Avatar

Punyajoy Saha

PhD student

IIT Kharagpur

Biography

Punyajoy Saha is a PMRF research scholar in the Department of Computer Science and Engineering at IIT Kharagpur, West Bengal. Currently, he is doing research under the supervision of Prof. Animesh Mukherjee. He is also a member of the research group CNeRG.

Research - His current research interests lie in the intersection of computational social science and natural language processing. He is working in developing better human-in-the-loop mitigation strategies for hate speech and other forms of harmful speech like fear speech. The collection of his efforts as well as his colleagues towards the goal of creating a more inclusive and accepting online environment is present at this website

Reviewer assignment - AAAI 2021, ACL 2020, ACL-IJCNLP 2021, EACL 2021, AAAI 2022, AAAI 2023

Teaching Assistantship - AI and Ethics,Spring 2021 (CS60016), Algorithms-1,Autumn 2021 (CS21003), AI and Ethics,Spring 2022 (CS60016), Information retrieval 2023 (CS60092)

Outside Teaching Assitantship - As a part of PMRF deliverables, I am also a teaching assistant for the NPTEL-NLP course, videos are uploaded weekly here and taking a course on “Programming in C” in Hijli College, Kharagpur (materials to be updated soon).

Interests

  • Natural Language Processing
  • Social Computing
  • Hate speech analytics
  • Counter speech generation

Education

  • PhD in Computer Science and Engineering, 2020 - Present

    Indian Institute of Kharagpur,West Bengal

  • B.Tech in Computer Science and Technology, 2015-19

    Indian Institute of Engineering Science Technology, Shibpur

  • CBSE XII Science, 2014-15

    Kendriya Vidyalaya R.K Puram Sec-4 New Delhi

News

Recent updates

Paper on Fear speech in Indian public Whatsapp groups to be presented at ARCS 2022
Invited talk in Understanding and Automating Counterspeech workshop organised by CRASSH
Our paper hateful user detection paper has won the highly presitgious Ted Nelson Newcomer Award at ACM Hypertext 2021.
1st in both subtasks of Abusive and Threatening Language Detection for Tweets in Urdu: Abusive Language Detection at FIRE 2021
Paper on Fear speech in Indian public Whatsapp groups to be presented at IC2S2 2021
Tutorial on 'Hate speech- Detection, Mitigation and beyond' accepted at ICWSM 2021
1st, 2nd and 1st in offensive language detection in Tamil, Kannada and Malayalam respectively in Dravidian workshop at EACL2021
Paper on Fear speech in Indian public Whatsapp groups is accepted in WWW 2021
Paper on Explainable hate speech dataset accepted at AAAI 2021
World rank 11th in Hatememe detection competition organised by Facebook
Paper on temporal effects of hate speech got accepted in CSCW 2020
Hate alert covered by Huggingface CEO in twitter
Paper on multilingual hate speech detection accepted at ECML PKDD 2020
Winner of Facebook and Social Science One data grant
Joined PhD in January 2020
Misogyny detection recognised by MIT Technical Review

Recent & Upcoming Talks

Echoes of Fear: Unraveling the Presence of Fear Speech in Social Media Platforms @CyberVSR

This talk explores the impact of harmful speech, particularly fear speech, on social media platforms such as Gab and Whatsapp in the US …
Aug 2, 2023 7:00 PM — 8:00 PM
Slides

Hate speech: Detection, Mitigation and Beyond @WSDM

In this translation style tutorial presented at WSDM 2023, we present an exposition of hate speech detection and mitigation and also …
Mar 27, 2023 8:30 AM — 11:30 AM
Project Slides Follow

Enabling moderation of harmful content in online social media platforms@CLAWS group

Social media has become an integral part of our everyday life. It allows the rapid dissemination of information and opinions. However, …
Jan 19, 2023 8:30 PM — 9:30 PM
Project Slides

Generating counter speech with CounterGeDi@CredCo

Recently, many studies have tried to create generation models to assist counter speakers by providing counterspeech suggestions for …
May 12, 2022 1:00 PM — 2:00 PM
Slides

Hate speech: Detection, Mitigation and Beyond @AAAI

In this translation style tutorial presented at AAAI 2022, we present an exposition of hate speech detection and mitigation and also …
Feb 23, 2022 10:00 PM — 11:30 PM
Project Slides Follow
See all talks

Publications

List of all publications till date

Mithun Das, Saurabh Kumar Pandey, Shivansh Sethi, Punyajoy Saha, Animesh Mukherjee (2024). Low-Resource Counterspeech Generation for Indic Languages: The Case of Bengali and Hindi. EACL 2024.

Preprint PDF

Sarthak Roy, Ashish Harshavardhan, Animesh Mukherjee, Punyajoy Saha (2023). Probing LLMs for hate speech detection: strengths and vulnerabilities. EMNLP 2023.

Preprint PDF Source Document

Punyajoy Saha, Divyanshu Sheth, Kushal Kedia, Binny Mathew, Animesh Mukherjee (2023). Rationale-Guided Few-Shot Classification to Detect Abusive Language. ECAI 2023.

Preprint Code

Punyajoy Saha, Kiran Garimella, Narla Komal Kalyana, Saurabh Kumar Pandeya, Pauras Mangesh Meher, Binny Mathew, Animesh Mukherjee (2023). On the rise of fear speech in online social media. PNAS.

Preprint PDF Code

Mithun Das, Rohit Raj, Punyajoy Saha, Binny Mathew, Manish Gupta, Animesh Mukherjee (2023). HateMM: A Multi-Modal Dataset for Hate Video Classification. ICWSM 2023.

Preprint PDF Code

Piush Aggarwal, Pranit Chawla, Mithun Das, Punyajoy Saha, Binny Mathew, Torsten Zesch, Animesh Mukherjee (2023). HateProof: Are Hateful Meme Detection Systems really Robust?. WWW 2023.

Preprint PDF Source Document

Vikram Gupta, Sumegh Roychowdhury, Mithun Das, Somnath Banerjee, Punyajoy Saha, Binny Mathew, Hastagiri Prakash Vanchinathan, Animesh Mukherjee (2022). Multilingual Abusive Comment Detection at Scale for Indic Languages. NeurIPS 2022 Datasets and Benchmarks.

Preprint Code

Punyajoy Saha, Kanishk Singh, Adarsh Kumar, Binny Mathew, Animesh Mukherjee (2022). CounterGeDi: A controllable approach to generate polite, detoxified and emotional counterspeech. IJCAI 2022.

Preprint Code

Mithun Das, Punyajoy Saha, Binny Mathew, Animesh Mukherjee (2022). HateCheckHIn: Evaluating Hindi Hate Speech Detection Models. LREC 2022.

Preprint Code

Mithun Das, Somnath Banerjee, Punyajoy Saha (2021). Abusive and Threatening Language Detection in Urdu using Boosting based and BERT based models: A Comparative Approach. FIRE’21.

Code Source Document

Somnath Banerjee, Mithun Das, Punyajoy Saha (2021). Exploring Transformer Based Models to Identify Hate Speech and Offensive Content in English and Indo-Aryan Languages. FIRE’21.

Source Document

Mithun Das, Punyajoy Saha, Ritam Dutt, Pawan Goyal, Animesh Mukherjee, Binny Mathew (2021). You too Brutus! Trapping Hateful Users in Social Media: Challenges, Solutions & Insights. In ACMHT-2021.

Preprint PDF Code Dataset Slides

Debjoy Saha, Naman Paharia, Debajit Chakraborty, Punyajoy Saha, Animesh Mukherjee (2021). Hate-Alert@ DravidianLangTech-EACL2021: Ensembling strategies for Transformer-based Offensive language Detection. DravidianLangTech-EACL2021.

PDF Code Dataset Source Document

Punyajoy Saha, Binny Mathew, Kiran Garimella, Animesh Mukherjee (2021). Short is the Road that Leads from Fear to Hate: Fear Speech in Indian WhatsApp Groups. WWW 2021.

Preprint PDF Code Dataset Slides Video

Poojitha Maheshappa, Binny Mathew, Punyajoy Saha (2021). Using Knowledge Graphs to Improve Hate Speech Detection. CODS COMAD 2021.

PDF Source Document DOI

Binny Mathew, Punyajoy Saha, Seid Muhie Yimam, Chris Biemann, Pawan Goyal, Animesh Mukherjee (2020). HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection. AAAI 2021.

Preprint PDF Code Dataset Slides

Binny Mathew, Anurag Illendula, Punyajoy Saha, Soumya Sarkar, Pawan Goyal, Animesh Mukherjee (2020). Hate begets Hate: A Temporal Study of Hate Speech. CSCW 2020.

Preprint PDF Code

Sai Saketh Aluru, Binny Mathews, Punyajoy Saha, Animesh Mukherjee (2020). Deep Learning Models for Multilingual Hate Speech Detection. ECML-PKDD 2020.

Preprint PDF Code Dataset Slides

Punyajoy Saha, Binny Mathews, Animesh Mukherjee, Pawan Goyal (2019). HateMonitors: Language Agnostic Abuse Detection in Social Media. CEUR.

Preprint PDF Code Dataset

Binny Mathews, Punyajoy Saha, Hardik Tharad, Subham Rajgaria, Prajwal Singhania, Suman Kalyan Maity, Animesh Mukherjee, Pawan Goyal (2019). Thou shalt not hate: Countering Online Hate Speech. In ICWSM-2019.

PDF Code Dataset Source Document

Punyajoy Saha, Binny Mathews, Animesh Mukherjee, Pawan Goyal (2018). Hateminers: Detecting Hatespeech against Women. Arxiv.

Preprint PDF Code Dataset

Chetan Ralekar, Punyajoy Saha, Tapan K. Gandhi, Santanu Chaudhury (2018). Effect of Devanagari Font Type in Reading Comprehension: An Eye Tracking Study. In IHCI-2019.

PDF

Sandipan Sarma, Punyajoy Saha, Jaya Sil (2018). Document Categorization Using Graph Structuring. In ICCACP-2018.

PDF Code

Experience

Intern and Job Experiences

 
 
 
 
 

Research Scholar

CNeRG Lab, CSE Department, IIT Kharagpur

Jan 2020 – Present Kharagpur, West Bengal

Broad Research Area - NLP and Social Computing.

Current Research Area - Detection, analysis and countering of hate speech

Courses done during PhD - Natural language processing, Deep learning, High performance computing, Reinforcement learning, Social computing, AI and Ethics

 
 
 
 
 

Research Intern

Language Technology group, University of Hamburg

Aug 2019 – Sep 2019 Hamburg, Germany
worked on developing the backend for Forum 4.0 project which a site for new comments analysis
 
 
 
 
 

Junior Research Fellow

CNERG Labs, IIT Kharagpur

May 2019 – Dec 2019 Kharagpur
Worked in the project Cognitive Stimuli: Detecting and Generating Exaggeration in Online and Social Media Content (CSAM), funded by Adobe Systems Inc
 
 
 
 
 

Research-Intern

CNERG Labs, IIT Kharagpur

May 2018 – Apr 2019 Kharagpur

Work included detection of hate speech in online social media.

Achievements:

  • Part of the winning team “Hateminers” at AMI Evalita-2018.
  • Member of the project which is releasing the first ever dataset for analysis of counterspeech at ICWSM-2019
 
 
 
 
 

Software-Intern

Bluezeal Labs, IIT Bombay

Dec 2017 – Mar 2018 Shibpur, Howrah
Part of the artificial intelligence team of this start-up. Responsibilities include:

  • Developing optical character recognition
  • Developing web scraping and chatbot solutions.
 
 
 
 
 

Research-Intern

Neurocomputing Lab ,IIT Delhi

May 2017 – Jul 2018 New Delhi, Delhi

Work included analysis of Hindi Letters using eye tracking and neural networks.

Achievements include:

  • Part of the project “Analysis of hindi fonts on reading comprehension” , later a paper on this was selected at IHCI-2018. Major contribution include error reduction of eye tracker using machine learning

Contact

  • +919968982501
  • CNeRG Lab, CSE Department, IIT Kharagpur, Kharagpur, West Bengal 721302
  • Enter CSE Dept and take the stairs to Room 205 on 1st Floor
  • Weekdays from 10:00 am to 5:00 pm
  • Book an appointment
  • DM Me