Microsoft Research – Emerging Technology, Computer, and Software Research

Christopher Bishop, Technical Fellow and Director, Microsoft Research AI4Science

AI4Science to empower the fifth paradigm of scientific discovery

confidential computing hero

Confidential Containers: Verifiably secure computation in the cloud

Three bar plots. The first plot shows that the model size of XTC-BERT is 32 times smaller than that of BERT, and two dots show the accuracy of BERT and XTC-BERT, which are 83.95 and 83.44, respectively. The second one shows that INT8 using ZeroQuant can be 2.6 times faster than Baseline with FP16 using PyTorch and ZeoQuant can reduce the number of GPUs for inference from 2 to 1, which in total provides 5.2 times efficiency. It also shows that ZeroQuant has 50.4 accuracy compared to 50.5 using Baseline PyTorch. The third plot shows that ZeroQuant is more than 5000 times cheaper than baseline to compress a model, and the accuracy of ZeroQuant is 42.26 compared to 42.35 of baseline.

DeepSpeed Compression: A composable library for extreme compression and zero-cost quantization

danah boyd featured in the “People of ACM”

ACMAug 9, 2022

Make greenhouse-gas accounting reliable — build interoperable systems

NatureJul 26, 2022

Using AI and machine learning to kickstart climate change fightback

IT ProJul 19, 2022

Bichlien Nguyen is one of the ACS 2022 Talented Twelve

C&ENJul 15, 2022

See all news & awards