Dr. Song Han
MIT / HAN Lab
Associate Professor of EECS at MIT and Director of HAN Lab. Co-founder of OmniML (acquired by NVIDIA). Pioneer of Deep Compression, AWQ (Activation-aware Weight Quantization), SmoothQuant, and TinyEngine, pioneering modern efficient deep learning and on-device foundation model execution.
Areas of focus
Professional niches
Proof of Work
SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models
Invented SmoothQuant, an algorithm that mathematically splits quantization difficulty between activations and weights via per-channel scaling transformations, unlocking 8-bit weight and activation (W8A8) tensor-core acceleration.
Requires support for per-token dynamic scaling or fixed calibration scale tensors in runtime engines.
Context: INT8 matrix multiplications on NVIDIA Ampere/Ada Lovelace tensor cores.
View missionAWQ: Activation-aware Weight Quantization for On-Device LLM Compression and Acceleration
Published the MLSys Best Paper proving that protecting the top 1% of salient weight channels based on activation magnitudes enables hardware-friendly 4-bit weight-only quantization without costly backpropagation or retraining.
Requires representative calibration datasets to accurately identify activation outliers across specialized domains.
Context: 4-bit integer weights, FP16 activations, evaluated across LLaMA-1/2, Vicuna, StarCoder.
View mission