Dr. Jehanzeb Mirza

MIT CSAIL

Research Expertise

Computer Vision
Machine Learning
Deep Learning
Multi-Modal Learning

About

Hi, I am Jehanzeb Mirza. I am a Postdoctoral Researcher at [MIT CSAIL](https://www.csail.mit.edu/), in the Spoken Language Systems Group, led by Dr. [James Glass](https://www.csail.mit.edu/person/jim-glass). I received my Ph.D. in Computer Science (Computer Vision) from [TU Graz, Austria](https://www.tugraz.at/home), where I was advised by Professor [Horst Bischof](https://scholar.google.com/citations?user=_pq05Q4AAAAJ&hl=en), and Professor [Serge Belongie](https://sergebelongie.github.io/) served as an external referee. I am particularly interested in self-supervised learning for uni-modal models and multi-modal learning for vision-language models, with a focus on improving fine-grained understanding.

Legacy Map

Full View

Publications

The Norm Must Go On: Dynamic Unsupervised Domain Adaptation by Normalization
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
2022
An Efficient Domain-Incremental Learning Approach to Drive in All Weather Conditions
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
2022
Robustness of Object Detectors in Degrading Weather Conditions
2021 IEEE International Intelligent Transportation Systems Conference (ITSC)
2021
MAtch, eXpand and Improve: Unsupervised Finetuning for Zero-Shot Action Recognition with Language Knowledge
2023 IEEE/CVF International Conference on Computer Vision (ICCV)
2023
Video Test-Time Adaptation for Action Recognition
2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
2023
ActMAD: Activation Matching to Align Distributions for Test-Time-Training
2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
2023
Towards Multimodal In-context Learning for Vision and Language Models
Lecture Notes in Computer Science
2025
MATE: Masked Autoencoders are Online 3D Test-Time Learners
2023 IEEE/CVF International Conference on Computer Vision (ICCV)
2023
Meta-prompting for Automating Zero-Shot Visual Recognition with LLMs
Lecture Notes in Computer Science
2024
Comment on the Paper Titled ’The Origin of Quantum Mechanical Statistics: Insights from Research on Human Language’ (arXiv preprint arXiv:2407.14924, 2024)
Unknown Venue
2024
ConMe: Rethinking Evaluation of Compositional Reasoning for Modern VLMs
Advances in Neural Information Processing Systems 37
2024
Can Biases in ImageNet Models Explain Generalization?
2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
2024
Comparison Visual Instruction Tuning
2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
2025
Preprint site arXiv is banning computer-science reviews: here’s why
Nature
2025
TTT-KD: Test-Time Training for 3D Semantic Segmentation Through Knowledge Distillation From Foundation Models
2025 International Conference on 3D Vision (3DV)
2025
Exploring Modality Guidance to Enhance VFM-Based Feature Fusion for UDA in 3D Semantic Segmentation
2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
2025
ARXİV SƏNƏDLƏRİNİN MÜHAFİZƏXANALARDA MÜHAFİZƏ QAYDALARI
ADMİU Elmi Əsərlər
2025
Test-Time Adversarial Detection and Robustness for Localizing Humans Using Ultra Wide Band Channel Impulse Responses
2023 31st European Signal Processing Conference (EUSIPCO)
2023
Sit Back and Relax: Learning to Drive Incrementally in All Weather Conditions
2023 IEEE Intelligent Vehicles Symposium (IV)
2023
Influence Prediction in Collaboration Networks: An Empirical Study on arXiv
Unknown Venue
2025
Evaluation of Spatio-Temporal Small Object Detection in Real-World Adverse Weather Conditions
2025 IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW)
2025
DRT: Detection Refinement for Multiple Object Tracking
Proceedings of the British Machine Vision Conference 2021
2021
Detector-Free Weakly Supervised Grounding by Separation
2021 IEEE/CVF International Conference on Computer Vision (ICCV)
2021
Semi-Supervised Audio-Visual Action Recognition with Audio Source Localization Guided Mixup
2025 IEEE 35th International Workshop on Machine Learning for Signal Processing (MLSP)
2025
Affine calibration from moving objects
Proceedings Eighth IEEE International Conference on Computer Vision. ICCV 2001
Shape-Biased Texture Agnostic Representations for Improved Textureless and Metallic Object Detection and 6D Pose Estimation
2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
2025
Online Continual Learning of Diffusion Models: Multi-Mode Adaptive Generative Distillation
2025 IEEE International Conference on Image Processing (ICIP)
2025
Enhancing Zero-Shot Vision Models by Label-Free Prompt Distribution Learning and Bias Correcting
Advances in Neural Information Processing Systems 37
2024

Education

TU Graz, Austria

Ph.D. in Computer Vision, Computer Vision / 2024

KIT, Germany

MS in ETIT / 2020

NUST, Pakistan

BS in EE / 2017

Graz University of Technology

Ph.D. in Computer Science / 2024

Karlsruhe Institute of Technology

M.Sc. in Electrical Engineering and Information Technology / 2020

National University of Science and Technology

B.Sc. in Electrical Engineering / 2017

Experience

Massachusetts Institute of Technology (MIT)

Postdoctoral Researcher / November, 2024December

Leading research on multimodal learning combining speech vision and language for scalable AI systems. Designing and evaluating methods to improve fine-grained reasoning in large language and vision-language models.

Graz University of Technology

Computer Vision Project Assistant / January, 2021October, 2024

Developed self-supervised and unsupervised learning techniques to improve neural network robustness to distribution shifts at test time. Conducted extensive research on LLMs and multimodal VLMs resulting in multiple publications at NeurIPS ICCV and CVPR.

Sony AI

Research Scientist Intern / May, 2024August, 2024

Designed multimodal learning methods integrating vision audio and language signals. Prototyped and evaluated models for cross-modal understanding in real-world scenarios.

Intel Labs

Master Thesis Researcher / January, 2020July, 2020

Evaluated robustness of state-of-the-art 2D and 3D object detectors for autonomous driving under adverse weather.

C++ Developer Intern / October, 2019December, 2019

Implemented state estimation using Unscented Kalman Filter in C++ and OpenCV for real-time object tracking.

Intel

Platform Application Engineer Intern / March, 2019August, 2019

Built an automation framework including PCB design and microcontroller integration to streamline internal workflows.

Join Jehanzeb on NotedSource!
Join Now

At NotedSource, we believe that professors, post-docs, scientists and other researchers have deep, untapped knowledge and expertise that can be leveraged to drive innovation within companies. NotedSource is committed to bridging the gap between academia and industry by providing a platform for collaboration with industry and networking with other researchers.

For industry, NotedSource identifies the right academic experts in 24 hours to help organizations build and grow. With a platform of thousands of knowledgeable PhDs, scientists, and industry experts, NotedSource makes connecting and collaborating easy.

For academic researchers such as professors, post-docs, and Ph.D.s, NotedSource provides tools to discover and connect to your colleagues with messaging and news feeds, in addition to the opportunity to be paid for your collaboration with vetted partners.

Expert Institutions
NotedSource has experts from Stanford University
Expert institutions using NotedSource include Oxfort University
Experts from McGill have used NotedSource to share their expertise
University of Chicago experts have used NotedSource
MIT researchers have used NotedSource
Proudly trusted by
Microsoft uses NotedSource for academic partnerships
Johnson & Johnson academic research projects on NotedSource
ProQuest (Clarivate) uses NotedSource as their industry academia platform
Slamom consulting engages academics for research collaboration on NotedSource
Omnicom and OMG find academics on notedsource
Unilever research project have used NotedSource to engage academic experts

Connect with researchers and scientists like Dr. Jehanzeb Mirza on NotedSource to help your company with innovation, research, R&D, L&D, and more.