About

Portrait of Sarthak Kumar Maharana

I am a CS PhD Candidate at The University of Texas at Dallas, advised by Dr. Yunhui Guo. Prior to this, I earned my MS in Electrical Engineering from the University of Southern California. I completed my BTech in Electrical and Electronics Engineering at the International Institute of Information Technology (IIIT), Bhubaneswar, India.

I'm interested in building adaptive, efficient, and robust AI systems that keep learning when the world changes. My work spans continual learning, multimodal perception, and real-time motion planning and reasoning for autonomous driving.

Areas of interest: Computer Vision, Multimodal AI, Physical AI.

Sub areas: Continual Learning/Adaptation, Robust and Efficient Perception, On-Device AI, Autonomous Driving.

Open to research conversations & collaborations. Feel free to reach out!

Publications

My full publication record is available on Google Scholar.

2026

Towards Continual Test-Time Adaptation of Vision-Language Models in Open-Vocabulary Semantic Segmentation

Chandler Timm C Doloriel, Yunbei Zhang, Sarthak Kumar Maharana, Muhammad Salman Siddiqui, Tor Kristian Stevik, Fadi Al Machot, Kristian Hovde Liland, Habib Ullah

arXiv, 2026

Audio-Visual Continual Test-Time Adaptation without Forgetting

Sarthak Kumar Maharana, Akshay Mehra, Bhavya Ramakrishna, Yunhui Guo, Guan-Ming Su

European Conference on Computer Vision (ECCV), 2026

Continual Test-Time Adaptation in Computer Vision: Methods, Benchmarks, and Future Directions

Sarthak Kumar Maharana*, Shambhavi Mishra*, Yunbei Zhang*, Shuaicheng Niu, Taki Hasan Rafi, Jihun Hamm, Marco Pedersoli, Jose Dolz, Yunhui Guo

Transactions on Machine Learning Research (TMLR), 2026

Variational Diffusion Unlearning: A Variational Inference Framework for Unlearning in Diffusion Models under Data Constraints

Subhodip Panda, Varun M S, Shreyans Jain, Sarthak Kumar Maharana, Prathosh AP

Transactions on Machine Learning Research (TMLR), 2026

2025

AVROBUSTBENCH: Benchmarking the Robustness of Audio-Visual Recognition Models at Test-Time

Sarthak Kumar Maharana, Saksham Singh Kushwaha, Baoming Zhang, Adrian Rodriguez, Songtao Wei, Yapeng Tian, Yunhui Guo

Conference on Neural Information Processing Systems (NeurIPS) Datasets & Benchmarks, 2025

BATCLIP: Bimodal Online Test-Time Adaptation for CLIP

Sarthak Kumar Maharana, Baoming Zhang, Leonid Karlinsky, Rogerio Feris, Yunhui Guo

International Conference on Computer Vision (ICCV), 2025

PALM: Pushing Adaptive Learning Rate Mechanisms for Continual Test-Time Adaptation

Sarthak Kumar Maharana, Baoming Zhang, Yunhui Guo

AAAI Conference on Artificial Intelligence (AAAI), 2025 · Oral

SELECT: A Submodular Approach for Active LiDAR Semantic Segmentation

Ruiyu Mao, Sarthak Kumar Maharana, Xulong Tang, Yunhui Guo

arXiv, 2025

2024

STONE: A Submodular Optimization Framework for Active 3D Object Detection

Ruiyu Mao, Sarthak Kumar Maharana, Rishabh K Iyer, Yunhui Guo

Conference on Neural Information Processing Systems (NeurIPS), 2024

Not Just Change the Labels, Learn the Features: Watermarking Deep Neural Networks with Multi-View Data

Yuxuan Li, Sarthak Kumar Maharana, Yunhui Guo

European Conference on Computer Vision (ECCV), 2024

Acoustic-to-Articulatory Inversion for Dysarthric Speech: Are Pre-trained Self-Supervised Representations Favorable?

Sarthak Kumar Maharana, Krishna Kamal Adidam, Shoumik Nandi, Ajitesh Srivastava

ICASSP SASB Workshop, 2024

2021

Acoustic-to-Articulatory Inversion for Dysarthric Speech by Using Cross-Corpus Acoustic-Articulatory Data

Sarthak Kumar Maharana, Aravind Illa, Renuka Mannem, Yamini Belur, Preetie A. Shetty, Veeramani Preethish Kumar, Seena Vengalil, Kiran Polavarapu, Atchayaram Nalini, Prasanta Ghosh

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021

Miscellaneous

Experience

I work across academic and industrial research, connecting fundamental questions.

NEC Laboratories America, Inc.

Research Scientist Intern

Real-time reasoning and closed-loop trajectory planning for autonomous driving in long-tailed scenes.

Dolby Laboratories, Inc.

PhD Research Intern

Robust audio-visual learning in continual learning settings.

The University of Texas at Dallas

PhD Candidate · Computer Science

Advised by Dr. Yunhui Guo in the Data Efficient Intelligent Learning Lab.

Workshops

News

Received an ECCV 2026 travel award; will also review for AAAI 2027.

AVReCAP was accepted to ECCV 2026.

I'll be joining NEC Laboratories America to work on real-time reasoning and planning for autonomous driving.

AVROBUSTBENCH was accepted to NeurIPS Datasets & Benchmarks.

Received a travel grant from the ICCV community!

BATCLIP was accepted to ICCV 2025.

Old news 2024 — 2026

Passed my qualifying examination and became a PhD candidate.

PALM was accepted to AAAI 2025 as an oral presentation.

Variational Diffusion Unlearning was accepted to the NeurIPS SafeGenAI Workshop.

Our submodular optimization work for active 3D object detection was accepted to NeurIPS 2024.

Contact

I’m always glad to discuss research collaborations and emerging ideas. I have a knack for finding interesting problems or digging into the teeny-tiny details of a paper. Feel free to contact!

Start a conversation