Microsoft Research
In plain wordsMicrosoft's research division, with laboratories on several continents, whose scientists have produced influential work on speech, vision and language AI.
The research organization of Microsoft, with laboratories in Redmond, Cambridge (UK), Beijing, New York, Montreal, Bangalore and elsewhere. Its contributions to AI include residual networksResNet, prop. n.: A family of image-recognition networks from 2015 whose shortcut connections allowed far deeper networks than before, over 100 layers, and became a standard building block across AI., developed at Microsoft Research Asia, which won the 2015 ImageNet competition and made very deep networks trainable; a conversational speech recognizer reported in 2016 to match professional transcribers on the Switchboard benchmark; the DeepSpeed library for large-scale training; and the Phi series of small language modelssmall language model, n.: A language model built to be compact enough to run cheaply and quickly, even on a laptop or phone, trading some knowledge and skill for speed, privacy and lower cost. trained largely on curated and synthetic text. Its researchers' early study of GPT-4, Sparks of AGISparks of AGI, prop. n.: A 2023 Microsoft research paper arguing that an early version of GPT-4 showed the first glimmers of general, humanlike intelligence. Its claim provoked sharp debate about how to judge AI abilities. (2023), was widely read and widely disputed. Microsoft Research is distinct from the Microsoft AI consumer division created in 2024 and from Microsoft's commercial partnership with OpenAIOpenAI, prop. n.: The American company that makes ChatGPT and the GPT models. It began in 2015 as a nonprofit research lab and is now a for-profit company overseen by a nonprofit foundation., begun in 2019.
Origin
Founded in 1991 after Nathan Myhrvold, then a senior Microsoft technology executive, persuaded Bill Gates to create a basic-research laboratory; Richard (Rick) Rashid left Carnegie Mellon University to be its first director. Microsoft Research Cambridge opened in 1997 and Microsoft Research China, later renamed Microsoft Research Asia, in Beijing in 1998.
Related
- ResNetResNet, prop. n.: A family of image-recognition networks from 2015 whose shortcut connections allowed far deeper networks than before, over 100 layers, and became a standard building block across AI.
- Sparks of AGISparks of AGI, prop. n.: A 2023 Microsoft research paper arguing that an early version of GPT-4 showed the first glimmers of general, humanlike intelligence. Its claim provoked sharp debate about how to judge AI abilities.
- small language modelsmall language model, n.: A language model built to be compact enough to run cheaply and quickly, even on a laptop or phone, trading some knowledge and skill for speed, privacy and lower cost.
- OpenAIOpenAI, prop. n.: The American company that makes ChatGPT and the GPT models. It began in 2015 as a nonprofit research lab and is now a for-profit company overseen by a nonprofit foundation.
- Mustafa SuleymanMustafa Suleyman, prop. n.: A British entrepreneur who co-founded DeepMind and Inflection AI, leads Microsoft's effort to build its own advanced AI models, and has written about the difficulty of containing powerful AI.
- ImageNetImageNet, prop. n.: A huge collection of more than 14 million labeled photos, sorted into object categories, whose yearly recognition contest from 2010 to 2017 sparked the deep learning boom.
Sources
- He, K., Zhang, X., Ren, S. & Sun, J. (2016). Deep Residual Learning for Image Recognition. Proceedings of CVPR 2016.
Mentioned in
AutoGenAutoGen, prop. n.: A free Microsoft software framework for building teams of AI agents that solve problems by messaging one another, writing and running code, and asking people for help when needed., Inflection AIInflection AI, prop. n.: An AI company founded in 2022 that made Pi, a friendly personal chatbot. In 2024 its leaders and most of its staff left to join Microsoft., Sparks of AGISparks of AGI, prop. n.: A 2023 Microsoft research paper arguing that an early version of GPT-4 showed the first glimmers of general, humanlike intelligence. Its claim provoked sharp debate about how to judge AI abilities.