FADE Research Group

223, CSE Department,

Indian Institute of Technology,

Jodhpur, Rajasthan, 342030

The Fairness And Data Efficiency (FADE) Research Group at the Indian Institute of Technology Jodhpur, led by Dr Pratik Mazumder, focuses on developing equitable AI systems and advancing deep learning techniques for data-constrained environments, with applications in bias mitigation and medical image processing.

Note: We are looking for motivated PhD, B.Tech., and M.Tech. students with at least some experience in Deep Learning.

news

Mar 11, 2026	Paper Accepted at IEEE Guwahati Subsection Conference 2026: Auxiliary Dense Feature Alignment for Medical Incremental Learning (AuxDenseFeatAlign), Arnab Roy*, Shubham Bagwari, Pratik Mazumder, IEEE Guwahati Subsection Conference 2026
Mar 05, 2026	Paper Accepted at ICLR Workshop: Semantic-Anchored, Class Variance-Optimized Clustering for Robust Semi-Supervised Few-Shot Learning, Souvik Maji* , Rhythm Baghel, Pratik Mazumder, International Conference on Learning Representations 2026, Geometry-grounded Representation Learning and Generative Modeling Workshop.
Feb 15, 2026	Paper Accepted at KBS 2026: Debiased Incremental Learning (DIL): A Novel Framework for Mitigating Bias in Incremental Learning, Shubham Bagwari, Pratik Mazumder, Knowledge-Based Systems (KBS), 2026.

Selected Publications

Debiased incremental learning (DIL): A novel framework for mitigating bias in incremental learning

Shubham Bagwari^* and Pratik Mazumder^*

Knowledge-Based Systems, 2026

Bib

@article{2026_KBS_bagwari_dil,
  title = {Debiased incremental learning (DIL): A novel framework for mitigating bias in incremental learning},
  author = {Bagwari, Shubham and Mazumder, Pratik},
  journal = {Knowledge-Based Systems},
  year = {2026},
  publisher = {Elsevier}
}

Semantic-Anchored, Class Variance-Optimized Clustering for Robust Semi-Supervised Few-Shot Learning

Souvik Maji^*, Rhythm Baghel^*, and Pratik Mazumder^*

In Geometry-grounded Representation Learning and Generative Modeling Workshop, International Conference on Learning Representations (ICLRw), 2026

Bib

@inproceedings{2026_iclrw_maji_semantic,
  title = {Semantic-Anchored, Class Variance-Optimized Clustering for Robust Semi-Supervised Few-Shot Learning},
  author = {Maji, Souvik and Baghel, Rhythm and Mazumder, Pratik},
  booktitle = {Geometry-grounded Representation Learning and Generative Modeling Workshop, International Conference on Learning Representations (ICLRw)},
  year = {2026}
}

Leveraging Joint Incremental Learning Objective with Data Ensemble for Class Incremental Learning

Pratik Mazumder, Mohammed Asad Karim, Indu Joshi, and 1 more author

Neural Networks, 2024

Bib

@article{2024_NN_mazumder_leveraging,
  title = {Leveraging Joint Incremental Learning Objective with Data Ensemble for Class Incremental Learning},
  author = {Mazumder, Pratik and Karim, Mohammed Asad and Joshi, Indu and Singh, Pravendra},
  journal = {Neural Networks},
  year = {2024},
  publisher = {Elsevier}
}

Rectification-based Knowledge Retention for Task Incremental Learning

Pratik Mazumder, Pravendra Singh, Piyush Rai, and 1 more author

IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

Bib

@article{2024_TPAMI_mazumder_rectification,
  title = {Rectification-based Knowledge Retention for Task Incremental Learning},
  author = {Mazumder, Pratik and Singh, Pravendra and Rai, Piyush and Namboodiri, Vinay P},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2024},
  publisher = {IEEE}
}

Protected Attribute Guided Representation Learning for Bias Mitigation in Limited Data

Pratik Mazumder and Pravendra Singh

Knowledge-Based Systems, 2022

Bib

@article{2022_KBS_mazumder_protected,
  title = {Protected Attribute Guided Representation Learning for Bias Mitigation in Limited Data},
  author = {Mazumder, Pratik and Singh, Pravendra},
  journal = {Knowledge-Based Systems},
  volume = {238},
  pages = {107926},
  year = {2022},
  publisher = {Elsevier}
}

Hybrid Sample Synthesis-based Debiasing of Classifier in Limited Data Setting

Piyush Arora and Pratik Mazumder

In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024

Bib

@inproceedings{2024_wacv_arora_hybrid,
  title = {Hybrid Sample Synthesis-based Debiasing of Classifier in Limited Data Setting},
  author = {Arora, Piyush and Mazumder, Pratik},
  booktitle = {IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
  year = {2024}
}

Few-Shot Lifelong Learning

Pratik Mazumder, Pravendra Singh, and Piyush Rai

In AAAI Conference on Artificial Intelligence (AAAI), 2021

Bib

@inproceedings{2021_aaai_mazumder_few,
  title = {Few-Shot Lifelong Learning},
  author = {Mazumder, Pratik and Singh, Pravendra and Rai, Piyush},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2021}
}

Calibrating cnns for lifelong learning

Pravendra Singh, Vinay Kumar Verma, Pratik Mazumder, and 2 more authors

In git Advances in Neural Information Processing Systems, 2020

Bib

@inproceedings{2020_neurips_singh_calibrating,
  title = {Calibrating cnns for lifelong learning},
  author = {Singh, Pravendra and Verma, Vinay Kumar and Mazumder, Pratik and Carin, Lawrence and Rai, Piyush},
  booktitle = {git Advances in Neural Information Processing Systems},
  pages = {13702--13713},
  year = {2020}
}