In the news | Adaptive Biotech

Imagine if a simple blood test could provide accurate and early diagnosis of thousands of different diseases simultaneously? That is no easy task, but it may be possible using new technology that can leverage how our body naturally detects diseases…

In the news | Microsoft On the Issues

“What can we do?” Those words continued to echo as I returned home from the World Economic Forum in Davos in early 2019. The director of a major news organization asked me that question point blank after we reviewed several deepfake…

In the news | Microsoft Transform

B.J. Moore knew he had to stick to his gut. He had just made a crucial choice as the new chief information officer at Providence, one of the nation’s largest health systems, spanning 51 hospitals and 1,085 clinics in seven…

Microsoft Research Blog

| Batu Aytemiz, Mikhail Jacob, Sam Devlin, and Katja Hofmann

In video games, nonplayer characters, bots, and other game agents help bring a digital world and its story to life. They can help make the mission of saving humanity feel urgent, transform every turn of a corner into a gamer’s…

Articles

Developing usable and valuable features is not much use if the feature is going to be turned off as soon as it hits the enterprise environment. They implored us to think of the stakeholder cycles that each new feature must…

In the news | Microsoft Customer Stories

Children’s Mercy Kansas City, an award-winning hospital and research institute, manages one of the leading genome sequencing centers in the United States. To better support researchers, Children’s Mercy is working with Microsoft and Pacific Biosciences to create a scalable, sharable,…

In the news | WIRED

A chess program that learns from human error might be better at working with people or negotiating with them. It took about 50 years for computers to eviscerate humans in the venerable game of chess. A standard smartphone can now…

In the news | The Economist

But as the software improves, access to material gets harder.

An visual comparing public image classification APIs to the proposed denoised smoothing framework applied to a public image classification API. In the first sequence, an image of an elephant is input into a public image classification API, represented by an arrow leading to a gray square labeled as such. An arrow from the square points to a correct prediction of elephant, enclosed in a green square, with the words “no robustness guaranteed” under it. In the second sequence, an arrow points from an image of an elephant to six noisy copies of the image. An arrow then points from the copies to a square labeled “Custom-trained Denoiser,” which outputs six clean versions of the images. An arrow points from the clean copies to a square labeled “Public Image Classification API.” The classifier provides predictions for each copy, of which four correctly identify their respective images as elephant. Adjacent to the predictions is a pie chart labeled “Majority Rules!” with one-third of the pie in red and two-thirds in green. Arrows point to the output of the process: a final prediction of elephant, enclosed in a green square, and a strong robustness guarantee, denoted by the words “certified radius” enclosed in a green square.
Microsoft Research Blog

| Hadi Salman

Editor’s note: This post and its research are the result of the collaborative efforts of a team of researchers comprising former Microsoft Research Engineer Hadi Salman (opens in new tab), CMU PhD student Mingjie Sun (opens in new tab), Researcher…