Tushar Prakash

I am a ML researcher at Sony Research India, working on personalization and optimization. Previously, I worked on domain adaptation with the STARS Research Team at Inria Centre at Université Côte d’Azur. I also interned at the University of Alberta, where I worked on adversarial robustness. I have spent a "bit" of my time at Microsoft as SDE intern, developing testbeds for 5G infrastructure. As an undergraduate student, I studied Electronics and Communication Engineering at the Delhi Technological University (India).

For any inquiries, feel free to reach out to me via mail!

CV Mail Twitter Scholar Github LinkedIn

Awards: “In 2018, I graduated at the top of my class from high school. I completed my undergraduate studies in 2023, graduating with First Division with Distinction. In 2022, I was selected for the Mitacs Global Research Internship (GRI) at the University of Alberta. In 2023, I was honored with the Best Paper Award for CR-SoRec at Sony’s annual conference for my contributions to research.”

Publications

MVBN: Enhancing Social Recommendation with Multi-View BERT Network
Tushar Prakash, Raksha Jalan, Naoyuki Onoe
In Proceedings of the 23th IEEE International Conference on Data Mining (ICDM), 2023
Project Page / Paper /

@InProceedings{tushar2023icdm, 
	author = {Tushar Prakash and Raksha Jalan and Naoyuki Onoe}, 
	title = {MVBN: Enhancing Social Recommendation with Multi-View BERT Network}, 
	booktitle = {In Proceedings of the 23th IEEE International Conference on Data Mining (ICDM)}, 
	year = {2023}, 
}

CR-SoRec: BERT driven Consistency Regularization for Social Recommendation
Tushar Prakash*, Raksha Jalan*, Brijraj Singh, Naoyuki Onoe
In Proceedings of the 17th ACM Conference on Recommender Systems (RecSys), 2023
Project Page / Paper /

@InProceedings{tushar2023recsys, 
	author = {Tushar Prakash and Raksha Jalan and Brijraj Singh and Naoyuki Onoe}, 
	title = {CR-SoRec: BERT driven Consistency Regularization for Social Recommendation}, 
	booktitle = {In Proceedings of the 17th ACM Conference on Recommender Systems (RecSys)}, 
	year = {2023}, 
}

LLM-BRec: Personalizing Session-based Social Recommendation with LLM-BERT Fusion Framework
Raksha Jalan*, Tushar Prakash*, Niranjan Pedanekar
In Proceedings of Gen-IR Workshop, 47th ACM SIGIR Conference on Research and Development in Information Retrieval, 2024
Project Page / Paper /

@InProceedings{jalan2024sigir, 
	author = {Raksha Jalan and Tushar Prakash and Niranjan Pedanekar}, 
	title = {LLM-BRec: Personalizing Session-based Social Recommendation with LLM-BERT Fusion Framework}, 
	booktitle = {In Proceedings of Gen-IR Workshop, 47th ACM SIGIR Conference on Research and Development in Information Retrieval}, 
	year = {2024}, 
}

Dynamic Task-adaptive Meta Optimization for User Cold-Start Recommendation
Raksha Jalan*, Tushar Prakash*, Niranjan Pedanekar
In Submission at 47th European Conference on Information Retrieval (ECIR), 2024
Project Page / Paper /

@InProceedings{jalan2024ecir, 
	author = {Raksha Jalan and Tushar Prakash and Niranjan Pedanekar}, 
	title = {Dynamic Task-adaptive Meta Optimization for User Cold-Start Recommendation}, 
	booktitle = {In Submission at 47th European Conference on Information Retrieval (ECIR)}, 
	year = {2024}, 
}

Meta Ordinal Regression with Dynamic Heterogeneous Graphs and Normalized Ratings for User Cold-Start Scenario
Raksha Jalan*, Tushar Prakash*, Niranjan Pedanekar
In Submission at 47th European Conference on Information Retrieval (ECIR), 2024
Project Page / Paper /

@InProceedings{jalan2024ecirs, 
	author = {Raksha Jalan and Tushar Prakash and Niranjan Pedanekar}, 
	title = {Meta Ordinal Regression with Dynamic Heterogeneous Graphs and Normalized Ratings for User Cold-Start Scenario}, 
	booktitle = {In Submission at 47th European Conference on Information Retrieval (ECIR)}, 
	year = {2024}, 
}

Unsupervised domain alignment of fingerprint denoising models using pseudo annotations
Indu Joshi, Tushar Prakash, Rohit Kumar, Antitza Dantcheva, Sumantra Roy, Prem Kalra
In Springer's Multimedia Tools and Applications, vol. 83, pp. 38 167–38 192, 2023
Project Page / Paper /

@InProceedings{Joshi2024mta, 
	author = {Indu Joshi and Tushar Prakash and Rohit Kumar and Antitza Dantcheva and Sumantra Roy and Prem Kalra}, 
	title = {Unsupervised domain alignment of fingerprint denoising models using pseudo annotations}, 
	booktitle = {In Springer's Multimedia Tools and Applications, vol. 83, pp. 38 167–38 192}, 
	year = {2023}, 
}

Context-Aware Restoration of Noisy Fingerprints
Indu Joshi, Tushar Prakash, Rohit Kumar, B. Jaiswal, Antitza Dantcheva, Sumantra Roy, Prem Kalra
In IEEE Sensors Letters, vol. 6, pp. 101–104, 2022
Project Page / Paper /

@InProceedings{Joshi2024ieee, 
	author = {Indu Joshi and Tushar Prakash and Rohit Kumar and B. Jaiswal and Antitza Dantcheva and Sumantra Roy and Prem Kalra}, 
	title = {Context-Aware Restoration of Noisy Fingerprints}, 
	booktitle = {In IEEE Sensors Letters, vol. 6, pp. 101–104}, 
	year = {2022}, 
}

Leveraging Explainable Artificial Intelligence for Understanding the Effect of Model Capacity on Training Dataset Size
Tushar Prakash, Tashvik Dhamija, Rohit Kumar, Jeebananda Panda
In Proceedings of the 17th IEEE International Conference on Service Operations and Logistics, and Informatics (SOLI), 2022
Project Page / Paper /

@InProceedings{tushar2024soli, 
	author = {Tushar Prakash and Tashvik Dhamija and Rohit Kumar and Jeebananda Panda}, 
	title = {Leveraging Explainable Artificial Intelligence for Understanding the Effect of Model Capacity on Training Dataset Size}, 
	booktitle = {In Proceedings of the 17th IEEE International Conference on Service Operations and Logistics, and Informatics (SOLI)}, 
	year = {2022}, 
}

Homepage Template

Feel free to use this website as a template! It is fully responsive and very easy to use and maintain as it uses a python script that crawls your bib files to automatically add the papers and talks. If you find it helpful, please add a link to my website - I will also add a link to yours (if you want). Checkout the github repository for instructions on how to use it.
⚛ ⚛ ⚛