learn to become our own pillars. That quiet strength he passed on to us will now carry his legacy forward. Today, as per CS core ranking, our lab ranks in the top-3 in India for NLP research.

Prof. Pushpak’s life was a gift — to his students, to this lab, and to the broader world of NLP. His vision, kindness, and unwavering principles will continue to guide us, in both academia and industry.

Rest in peace, Sir . You will forever remain in our hearts and in the soul of the AI-NLP-ML Lab.

“जैसे पुष्पक विमान अनगिनत रत्नों से अलंकृत था, वैसे ही प्रो. पुष्पक अपने ज्ञान, सिद्धांतों और मूल्यों से हमें समृद्ध कर गए।”

The Artificial Intelligence–Natural Language Processing–Machine Learning (AI-NLP-ML) Research Group at IIT Patna (https:...
30/08/2025

The Artificial Intelligence–Natural Language Processing–Machine Learning (AI-NLP-ML) Research Group at IIT Patna (https://lnkd.in/d7e7XDrh) invites applications for multiple positions under a prestigious R&D project titled "Centre for Indian Language Data (COIL-D)" funded by the flagship mission, Bhashini, MeitY, Government of India.

Applications are invited for the post of Office Assistant in the prescribed format only for the following assignments in a purely time-bound research project undertaken at this institute. Please apply early and carefully, as there are a limited number of seats.

ADVERTISEMENT NO: R&D/1155/CIL/511, DATED: 26.08.2025
Name of the Posts:
Office Assistant – 01 Post

For more details, please find the attached advertisement with this post.

The Artificial Intelligence–Natural Language Processing–Machine Learning (AI-NLP-ML) Research Group at IIT Patna (https:...
30/08/2025

The Artificial Intelligence–Natural Language Processing–Machine Learning (AI-NLP-ML) Research Group at IIT Patna (https://lnkd.in/d7e7XDrh) invites applications for multiple positions under a prestigious R&D project titled "Centre for Indian Language Data (COIL-D)" funded by the flagship mission, Bhashini, MeitY, Government of India.

We are looking for bright and motivated candidates who are passionate about Artificial Intelligence, Natural Language Processing, and Machine Learning for Technical positions.
Applications are invited in the prescribed format only for the following assignments in a purely time-bound research project undertaken at this institute. Please apply early and carefully, as there are a limited number of seats.

ADVERTISEMENT NO: R&D/1155/CIL/512, DATED: 26.08.2025
Name of the Posts:
Senior Research Associate (Technical) – 01 Post

Junior Research Associate (Technical) – 02 Post

Junior Research Associate (Odia Language) – 01 Post

For more details, please find the attached advertisement with this post.

27/08/2025

From our research group, we have made the code and dataset available for our papers accepted at the Conference of the North American Chapter of the Association for Computational Linguistics (NAACL 2025)! 🎉

Paper 1: MixRevDetect – Towards Detecting AI-Generated Content in Hybrid Peer Reviews
Authors: S. Kumar, S. Garg, S. Sengupta, T. Ghosal, and A. Ekbal
Venue: NAACL 2025 (Main Conference)

Summary:
MixRevDetect introduces a novel method to detect AI-generated content within peer reviews, even when reviews contain both human- and AI-written segments. Unlike earlier methods that treat reviews as a single block, MixRevDetect pinpoints fine-grained AI-generated points, ensuring the integrity of academic evaluations. It significantly outperforms existing detection tools with an F1 score of 88.86%.

Paper Link :https://aclanthology.org/2025.naacl-short.79.pdf
Code and Dataset : https://github.com/sandeep82945/AI-text-Points

Paper 2: MENDER – Multi-hop Commonsense and Domain-specific CoT Reasoning for Knowledge-grounded Empathetic Counseling of Crime Victims
Authors: A. Hossain, P. Priya, A. M. Tripathi, P. Verma, and A. Ekbal
Venue: NAACL 2025 (Student Research Workshop)

Summary:
MENDER is a framework for knowledge-grounded empathetic counseling of crime victims. It leverages multi-hop reasoning with LLMs by combining commonsense and domain-specific knowledge through dual reasoning chains (Commonsense & Domain Knowledge-driven CoT). On the POEM dataset, MENDER produces coherent and empathetic responses effectively.

Paper Link :https://aclanthology.org/2025.naacl-srw.49.pdf
Code and Dataset : https://github.com/Abid839/MENDER/tree/main

Call for Papers – IEEE BuildSEC’25We are delighted to announce that IIT Patna is organizing IEEE BuildSEC’25, to be held...
22/08/2025

Call for Papers – IEEE BuildSEC’25

We are delighted to announce that IIT Patna is organizing IEEE BuildSEC’25, to be held from 1st–3rd December 2025.

📝 Important Dates
▪️ Paper Registration Deadline: 25th August 2025
▪️ Paper Submission Link: https://lnkd.in/g8yjJSvS
📍 Venue: IIT Patna, Bihar, India

Key Highlights:
• The conference goes beyond technical discussions to foster broader dialogue.
• It provides a unique platform to collectively address today’s most critical cybersecurity and privacy challenges.
• It explores how cutting-edge technologies can be leveraged to enhance security and resilience.
• It emphasizes the importance of user studies in safeguarding citizens in cyberspace.

We invite researchers, practitioners, and industry experts to submit their research papers and join us in this exciting conference!

The AI-NLP-ML Research Group at IIT Patna (https://lnkd.in/d7e7XDrh) invites applications for multiple positions under a...
22/08/2025

The AI-NLP-ML Research Group at IIT Patna (https://lnkd.in/d7e7XDrh) invites applications for multiple positions under a prestigious R&D project titled
"VIDYAAPATI: A Bidirectional Machine Translation System involving Bengali, Konkani, Maithili, and Hindi," funded by the flagship mission, Bhashini, MeitY, Government of India.

We are looking for bright and motivated candidates who are passionate about Artificial Intelligence, Natural Language Processing, and Machine Learning. Applications are invited in the prescribed format only. Please apply early and carefully, as there are a limited number of seats.

ADVERTISEMENT NO: R&D/690/MeitY/507, DATED: 12-08-2025
Name of the Posts:
Senior Research Fellow (Technical) – 02 Posts
Junior Research Fellow (Technical) – 03 Posts

For more details, please find the attached advertisement with this post

22/08/2025

From our group, following papers have been accepted in EMNLP 2025 :

S. Kumar, T. Ghosal, V. Goyal, & A. Ekbal (2025). Can Large Language Models Unlock Novel Scientific Research Ideas? In EMNLP 2025 (Main)

Summary: This study examines the ability of Large Language Models (LLMs) to generate future research ideas from scientific papers. Four LLMs—Claude-2, GPT-4, GPT-3.5, and Gemini—were evaluated across five domains (Chemistry, Computer Science, Economics, Medicine, and Physics). Results show that Claude-2 and GPT-4 produce ideas more aligned with authors’ perspectives, with Claude-2 generating the most diverse ideas. A human evaluation further assessed novelty, relevance, and feasibility, providing insights into both the strengths and limitations of LLMs in research idea generation. The datasets and code are made publicly available to support further exploration.

P. Priya, S. Dudhate, D V Yasheshbhai & A. Ekbal (2025). We Argue to Agree: Towards Personality-Driven Argumentation-Based Negotiation Dialogue Systems for Tourism. In EMNLP 2025 (Findings)

Summary: This work introduces the Personality-driven Argumentation-based Negotiation Dialogue Generation (PAN-DG) task and presents PACT, a dataset of personality-driven negotiation conversations in the tourism sector. Generated using LLMs, PACT incorporates three personality profiles—Argumentation, Preference, and Buying Style—to simulate diverse negotiation scenarios. Automatic and human evaluations confirm the dataset’s quality, and experiments show that fine-tuned LLMs outperform pre-trained ones in generating rational, personality-aligned negotiation responses. These findings highlight PACT’s potential to advance personalization and reasoning in negotiation dialogue systems.

We are inviting applications for rolling vacancies for Language Experts under the COIL-D Project (Funded by MeitY) at II...
08/08/2025

We are inviting applications for rolling vacancies for Language Experts under the COIL-D Project (Funded by MeitY) at IIT Patna for the following language pairs:

1. Hindi → Maithili
2. Hindi → Odia
3. Hindi → Santhali (Ol Chiki script)
4. Hindi → Bengali

Selected candidates will work remotely as freelancers and will be handsomely compensated.

If you're interested, please apply through the form below:

This link will take you to a page that’s not on LinkedIn

16/07/2025

The following paper from our group have been accepted at ACM Transactions on Asian and Low-Resource Language Information Processing (TALLIP) :

B. Gain, D. Bandyopadhyay, S. Mukherjee, C. Adak, & A. Ekbal. (2025). Impact of Visual Context on Noisy Multimodal NMT: An Empirical Study for English to Indian Languages, ACM Trans. Asian Low-Resour. Lang. Inf. Process. Just Accepted (July 2025). https://doi.org/10.1145/3748311

Summary: This paper investigates the effect of visual context in Multimodal Neural Machine Translation (MMT) from English to Hindi, Bengali, and Malayalam, particularly under noisy textual conditions. It finds that while images contribute little in clean, high-resource settings, their utility grows in noisy scenarios—cropped image features help in low-noise cases, and full images perform better in high-noise settings. Surprisingly, even random images offer comparable improvements, suggesting that visual input primarily serves as a form of regularization rather than providing semantically meaningful aid. The study emphasizes that current MMT models do not leverage image information effectively and calls for better datasets and architectures to unlock the true potential of multimodal translation

Paper available at : https://dl.acm.org/doi/abs/10.1145/3748311 (Open Access)

Address

AI-NLP-ML Lab Room Number: 510, 5th Floor, Block III, Department Of Computer Science And Engineering Indian Institute Of Technology Patna
Bihta
801103

Telephone

+91-612-3028338

Website

Alerts

Be the first to know and let us send you an email when AI-NLP-ML Group, IIT Patna posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Share