Research

My research interest lies in the fields of computer vision, graphics, and multimodal interaction, with a specific focus on problems associated with virtual humans, embodied intelligent agents, and human-agent interaction. My past research includes designing generative models for animating human faces, hand and body gestures, and collective group activities. Additionally, I’ve extensively researched the capability of LLMs for conversational avatars, digital storytelling, narrative graphs, and autonomous social agents. Another branch of my research focus is to unleash the potential of synthetic data, powered by CG render engines and generative diffusion models, to ultimately address data scarcity and improve model generalizability in computer vision tasks, leading to more robust and reliable AI systems.

Contact me at chejui.chang@rutgers.edu or check out my LinkedIn

CURRICULUM VITAE

Education

• PhD, Computer Science, Rutgers University, 2020-present

• MS, Communication Engineering, National Taiwan University, 2016-2018

• BS, Physics, National Taiwan University, 2012-2016

Publications

• **New** C. -J. Chang, D. Li, S. Moon, M. Kapadia, "On the Equivalency, Substitutability, and Flexibility of Synthetic Data," arXiv preprint (SynData4CV @ CVPR24)

• **New** C. -J. Chang, D. Li, D. Patel, P. Goel, H. Zhou, S. Moon, S. S. Sohn, S. Yoon, V. Pavlovic, M. Kapadia, "Learning from Synthetic Human Group Activities," In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024 [Project Page]

• C. -J. Chang, S. S. Sohn, S. Zhang, R. Jayashankar, M, Usman, M. Kapadia, "The Importance of Multimodal Emotion Conditioning and Affect Consistency for Embodied Conversational Agents," ACM Conference on Intelligent User Interfaces 2023

• C. -J. Chang, S. Zhang, and M. Kapadia, "The IVI Lab entry to the GENEA Challenge 2022–A Tacotron2 based method for co-speech gesture generation with locality-constraint attention mechanism," International Conference on Multimodal Interaction, 2022 (Reproducibility Award, Best Appropriateness!!)

• C. -J. Chang, L. Zhao, S. Zhang, and M. Kapadia, "Disentangling Audio Content and Emotion with Adaptive Instance Normalization for Expressive Facial Animation Synthesis," Computer Animation and Virtual Worlds, 2022.

• C. -J. Chang, "Transfer Learning from Monolingual ASR to Transcription-free Cross-lingual Voice Conversion," arXiv preprint arXiv:2009.14668, 2020.

• C. -J. Chang, and S. -K. Jeng, "Acoustic Anomaly Detection Using Multilayer Neural Networks and Semantic Pointers," Journal of Information Science and Engineering 2020.

• C. -J. Chang, "Humanoid Auditory Processing for Acoustic Anomaly Detection,” National Taiwan University, Master Thesis, 2018.

[Google Scholar]

Experience

• PhD Research Intern @ Roblox, Summer 2023 - Present

• Graduate Assistant @ Rutgers University, Summer 2022 - Present

• Teaching Assistant, CS439 Introduction to Data Science @ Rutgers University, Spring 2022

• Teaching Assistant, CS523 Computer Graphics @ Rutgers University, Fall 2021

• Teaching Assistant, CS206 Introduction to Discrete Structures II @ Rutgers University, Summer 2021

• Teaching Assistant, CS205 Introduction to Discrete Structures I @ Rutgers University, Spring 2021

• AI Scientist, Red Pill Lab, 2019-2020

• Machine Learning Engineer Intern @ Aeolus Robotics, 2018

• Research Assistant @ Academia Sinica, 2018

• Research Assistant @ National Taiwan University, 2016

Referee Experience

• Program Committee, International Conference on Multimodal Interaction (ICMI) 2024

• Program Committee, International Conference on Autonomous Agents and Multiagent Systems (AAMAS) 2022

• International Conference on Multimodal Interaction (ICMI) 2023

• ACM Siggraph/Eurographics Symposium on Computer Animation (SCA) 2022

• European Association for Computer Graphics (Eurographics) 2022

• Association for the Advancement in Artificial Intelligence (AAAI) Conference 2022

• IEEE International Conference on Artificial Intelligence and Virtual Reality (AIVR) 2021.

• Conference on Neural Information Processing Systems (NIPS) 2021.

• IEEE International Conference on Data Mining (ICDM) 2021.

• International Conference on Computer Animation and Social Agents (CASA) 2021.

• ACM Siggraph/Eurographics Symposium on Computer Animation (SCA) 2021.

• IEEE Transactions on Visualization and Computer Graphics (TVCG) 2021.

Speaking

• "Learning from Synthetic Data for Human-Centered Tasks", at National Taiwan University, 2024.

• "The Importance of Multimodal Emotion Conditioning and Affect Consistency for Embodied Conversational Agents", at International Conference on Intelligent User Interfaces, Sydney, 2023.

• "The IVI Lab entry to the GENEA Challenge 2022 – A Tacotron2 Based Method for Co-Speech Gesture Generation With Locality-Constraint Attention Mechanism", at International Conference on Multimodal Interaction, 2022.

• "Animating Expressive Human Faces: From Parametric 3D Models To Data-Driven Approaches", at Rutgers University, 2022.

• "Disentangling Audio Content and Emotion with Adaptive Instance Normalization for Expressive Facial Animation Synthesis", at International Conference on Computer Animation and Social Agents, 2022.

• "3D Facial Reconstruction from Monocular Videos and its Applications", at Rutgers University, 2021

Patent

• Method for converting voice into virtual face image (CN112992120A)

Past Projects

• LLM for Digital Storytelling, narrative graphs, and autonomous social agents

• Real-time Body Tracking for Personalized Avatars

• 3D Group Activity Generation

• Synthetic Data Generation for Human Motions and Group Activities

• Evaluation of Emotion Perception for Multimodal ECAs

• Audio-Driven Gesture Generation

• Emotion-Conditioned Speech-Driven Facial Animation Synthesis

• Speech-Driven Lip Synchronization