Pratinav Seth

Pratinav Seth

Machine Learning Research Enthusiast

Publications

RSM-NLP at BLP-2023 Task 2: Bangla Sentiment Analysis using Weighted and Majority Voted Fine-Tuned Transformers

Published in Accepted at Proceedings of the 1st Workshop on Bangla Language Processing (BLP 2023), EMNLP, 2023

we benchmarked various pre-trained Transformer based models such as BERT, RoBERTa and DeBERTaV3 and their majority voting Ensemble Models. This detection and classification of online sexism were done under Task 10 at SemEval 2023-Explainable Detection of Online Sexism (EDOS). We also used Focal loss to deal with class imbalance.

Recommended citation: Pratinav Seth, Rashi Goel, Komal Mathur and Swetha Vemulapalli. RSM-NLP at BLP-2023 Task 2: Bangla Sentiment Analysis using Weighted and Majority Voted Fine-Tuned Transformers. Conditionally Accepted at Proceedings of the 1st Workshop on Bangla Language Processing (BLP 2023), EMNLP, Association for Computational Linguistics. https://scholar.google.com/citations?user=DwBn1fcAAAAJ&hl=en&authuser=1

Refuseg: Regularized multi-modal fusion for precise brain tumour segmentation

Published in Accepted at 9th Brain Lesion (BrainLes) workshop - the satellite event of the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) 2023, 2023

Our proposed method can help address the challenges posed by artifacts in medical imagery due to data acquisition errors (such as patient motion) or a reconstruction algorithm’s inability to represent the anatomy while ensuring a trade-off in accuracy. Our proposed regularization module makes it robust to these scenarios and ensures the reliability of lesion segmentation.

Recommended citation: Aditya Kasliwal, Sankarshanaa Sagaram, Laven Srivastava, Pratinav Seth, and Adil Khan. Refuseg: Regularized multi-modal fusion for precise brain tumour segmentation. Accepted at 9th Brain Lesion (BrainLes) workshop - the satellite event of the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) 2023. URL: https://aps.arxiv.org/pdf/2308.13883.pdf https://aps.arxiv.org/pdf/2308.13883.pdf

SSS at SemEval-2023 task 10: Explainable detection of online sexism using majority voted fine-tuned transformers

Published in Accepted at 17th International Workshop on Semantic Evaluation (SemEval-2023), Association for Computational Linguistics, 2023

we benchmarked various pre-trained Transformer based models such as BERT, RoBERTa and DeBERTaV3 and their majority voting Ensemble Models. This detection and classification of online sexism were done under Task 10 at SemEval 2023-Explainable Detection of Online Sexism (EDOS). We also used Focal loss to deal with class imbalance.

Recommended citation: Sriya Rallabandi, Sanchit Singhal, and Pratinav Seth. SSS at SemEval-2023 task 10: Explainable detection of online sexism using majority voted fine-tuned transformers. In Proceedings of the 17th International Workshop on Semantic Evaluation (SemEval-2023), pages 1231–1236, Toronto, Canada, July 2023. Association for Computational Linguistics. URL: https://aclanthology.org/2023.semeval-1.171, doi:10.18653/v1/2023.semeval-1.171 https://aclanthology.org/2023.semeval-1.171

Corefusion: Contrastive regularized fusion for guided thermal super-resolution

Published in Accepted at 19th IEEE Workshop on Perception Beyond the Visible Spectrum,CVPR 2023, 2023

We present a novel data fusion framework and regularization technique for Guided Super Resolution of Thermal images. The proposed architecture is computationally in-expensive and lightweight with the ability to maintain performance despite missing one of the modalities, i.e., high-resolution RGB image or the lower-resolution thermal image, and is designed to be robust in the presence of missing data. The proposed method presents a promising solution to the frequently occurring problem of missing modalities in a real-world scenario.

Recommended citation: Aditya Kasliwal, Pratinav Seth, Sriya Rallabandi, and Sanchit Singhal. Corefusion: Contrastive regularized fusion for guided thermal super-resolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop (CVPRW), pages 507–514, 2023. URL: https://ieeexplore.ieee.org/document/10208919 https://ieeexplore.ieee.org/document/10208919

Uncertainty-aware test-time augmented ensemble of berts for classification of common mental illnesses on social media posts

Published in The First Tiny Papers Track at ICLR 2023, Tiny Papers @ ICLR 2023, Kigali, Rwanda, 2023

Uncertainty-Aware Test-Time Augmented Ensembling of BERTs for producing reliable and well-calibrated predictions to classify six possible types of mental illnesses- None, Depression, Anxiety, Bipolar Disorder, ADHD, and PTSD by analyzing unstructured user data on Reddit.

Recommended citation: Pratinav Seth(∗) and Mihir Agarwal(∗) . Uncertainty-aware test-time augmented ensemble of berts for classification of common mental illnesses on social media posts. In Krystal Maughan, Rosanne Liu, and Thomas F. Burns, editors, (* - equal contribution) The First Tiny Papers Track at ICLR 2023, Tiny Papers @ ICLR 2023, Kigali, Rwanda, May 5, 2023, 2023. URL: https://openreview.net/pdf?id=a9VgV-hywP https://arxiv.org/pdf/2304.04539.pdf

Published in , 1900

Uncertainty Aware Test Time Augmented Ensemble for PIRC Diabetic Retinopathy Detection

Published in Accepted at the Medical Imagery Meets NeurIPS Workshop,NeurIPS 2022, 2022

Proposed a methodology UATTA-ENS to produce reliable and well-reliable calibrated predictions for 5 class Diabetic Retinopathy classifications.

Recommended citation: Pratinav Seth, Adil Hamid Khan(∗), Ananya Gupta(∗), Saurabh Mishra(∗), and Akshat Bhandhari. Uatta-ens: Uncertainty aware test time augmented ensemble for pirc diabetic retinopathy detection. (* - equal contribution) Medical Imagery Meets NeurIPS Workshop,NeurIPS 2022, 2022. URL: http://www.cse.cuhk.edu.hk/∼qdou/public/medneurips2022/95.pdf https://arxiv.org/abs/2211.03148

Performance Evaluation of Deep Segmentation Models on Landsat-8 Imagery

Published in Accepted at Tackling Climate Change with Machine Learning Workshop,NeurIPS 2022, 2022

We benchmark several state-of-the-art segmentation techniques to detect contrails in low-orbit satellite imagery.

Recommended citation: Akshat Bhandari(∗), Sriya Rallabandi(∗), Sanchit Singhal(∗), Aditya Kasliwal(∗), and Pratinav Seth. Performance evaluation of deep segmentation models on landsat-8 imagery. (* - equal contribution) Tackling Climate Change with Machine Learning Workshop, NeurIPS 2022, 2022. URL: https://www.climatechange.ai/papers/neurips2022/92 https://arxiv.org/abs/2211.14851