Research

My research streams center around the application of machine learning and deep learning to analyze large-scale unstructured data, including audio, text, and video, with the goal of inferring consumer behavior and enhancing business decision-making. These streams can be broadly categorized as three areas: Developing Deep Learning Artifacts from Unstructured Data: In this stream, I focus on developing novel deep learning algorithms inspired by computer science literature, which are tailored to business contexts by considering the specific characteristics of the data and the business domain Applying Deep Learning Artifacts for Consumer Behavior Inference and Business Decision-Making: This research stream involves utilizing the deep learning artifacts developed in the first stream. The developed algorithms are applied to large-scale datasets gathered from social media and crowd-based platforms Assessing the Impact of State-of-the-Art (SOTA) Algorithms on Business: The last research stream focuses on evaluating the impact of cutting-edge algorithms on business decision-making such as ChatGPT. Additionally, I explore how these algorithms can be fine-tuned or adapted based on specific business contexts and requirements to enhance overall business outcomes.

Computer Science Publications


Multimodal Co-attention Transformer for Video-Based Personality Understanding

Mingwei Sun and Kunpeng Zhang

2023 IEEE International Conference on Big Data, Sorrento, Italy, 2023, pp. 1450 -1459

Acceptance rate: 17.5%

Paper Code

Sec2Sec Co-Attention Transformer for Video Emotion Prediction

Mingwei Sun and Kunpeng Zhang

2024 IEEE International Conference on Acoustics, Speech and Signal Processing, Seoul, Korea, 2024, pp. 8255 - 8259

Paper Code

Under Review


Network-enhanced Multimodal Co-attention Learning for Short Video Popularity Prediction

Mingwei Sun and Kunpeng Zhang

Under review at IEEE/CVF Winter Conference on Applications of Computer Vision


Working Papers


Large Language Model for Business-Value Alignment

Mingwei Sun, Balaji Padmanabhan and Kunpeng Zhang

To be submmited


An Interpretable Multimodal Framework for Video-Based Apparent Personality Understanding

Mingwei Sun and Kunpeng Zhang

To be submmited to Information Systems Research


Research-in-Progress


Biases in Generative AI

With Che-Wei Liu and Jisu Cao

An Interpretable Reinforcement Learning from Human Feedback in Image-to-Video Generation

With Che-Wei Liu and Xiaowei Liu

The Impact of Face and Voice Alignment

With Cathy Liu and Xitong Li

Video Summarization

With Kunpeng Zhang

Video-Music Generation

With Kunpeng Zhang and Bingze Xu