Research focus
About the research
How can AI models be made efficient enough to run at lower computational cost?
Read the full research overview
Song Han is an associate professor with tenure at MIT EECS. His work includes the Efficient Inference Engine, which first introduced weight sparsity to modern AI chips. His innovations include TinyML and hardware-aware neural architecture search for deploying models on resource-constrained devices. His recent work on LLM quantization and acceleration has improved efficiency in LLM inference.
Selected publications
3 listedAWQ: Activation-aware Weight Quantization for On-Device LLM Compression and Acceleration
GetMobile Mobile Computing and Communications
FastComposer: Tuning-Free Multi-subject Image Generation with Localized Attention
International Journal of Computer Vision
Tiny Machine Learning: Progress and Futures [Feature]
IEEE Circuits and Systems Magazine
People & alumni
Team details haven’t been added. See the lab’s website for its current members.
Alumni details haven’t been added. Check the lab’s official website.
Sources & listing information
Find Research Labs maintains this page using public sources. The faculty member, lab, and university have not claimed or approved this listing. Listing a lab does not indicate endorsement, partnership, or a recruiting opportunity.
Request a correction or removal ↗ · About directory listings