Microsoft Research India

Microsoft Research Lab – India

新闻与深度文章

新闻报道 | The Economic Times

Microsoft Research India is creating tools to help preserve fast disappearing languages 

In 2010, Bo, a language of the Andaman I…

新闻报道 | Microsoft Stories India

Microsoft Research project helps languages survive — and thrive 

A woman named Boa Sr was the last link t…

新闻报道 | Future Tense

An A.I. Translation Tool Can Help Save Dying Languages. But at What Cost? 

A.I. language tools depend on data—and l…

a couple of people that are standing in the grass
Microsoft Research India Podcast

Road safety is a very serious public hea…

A flowchart showing the DoWhy library process. Input Data and Domain Knowledge are injected into the library, where they go through four steps: Model causal mechanisms; Identify target estimands; Estimate causal effect; and Refute estimate. The process produces the output labelled Causal effect.
微软研究院博客

| Emre Kiciman 和 Amit Sharma

Identifying causal effects is an integra…

a couple of people posing for the camera
Microsoft Research India Podcast

Sridhar Vedantham: Welcome to the MSR In…

A flowchart showing inputs pre-processed before being fed into large language models including GPT-3, Codex, and others. The post-process output is returned to the end-user for verification. If they find the output incorrect, it is edited by them, and the learning is fed back into the pre-process and post-process mechanisms to improve them further.
微软研究院博客

| Rahul Sharma, Suresh Parthasarathy, Sriram Rajamani, Nagarajan Natarajan, 和 Arun Iyer

Large pre-trained language models such a…

a couple of people posing for the camera
Microsoft Research India Podcast

 Episode 011 | January 18, 2022 Keratoc…

新闻报道 | Neowin

EzPC: Microsoft’s attempt to enhance data security in AI model validation 

Those who have worked in the domain of d…