ArtifactArena: Evaluating Models by What They Build in the Physical World
Making is the measure of intelligence: models design and build functional robots that compete head-to-head in a simulated arena, producing a continuously updated Elo ranking of what models can build rather than what they say.
Kushagra Tiwary *, David Mayo*, Nikhil Behari, Xiangzhou Sun, Abdulrahman Alabdulkareem, Isaac Galatzer-Levy, Boris Katz, Brian Cheung
Agent Systems for Academic Research Automation
Introduces verifier-matched autonomy: an AI scientist's autonomy is only meaningful relative to the verifiers available to check the claims it helps make.
Pierfrancesco Beneventano, Riccardo Neumarker, Theodoros Evgeniou, Marc Gong Bacvanski, Kushagra Tiwary , Emanuele Rimoldi, Mehdi Hajoub, Yulu Gan, Qianli Liao, Mahmoud Abdelmoneum, Liu Ziyin, Tomer Galanti, Andrea Pinto, Tomaso Poggio
[
Paper ,
(ICML 2026 AI for Science Workshop) ]
AI For Discovery
What if Eye...? Computationally Recreating the Evolution of Vision
We created a virtual petri dish where digital creatures evolve eyes from scratch, replaying millions of years of evolution to understand vision.
Kushagra Tiwary *, Aaron Young*, Zaid Tasneem, Tzofi Klinghoffer, Akshat Dave, Tomaso Poggio, Dan-Eric Nilsson, Brian Cheung, Ramesh Raskar
Designing Visual Intelligence for Science and Engineering
Perspective: How biological principles can guide the creation of new forms of artificial visual intelligence.
Kushagra Tiwary , Tzofi Klinghoffer, Aaron Young, Siddharth Somasundaram, Nikhil Behari, Akshat Dave, Brian Cheung, Dan-Eric Nilsson, Tomaso Poggio, Ramesh Raskar
[
MIT Press ,
(Impacts of Generative AI, MIT Press, 2024) ]
Perspective
AI For Discovery
Bridging the Data Provenance Gap Across Text, Speech, and Video
Largest and first-of-its-kind longitudinal audit of text, speech, and video datasets used to train AI models.
Shayne Longpre, Nikhil Singh, Manuel Cherep, Kushagra Tiwary , Joanna Materzynska, William Brannon, Robert Mahari, ..., Sara Hooker, Jad Kabbara
Consent in Crisis: The Rapid Decline of the AI Data Commons
A critical examination of the legal, ethical, and infrastructural unraveling of the AI data commons.
Collaborative publication across interdisciplinary researchers.
DecentNeRFs: Private and Decentralized 3D from Crowd Sourced Image Data
Can we use crowdsourced images to create 3D models of the world?
Zaid Tasneem, Akshat Dave, Abhishek Singh, Kushagra Tiwary , Praneeth Vepakomma, Ashok Veeraraghavan, Ramesh Raskar
SUNDIAL: 3D Satellite Understanding through Direct Ambient and Complex Lighting Decomposition
Can we use complex lighting to understand 3D objects?
Nikhil Behari, Akshat Dave, Kushagra Tiwary , William Yang, Ramesh Raskar
[
Paper ,
(CVPR'W 2024) ]
Computer Vision
DISeR: Designing Imaging Systems with Reinforcement Learning
Can we use reinforcement learning to design new imaging systems?
Tzofi Klinghoffer*, Kushagra Tiwary *, Nikhil Behari, Bhavya Agrawalla, Ramesh Raskar
Objects as Radiance Field Cameras
Can we convert arbitrary objects into cameras to recover 3D?
Kushagra Tiwary , Akshat Dave, Nikhil Behari, Tzofi Klinghoffer, Ashok Veeraraghavan, Ramesh Raskar
Towards Learning Neural Representations from Shadow
Can we learn hidden neural representations from binary shadow images?
Kushagra Tiwary , Tzofi Klinghoffer, Ramesh Raskar
MIT Museum
Experience First Signs of Vision during this special adults-only evening event featuring ocean-themed exhibits and activities.
When: Thursday, April 10, 2025, 6:00pm - 9:00pmWhere: MIT Museum, 75 Amherst Street, Cambridge, MATickets: $20 General, $10 MIT ID Holders
Museum of Science Boston
Explore the First Signs of Vision exhibit at the Museum of Science Boston, perfect for all ages to discover the evolutionary history of sight.
When: Wednesday, April 2, 2025, 1:00pm - 4:00pmWhere: Museum of Science, BostonTickets: Included with museum admission
A Perspective on Decentralizing AI
Discusses emerging approaches and motivations behind building decentralized AI systems and infrastructure.
Abhishek Singh, Pradyumna Chari, Charles Lu, Gauri Gupta, Ayush Chopra, Jonas Blanc, Tzofi Klinghoffer, Kushagra Tiwary , Ramesh Raskar
Physics vs. Learned Priors: Rethinking Camera and Algorithm Design for Task-Specific Imaging
Perspective paper contrasting physics-based priors with data-driven methods in imaging system design.
Tzofi Klinghoffer*, Siddharth Somasundaram*, Kushagra Tiwary *, Ramesh RaskarPresented at ICCP 2022
[
Paper ,
(ICCP 2022) ]
Perspective
Computer Vision
Decentralized Neural Radiance Fields from Crowdsourced Images
Decentralized 3D reconstruction from crowdsourced images while separating personal content from shared scene information.
Inventors: Akshat Dave, Ramesh Raskar, Abhishek Singh, Zaid Tasneem, Kushagra Tiwary , Ashok Veeraraghavan, Praneeth Vepakomma
[
US20250336144A1 · Application published
October 30, 2025 ]
Patent
Computer Vision
Efficient detection of structure and status of traffic lights
Patent for a Software 2.0 system design at Optimus Ride (self-driving car company).
Inventors: Kushagra Tiwary , Nuno Sucena Almeida
You, Your AI, and Your Research
A workshop tutorial on how to build ecosystems where AI can accelerate your research.
Slides Coming soon!
Neural Fields Beyond Conventional Cameras
Workshop on novel neural field architectures for unconventional imaging systems.
Organizers: Kushagra Tiwary , Zaid Tasneem, Akshat Dave, Ramesh Raskar
[
Website ,
(ECCV 2024) ]
Workshop
Computer Vision
Accelerating Discovery
Workshop at MIT Member's Week on AI systems for accelerating scientific discovery.
Organizers: Kushagra Tiwary , Tzofi Klinghoffer, Aaron Young, Ramesh Raskar
[(MIT Member's Week 2023) ]
Workshop
Artificial Cambrian Intelligence (ACI)
Open source code for evolving artificial visual systems.
Aaron Young*, Kushagra Tiwary *
[
GitHub ]
Open Source
AI For Discovery