Newly discovered principle reveals how adversarial training can perform robust deep learning
| Zeyuan Allen-Zhu and Yuanzhi Li
In machine learning, adversarial examples usually refer to natural inputs plus small, specially crafted perturbations that can fool the model into making mistakes. In recent years, adversarial examples have been repeatedly discovered in deep learning applications, causing public concerns about…
In the news | Official Microsoft Blog
Microsoft launches initiative to help 25 million people worldwide acquire the digital skills needed in a COVID-19 economy
Around the world, 2020 has emerged as one of the most challenging years in many of our lifetimes. In six months, the world has endured multiple challenges, including a pandemic that has spurred a global economic crisis. As societies reopen,…
In the news | MSPoweruser
Microsoft Research develop amazing photo restoration algorithm
Microsoft Research team Ziyu Wan, Bo Zhang, and more have developed a new AI-based algorithm for restoring old photos that suffer from severe degradation through a deep learning approach.
It's critical to make cautious and informed decisions about changes that impact users-- especially during times of extra stress. Prompted by the Covid-19 crisis, Microsoft's ExP team shares their thinking about effective use of A/B testing to mitigate uncertainty during…
In the news | Medium
‘Colorblind’ Tech is Killing Us: Why COVID-19 Tech Must Focus on Equity
COVID-19 has taken the lives of more than 120,000 people in the United States and debilitated millions more since arriving on our shores. While anyone can contract the virus, there is a pattern to who gets sick and dies from…
Hi there, Steven from the Soundscape Team here again. One of the most common questions we receive in customer feedback is how Microsoft Soundscape appears to announce callouts slightly differently when travelling in a vehicle or on public transport compared…
Microsoft Research EMEA and Latin America PhD Awards springboard new ideas across intercontinental research
| Clare Morgan
The 2020 EMEA and Latin America PhD Award recipients are undertaking research central to advancing today’s technology—training artificial agents to learn faster in new situations, designing new scalable machine learning algorithms for continual and federated learning, and several other compelling…
With their work-from-home guidance stretching into another month, Information Workers (IWs) are facing numerous challenges with their ability to communicate effectively with their teams. Since collaboration is central to IWs daily productivity and engagement, a team of user researchers studied…
Enhancing your photos through artificial intelligence
The amount of visual data we accumulate around the world is mind boggling. However, not all the images are captured by high-end DSLR cameras, and very often they suffer from imperfections. It is of tremendous benefit to save those degraded…