Seung Wook Kim

Papers

Total Citations

H-Index

About

Seung Wook Kim is a researcher specializing in generative modeling, neural simulation, and 3D scene synthesis, with a particular focus on bridging machine learning and real-world applications in robotics and virtual reality. His work sits at the intersection of computer vision, deep learning, and interactive environment generation, pushing the boundaries of what neural networks can learn and create. Kim's most notable contribution, **NeuralField-LDM** (2023), has garnered significant attention with over 43 citations, introducing a hierarchical latent diffusion framework capable of synthesizing complex, high-quality 3D environments — a breakthrough for VR and robotics simulation pipelines. His earlier work on **GameGAN** (2020) demonstrated that neural networks could learn to simulate dynamic environments simply by observing gameplay, removing the need for hand-crafted rules. Building on this, **DriveGAN** (2021) advanced controllable, high-fidelity neural simulation for autonomous driving scenarios, highlighting his sustained commitment to scalable, data-driven simulators. Across his publications, Kim has consistently tackled the challenge of making simulation more accessible and realistic through learned models. His research has meaningful implications for autonomous systems, game development, and synthetic data generation, establishing him as an emerging voice in generative AI for 3D and dynamic environments.