Rohit Agarwal

I am Rohit Agarwal, a 4th-Year Ph.D. student at the Department of Computer Science, UiT Tromsø, Norway. I work on building scalable AI architectures for complex biological systems. I work in the Bio-AI and 3DNanoscopy Lab under the supervision of Dilip K. Prasad, Alexander Horsch and Krishna Agarwal.

I am interested in making dynamic deep learning models where the number of inputs to a model keeps changing with streaming data. I call them Haphazard Inputs. Along with these, I have experience in computer vision and natural language processing. Recently, I am also exploring large language models (LLMs) with a focus on continual learning and quantization.

Selected Publications

Haphazard Inputs

Aux-Drop: Handling Haphazard Inputs in Online Learning Using Auxiliary Dropouts

Rohit Agarwal, Deepak Gupta, Alexander Horsch , and 1 more author

Transactions on Machine Learning Research, 2023

Reproducibility Certification

Bib HTML PDF Code

@article{agarwal2023auxdrop,
  title = {Aux-Drop: Handling Haphazard Inputs in Online Learning Using Auxiliary Dropouts},
  author = {Agarwal, Rohit and Gupta, Deepak and Horsch, Alexander and Prasad, Dilip K.},
  journal = {Transactions on Machine Learning Research},
  issn = {2835-8856},
  year = {2023},
  url = {https://openreview.net/forum?id=R9CgBkeZ6Z},
  note = {Reproducibility Certification},
}

Vision

Mabnet: Master Assistant Buddy Network With Hybrid Learning for Image Retrieval

Rohit Agarwal, Gyanendra Das, Saksham Aggarwal , and 2 more authors

In ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2023

Bib HTML Video Code

@inproceedings{agarwal2023mabnet,
  title = {Mabnet: Master Assistant Buddy Network With Hybrid Learning for Image Retrieval},
  author = {Agarwal, Rohit and Das, Gyanendra and Aggarwal, Saksham and Horsch, Alexander and Prasad, Dilip K},
  booktitle = {ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  pages = {1--5},
  year = {2023},
  organization = {IEEE},
  doi = {10.1109/ICASSP49357.2023.10094987},
}