Siva Manivasagam

I am a PhD student at the University of Toronto, advised by Raquel Urtasun, and a member of the Vector Institute. I am also a Sr. Staff Researcher and Technical Lead Manager at Waabi, leading the sensor simulation team.

My research lies at the intersection of computer vision, machine learning, and robotics. I build scalable sensor simulation and 3D reconstruction systems for the safe development and testing of autonomous driving.

Email / Google Scholar / CV

Research

My research goal is to build scalable and realistic sensor simulation systems for the safe development of autonomous vehicles. I have worked on LiDAR and camera simulation, 3D reconstruction of objects and large-scale scenes, world modeling, and safety-critical scenario generation.

	Diffusion-guided Generalizable Enhancer for Urban Scene Reconstruction Henry Che, Jingkang Wang, Yun Chen, Ze Yang, Sivabalan Manivasagam, Raquel Urtasun International Conference on Robotics and Automation (ICRA), 2026 project / paper A generalizable 3D enhancer for urban scene reconstruction using diffusion guidance.
	SaLF: Sparse Local Fields for Multi-Sensor Rendering in Real-Time Yun Chen, Matthew Haines, Jingkang Wang, Sahil Jain, Krzysztof Baron-Lis, Sivabalan Manivasagam, Ze Yang, Raquel Urtasun International Conference on Robotics and Automation (ICRA), 2026 project / paper / video A unified representation for multi-sensor simulation in real-time via rasterization and ray tracing.
	Flux4D: Flow-based Unsupervised 4D Reconstruction Jingkang Wang, Henry Che, Yun Chen*, Ze Yang, Lily Goli, Sivabalan Manivasagam, Raquel Urtasun Neural Information Processing Systems (NeurIPS), 2025 project / paper / video A simple, scalable framework for unsupervised generalizable 4D reconstruction of large-scale driving scenes.
	GenAssets: Generating in-the-wild 3D Assets in Latent Space Ze Yang, Jingkang Wang, Haowei Zhang, Sivabalan Manivasagam, Yun Chen, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2025 project / paper / video Reconstruct or generate 360° assets from in-the-wild images and LiDAR with UniSim + latent diffusion.
	G3R: Gradient Guided Generalizable Reconstruction Yun Chen, Jingkang Wang, Ze Yang, Siva Manivasagam, Raquel Urtasun European Conference on Computer Vision (ECCV), 2024 project / paper / video A generalizable large reconstruction model for outdoor dynamic scenes with real-time full camera simulation.
	UniCal: Unified Neural Sensor Calibration Ze Yang, George Chen, Haowei Zhang, Kevin Ta, Ioan Andrei Bârsan, Daniel Murphy, Siva Manivasagam, Raquel Urtasun European Conference on Computer Vision (ECCV), 2024 project / paper / video Unified neural approach to multi-sensor calibration for autonomous driving.
	LightSim: Neural Lighting Simulation for Urban Scenes Ava Pun, Gary Sun, Jingkang Wang*, Yun Chen, Ze Yang, Siva Manivasagam, Wei-Chiu Ma, Raquel Urtasun Neural Information Processing Systems (NeurIPS), 2023 project / paper / video A data-driven neural lighting simulation system for urban scenes with diverse, controllable, and realistic output.
	Adv3D: Generating Safety-Critical 3D Objects through Closed-Loop Simulation Jay Sarva*, Jingkang Wang, James Tu, Yuwen Xiong, Siva Manivasagam, Raquel Urtasun Conference on Robot Learning (CoRL), 2023 project / paper / video Closed-loop simulation to identify challenging object shapes and autonomy failure modes.
	Real-Time Neural Rasterization for Large Scenes Jeffrey Liu, Yun Chen, Ze Yang, Jingkang Wang, Siva Manivasagam, Raquel Urtasun International Conference on Computer Vision (ICCV), 2023 project / paper / video Neural texture features + MLP shader for real-time rendering (>120 FPS at 1080p).
	Towards Zero Domain Gap: A Comprehensive Study of Realistic LiDAR Simulation for Autonomy Testing Siva Manivasagam, Ioan Andrei Bârsan, Jingkang Wang, Ze Yang, Raquel Urtasun International Conference on Computer Vision (ICCV), 2023 project / paper / video Analyzing the impact of LiDAR sensor simulation fidelity on the full autonomy stack.
	UniSim: A Neural Closed-Loop Sensor Simulator Ze Yang, Yun Chen, Jingkang Wang, Siva Manivasagam, Wei-Chiu Ma, Anqi Joyce Yang, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2023 (Highlight) project / paper A data-driven closed-loop sensor simulator generating realistic counterfactual scenarios from a single driving pass.
	NeuSim: Reconstructing Objects in-the-wild for Realistic Sensor Simulation Ze Yang, Siva Manivasagam, Yun Chen, Jingkang Wang, Rui Hu, Raquel Urtasun International Conference on Robotics and Automation (ICRA), 2023 project / paper Physics-guided NeRF for object reconstruction from sparse, noisy in-the-wild data.
	Mending Neural Implicit Modeling for 3D Vehicle Reconstruction in the Wild Shivam Duggal, Zihao Wang, Wei-Chiu Ma, Sivabalan Manivasagam, Justin Liang, Shenlong Wang, Raquel Urtasun Winter Conference on Applications of Computer Vision (WACV), 2022 arXiv / paper Robust neural implicit shape reconstruction for vehicles from sparse, noisy real-world observations.
	CADSim: Robust and Scalable in-the-wild 3D Reconstruction for Controllable Simulation Jingkang Wang, Siva Manivasagam, Yun Chen, Ze Yang, Ioan Andrei Bârsan, Anqi Joyce Yang, Wei-Chiu Ma, Raquel Urtasun Conference on Robot Learning (CoRL), 2022 project / paper Inverse rendering of articulated geometry, PBR material, and lighting from CAD templates.
	Adversarial Attacks on Multi-Agent Communication James Tu, Tsun-hsuan Wang, Jingkang Wang, Siva Manivasagam, Mengye Ren, Raquel Urtasun International Conference on Computer Vision (ICCV), 2021 paper Adversarial robustness in cooperative multi-agent communication systems.
	GeoSim: Photorealistic Image Simulation with Geometry-Aware Composition Yun Chen, Frieda Rong, Shivam Duggal*, Shenlong Wang, Xinchen Yan, Siva Manivasagam, Shangjie Xue, Ersin Yumer, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2021 (Oral, Best Paper Award Candidate) project / paper Photorealistic image simulation via geometry-aware composition for self-driving.
	AdvSim: Generating Safety-Critical Scenarios for Self-Driving Vehicles Jingkang Wang, Ava Pun, James Tu, Siva Manivasagam, Abbas Sadat, Sergio Casas, Mengye Ren, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2021 project / paper Safety-critical scenario generation using realistic mixed-reality LiDAR simulation.
	S3: Neural Shape, Skeleton, and Skinning Fields for 3D Human Modeling Ze Yang, Shenlong Wang, Siva Manivasagam, Zeng Huang, Wei-Chiu Ma, Xinchen Yan, Ersin Yumer, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2021 paper Neural implicit fields for 3D human shape, skeleton, and skinning reconstruction.
	SceneGen: Learning to Generate Realistic Traffic Scenes Shuhan Tan, Kelvin Wong, Shenlong Wang, Siva Manivasagam, Mengye Ren, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2021 paper Learning to generate realistic traffic scene layouts for self-driving simulation.
	Recovering and Simulating Pedestrians in the Wild Ze Yang, Siva Manivasagam, Ming Liang, Bin Yang, Wei-Chiu Ma, Raquel Urtasun Conference on Robot Learning (CoRL), 2020 (Spotlight) paper Recovering 3D pedestrian models from LiDAR for realistic sensor simulation.
	V2VNet: Vehicle-to-Vehicle Communication for Joint Perception and Prediction Tsun-Hsuan Wang, Siva Manivasagam, Ming Liang, Bin Yang, Wenyuan Zeng, Raquel Urtasun European Conference on Computer Vision (ECCV), 2020 (Oral) paper Vehicle-to-vehicle communication network for joint perception and prediction in self-driving.
	Weakly-supervised 3D Shape Completion in the Wild Jiayuan Gu, Wei-Chiu Ma, Siva Manivasagam, Yuwen Xiong, Wenyuan Zeng, Zihao Wang, Hao Su, Raquel Urtasun European Conference on Computer Vision (ECCV), 2020 (Spotlight) paper Estimating shape and pose from partial LiDAR point clouds without ground-truth supervision.
	Deep Feedback Inverse Problem Solver Wei-Chiu Ma, Shenlong Wang, Jiayuan Gu, Siva Manivasagam, Antonio Torralba, Raquel Urtasun European Conference on Computer Vision (ECCV), 2020 (Spotlight) paper Solving inverse problems using the forward process as feedback for an iterative update model.
	LiDARsim: Realistic LiDAR Simulation by Leveraging the Real World Siva Manivasagam, Shenlong Wang, Kelvin Wong, Wenyuan Zeng, Mikita Sazanovich, Shuhan Tan, Bin Yang, Wei-Chiu Ma, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2020 (Oral) paper Realistic LiDAR sensor simulation leveraging real-world data for closed-loop evaluation.
	Physically Realizable Adversarial Examples for LiDAR Object Detection James Tu, Mengye Ren, Siva Manivasagam, Ming Liang, Bin Yang, Richard Du, Frank Cheng, Raquel Urtasun Conference on Computer Vision and Pattern Recognition (CVPR), 2020 paper Universal rooftop adversarial attack that hides vehicles from LiDAR object detectors.

Last updated: June 2026. Website template from Jon Barron.