Wenye (Bear) Xiong 熊闻野

I am currently a Visiting Undergraduate Student at Harvard University (Sep. 2025 – May 2026) and an undergraduate student at ShanghaiTech University (B.E. in Computer Science and Technology, Minor in Life Sciences, expected Jun. 2027).

My research interests lie in Multimodal Machine Learning, Computer Vision, and AI for Healthcare & Life Science, with a focus on compositional generative models, scene understanding, and spatial reasoning. Recently, I have been working on Compositional Scene Generation: learning disentangled representations from structured scene descriptions (e.g., scene graphs / object-relation specifications) to enable flexible control over diffusion-based generation and reasoning.

For more information, please see my Curriculum Vitae.

Research Interests

• Multimodal Machine Learning
• Computer Vision
• AI for Healthcare & Life Science
• Compositional Generative Models
• Scene Understanding & Spatial Reasoning

Education

• ShanghaiTech University — B.E. in Computer Science and Technology, Minor in Life Sciences (Sep. 2023 – Jun. 2027 expected)
  • GPA: 3.8/4.0 (Rank 13/172 in CS major; 19/267 in school)
• Harvard University — Visiting Undergraduate Student (Sep. 2025 – May 2026)
  • GPA: 4.0/4.0
  • Coursework includes: Computer Vision, Planning and Learning Methods in AI, Signal Processing (MIT cross-registration)

Experience

• Embodied Minds Lab, Harvard University & Kempner Institute — Visiting Undergraduate Research Assistant (Sep. 2025 – present)
  Supervisors: Prof. Yilun Du & Dr. Ruojin Cai
  • Developing an inverse generative modeling framework for scene understanding (e.g., CLEVR) with relation-conditioned, composable diffusion models.
  • Building an inference pipeline that predicts discrete spatial relations by scoring candidate relations through diffusion denoising energy with Monte Carlo sampling.
  • Extending the framework toward real-world scenes with DiT and cross-attention to better capture complex object relations.
• Perception, Learning, and UnderStanding (PLUS) Lab, ShanghaiTech — Undergraduate Research Assistant (Jan. 2025 – Sep. 2025)
  Supervisor: Prof. Xuming He
  • Investigated Compositional Scene Generation with scene-graph-based diffusion models.
  • Studied disentangled representations from structured scene descriptions to enable controllable generation.
  • Explored classifier-free guidance and training-free methods for image generation.
• Virtual Reality and Visual Computing Center (VRVC), ShanghaiTech — Undergraduate Research Assistant (Jun. 2024 – Oct. 2024)
  Supervisor: Dr. Minzhang Li
  • Explored diffusion-model-based approaches in protein structure prediction.
  • Implemented an SE(3)-equivariant Transformer for 3D equivariant structure modeling.
  • Contributed to ShanghaiTech Fold, a diffusion-based model for all-atom biomolecular assembly.
• MakeSense, ShanghaiTech First SensUs Team — Co-Founder & Leader of Data Analysis Team (Aug. 2024 – Aug. 2025)
  • Developed a wearable biosensor-based device to continuously monitor AKI biomarkers.
  • Worked on enzyme-based creatinine sensing and a QCM (Quartz Crystal Microbalance) platform.
  • Team won First Place (Analytical Performance) at SensUs 2025.
• PACIFY project for iGEM 2024 — Team Member (Dec. 2023 – Oct. 2024) [wiki]
  • Conducted homology modeling and structure prediction (including AlphaFold2-based workflows).
  • Performed protein preparation and molecular dynamics simulation.
  • Developed a PID-based device concept to address itchiness without harming skin.

Selected Course Projects

• De Novo Design of Odorant Binding Proteins for Breast Cancer Detection (Dec. 2024 – Jan. 2025)
  Supervisor: Prof. Jiayi Dou
  • Designed OBPs targeting VOC biomarkers (hexanal, octanal, nonanal).
  • Built a de novo computational design pipeline (RFdiffusionAA + LigandMPNN).
  • Validated via AutoDock, PyRosetta, and ESMFold for affinity and stability.
• Neural Olfactory Sensing and Evaluation (NOSE) (May. 2025 – Jun. 2025)
  Supervisor: Prof. Yujiao Shi
  • Fine-tuned MoLFormer on GS-LF for odor prediction tasks.
  • Benchmarked against OpenPOM on Keller-2016 (classification + pleasantness prediction).
  • Achieved state-of-the-art or comparable performance on key metrics.

Awards & Honors

• Merit Student (Top 2% in school), ShanghaiTech University, 2023–2024
• AI Honor Class (Honors Degrees), ShanghaiTech University, 2024–2027 (expected)
• Gold Medal, International Genetically Engineered Machine Competition (iGEM), 2024
• First Place, Analytical Performance, SensUs Competition, 2025
• Outstanding Mentor Assistant, ShanghaiTech University, 2023

Publications

No publications yet.

News

• 2025.09 Joined Embodied Minds Lab (Harvard University & Kempner Institute) as a Visiting Undergraduate Research Assistant.
• 2025.05 Admitted to the Visiting Undergraduate Student (VUS) program at Harvard University for the 2025–2026 academic year.
• 2025.01 Joined the PLUS Lab of SIST, ShanghaiTech University, advised by Prof. Xuming He.
• 2025 Our MakeSense team won First Place (Analytical Performance) at SensUs 2025.
• 2024.12 Awarded Merit Student of ShanghaiTech University (Top 2% in school).
• 2024.10 Our PACIFY project won the Gold Medal at iGEM 2024.
• 2023.10 Joined the AI Honor Program / AI Honor Class of SIST, ShanghaiTech University.
• 2024.06 Joined VRVC as an undergraduate research assistant.
• 2024.02 Joined ShanghaiTech iGEM Team.
• 2024.01 Awarded the 2023 Outstanding Mentor Assistant.