Fuming (Felix) Yang

Research Fellow in Lichtman Lab at Harvard University

photo.jpg

fumingyang@fas.harvard.edu

Cambridge, MA 02138

Hi! I’m Fuming (Felix) Yang, a Research Fellow in the Lichtman Lab at Harvard University.

Research Interests

My research interests center on using AI/ML methods to build high-throughput imaging pipelines and on understanding how to extract biological insights from large brain datasets, while letting those insights, in turn, inform advances in AI/ML.

Current Work

I currently lead data-acquisition efforts at an electron microscopy facility at Harvard, funded by NIH-UM1 award as part of the High-throughput Integrative Mouse Connectomics (HI-MC) project—a collaboration between Harvard, Princeton, Google, MIT, the Allen Institute, the University of Cambridge, and Johns Hopkins University. Our goal is to reconstruct 10 mm³ of adult mouse cortex at nanometer resolution (~20 PB of data). I build, operate, and optimize the imaging pipeline, develop high-throughput quality-control workflows, and carry out downstream analyses. The pipeline combines a custom tissue mill with a 91-beam scanning electron microscope (Zeiss MultiSEM) to enable automated, record-speed imaging of large neural volumes.

Research Areas

  • Computer Vision & ML: Graph theory, image compression, 2D/3D reconstruction, NN architecture
  • Neuroscience: Connectomics, developmental neuroscience, synapses elimination
  • Software & Tools: High-throughput imaging pipelines, automated acqusition/QC workflows

“Shoot for the moon. Even if you miss, you land among stars”

news

Jan 13, 2026 🎉 Two papers accepted to ISBI 2026! Towards 1000-Fold Electron Microscopy Image Compression for Connectomics via VQ-VAE with Transformer Prior and WaferTools: A Software for Section Segmentation, ROI Registration, and Section Ordering!
Nov 24, 2025 🎉 DRIFT-EM: Direct retrieval of ultrathin serial sections by ionizer-facilitated transfer for volume electron microscopy submitting to Cell Reports Methods!
Nov 8, 2025 🎉 WaferTools - Software for Wafer-related Automatic Methods in Connectomics. submitted to ISBI 2026!
Oct 31, 2025 🏆 Our Project (Bio-inspired AI) has been nominated by Harvard University for the Kempner Institute Accelerator Award recognizing novel AI/ML methods, grateful to all collaborators!
Oct 31, 2025 🎉 Towards 1000-fold Electron Microscopy Image Compression for Connectomics via VQ-VAE with Transformer Prior submitted to arXiv and ISBI 2026!
Sep 30, 2025 🎉 Neuromuscular connectomes across development reveal synaptic ordering rules got accepted by Nature Neuroscience! Huge milestone for the team!
Sep 26, 2025 🎉 Budgeted Broadcast: An Activity-Dependent Pruning Rule for Neural Network Efficiency submitted to ICLR 2026!
Sep 16, 2025 🎉 Developmental Connectomics of the Mouse Cerebellum is submitting to Nature!
Sep 7, 2025 🎉 Sparse Graph Reconstruction and Seriation for Large-Scale Image Stacks submitted to arXiv!
May 9, 2025 🎉 Joined Lichtman Lab at Harvard University as a Research Fellow!
Aug 30, 2024 🏆 Department Head’s Fellowship and Academic Excellence Fellowship from Carnegie Mellon University!

selected publications

  1. MICCAI
    Sparse Graph Reconstruction and Seriation for Large-Scale Image Stacks
    Fuming Yang, Yaron Meirovitch, and Jeff W. Lichtman
    MICCAI (submitting) - ArXiv preprint, 2025
  2. ISBI
    Towards 1000-fold Electron Microscopy Image Compression for Connectomics via VQ-VAE with Transformer Prior
    Fuming Yang, Yicong Li, Hanspeter Pfister, and 2 more authors
    ISBI (in press) - ArXiv preprint, 2025
  3. ICLR
    Budgeted Broadcast: An Activity-Dependent Pruning Rule for Neural Network Efficiency
    Yaron Meirovitch*, Fuming Yang*, Jeff W. Lichtman, and 1 more author
    ICLR (under review) - ArXiv preprint, 2025
    *Equal contribution
  4. ISBI
    Software for Wafer-related Automatic Methods in Connectomics
    Fuming Yang, Yaron Meirovitch, Frederico Araújo, and 3 more authors
    ISBI (in press), 2025
  5. Nat Neurosci
    Neuromuscular connectomes across development reveal synaptic ordering rules
    Yaron Meirovitch, Kai Kang, Ryan W. Draft, and 12 more authors
    Nature Neuroscience (in press), 2025
  6. Cell Rep
    DRIFT-EM: Direct retrieval of ultrathin serial sections by ionizer-facilitated transfer for volume electron microscopy
    Nelson Medina, Joseph V Gogola, Kevin Boergens, and 5 more authors
    Cell Reports Methods(under review), 2025