🎓 Milestone - PhD Defended 🎓 I’m incredibly grateful to share that I successfully defended my PhD in Machine Learning at Luleå University of Technology on April 8th, 2025. 🧠 Thesis Title: Towards Robust and Domain-Aware Self-Supervised Representation Learning

I, Gautam Vashishtha, and Jithamanyu Settur are thrilled to share that our paper, "ASTrA: Adversarial Self-supervised Training with Adaptive-Attacks," has been accepted at #ICLR2025! *Equal contributions*

Presenting our work LCM: Log Conformal Maps for Robust Representation Learning to Mitigate Perspective Distortions at #ACCV2024

Thank you @STARSNetwork46

🚀 Incredibly excited to have been part of ECCV 2024 in Milan 🇮🇹 with friends across the globe!

Stay tuned for the release of source code, trained models, and benchmark soon! Looking forward to sharing more at #ECCV24 in #Milan and engaging with other fantastic works!

📊 Public Benchmark: ImageNet-PD, a dedicated benchmark for evaluating model robustness against perspective distortions.

🌐 Real-world Applications: Enhanced performance in multiple computer vision applications such as crowd counting, person re-identification, and fisheye image recognition.

🚀 Outcome: Improved model robustness in both supervised and self-supervised learning approaches against multiple benchmarks, working well with various architectures, including CNNs and transformers.

🔧 Proposed Method: Möbius transform based MPD method synthesizes perspective distortions with parameter controls, without needing camera parameters.

About Our Work: 📌 Problem: We show that the effect of perspective distortion is real and it causes unprecedented distribution shifts which degrades the robustness of learned representations.

I'm thrilled that our work titled "Möbius Transform for Mitigating Perspective Distortions in Representation Learning" has been accepted at @eccvconf. A huge thank you to my co-authors and mentors. An extra-mile to Prof. @ucfmshah for his invaluable guidance. #ECCV2024

Exciting times at #IEEE #ICIP2023 in Kuala Lumpur 🇲🇾. Presented our work & Kudos to contributors. 'Functional Knowledge Transfer with Self-supervised Representation Learning' Also It's wonderful opportunity to meet Prof. Mubarak Shah to get valuable advice here & in #ICCV2023.

Check out my sketch in the #ICIP 2023 Sketch Mosaic! https://t.co/vauZTbYERD

Honored to have met the legendary @ylecun at #ICCV2023 🇫🇷. I got a chance to join brief discussion and ask my question on self-supervision. You are the infinite inspiration! #AI #LeCun #SelfsupervisedLearning #ComputerVision #ICCV #phdlife

Thrilled to share my (on-going) #ICCV2023 Paris 🇫🇷 experience! Presented the recent research work at the ICCV OOD-CV workshop! Paper Title- Can Self-Supervised Representation Learning Methods Withstand Distribution Shifts and Corruptions? Paper- https://t.co/w8pYvOJK38

Wow, made it #chandrayaan3mission 🇮🇳🇮🇳 Incredible, already 56 million views for soft landing and 8 millions watched during live - https://t.co/VatZ7IDeK8 #Chandrayaan3

Two oral presentation at #IJCNN, #goldcoast #austraila Paper 1-Domain Adaptable Self-supervised Representation Learning on Remote Sensing Satellite Imagery Paper 2-Learning Self-Supervised Representations for Label Efficient Cross-Domain Knowledge Transfer on Diabetic RF Images

Attending #IJCNN #queensland

It's exciting to share the research work with diverse set of subject experts during #Research & #Innovation Week 2023 at University of the Western Cape, Cape Town. Thanks to South Africa - Sweden University Forum (SASUF)!