Bo Wu

Bo Wu is a Researcher at the MIT-IBM Watson AI Lab in Cambridge, MA, where he conducts pioneering research in deep learning, computer vision, natural language processing, and multimodal learning. Previously, he served as a postdoctoral research scientist at Columbia University after completing his Ph.D. at the Chinese Academy of Sciences (CAS) in Beijing, with additional research experience at Microsoft Research Asia (MSRA) and Academia Sinica.

His academic foundation includes:

• Ph.D. in Computer Science, Chinese Academy of Sciences (CAS)
• Research internships at Microsoft Research Asia and Academia Sinica

Wu's research focuses on advancing situated reasoning in real-world contexts, integrating neuro-symbolic approaches with deep learning for enhanced interpretability. His work spans video question answering, temporal forecasting, and multimodal understanding, with applications in social media prediction, enterprise AI, and personalized dialogue systems. He emphasizes bridging symbolic reasoning with neural networks to develop robust systems capable of handling open-world knowledge and dynamic environments.

Analysis of his recent publications reveals three dominant trends: the creation of novel benchmarks for situated video reasoning (STAR, SOK-Bench), development of efficient multimodal architectures for enterprise applications (Granite Vision), and personalization techniques for language models. His research consistently merges computer vision with linguistic understanding while addressing practical constraints like real-time processing and model compression, demonstrating strong industry-academia translation.

His scientific excellence is evidenced by prestigious recognitions including:

• IBM Master Inventor Award (2023)
• IBM Research Level-A Accomplishment Award (2021)
• ACL Best Demo Paper Award (2020)
• ICIP Prediction Challenge Champion (2020)
• Alibaba Global Vision AI Challenge Top 3 (2018)
• NIST TAC SM-KBP Top 1 (2019)

Wu actively mentors emerging talent, currently recruiting students for vision-language projects. He provides significant academic service as Area Chair for ACM Multimedia, Senior Program Committee Member for AAAI and IJCAI, and organizer of the SMP Challenge at ACM Multimedia since 2017. His leadership extends to CVPR workshops on Multimodal Foundations Models (MMFM) and Multimodal Video Content Understanding (MVCS), while serving on program committees for NeurIPS, CVPR, ACL, and other top-tier conferences.

As a core member of the MIT-IBM Watson AI Lab, Wu operates within a unique industry-academia ecosystem that fosters rapid translation of fundamental research into practical applications. His collaborative work with Chuang Gan and other researchers leverages IBM's computational resources and MIT's academic rigor, positioning him at the forefront of enterprise AI innovation where theoretical advances directly address real-world business challenges.