Microsoft Research Blog

The Microsoft Research blog provides in-depth views and perspectives from our researchers, scientists and engineers, plus information about noteworthy events and conferences, scholarships, and fellowships designed for academic and scientific communities.

AI for AI: Metareasoning for modular computing systems

A new document in a word processor can be a magical thing, a blank page onto which thoughts and ideas are put forth as quickly as we can input text. We can select words and phrases to underline and highlight and add images, shapes, and bulleted lists, and when we need editorial help, we can run a grammar and spell checker. The experience can feel so seamless at times that perhaps we don’t give much…

February 2020

Microsoft Research Blog

Turing-NLG: A 17-billion-parameter language model by Microsoft

This figure was adapted from a similar image published in DistilBERT. Turing Natural Language Generation (T-NLG) is a 17 billion parameter language model by Microsoft that outperforms the state of the art on many downstream NLP tasks. We present a demo of the model, including its freeform generation, question answering, and summarization capabilities, to academics for feedback and research purposes.  – This summary was generated by the Turing-NLG language model itself. Massive deep learning…

February 2020

Microsoft Research Blog

ZeRO & DeepSpeed: New system optimizations enable training models with over 100 billion parameters

The latest trend in AI is that larger natural language models provide better accuracy; however, larger models are difficult to train because of cost, time, and ease of code integration. Microsoft is releasing an open-source library called DeepSpeed, which vastly advances large model training by improving scale, speed, cost, and usability, unlocking the ability to train 100-billion-parameter models. DeepSpeed is compatible with PyTorch. One piece of that library, called ZeRO, is a new parallelized optimizer…

February 2020

Microsoft Research Blog

Project Rocket platform—designed for easy, customizable live video analytics—is open source

Thanks to advances in computer vision and deep neural networks (DNNs) in what can arguably be described as the golden age of vision, AI, and machine learning, video analytics systems—systems performing analytics on live camera streams—are becoming more accurate. This accuracy offers opportunities to support individuals and society in exciting ways, like informing homeowners when a package has been delivered outside their door, allowing people to give their pets the attention they need when out…

January 2020

Microsoft Research Blog

When bias begets bias: A source of negative feedback loops in AI systems

Is bias in AI self-reinforcing? Decision-making systems that impact criminal justice, financial institutions, human resources, and many other areas often have bias. This is especially true of algorithmic systems that learn from historical data, which tends to reflect existing societal biases. In many high-stakes applications, like hiring and lending, these decision-making systems may even reshape the underlying populations. When the system is retrained on future data, it may become not less but more detrimental to…

January 2020

Microsoft Research Blog

By making text-based games more accessible to RL agents, Jericho framework opens up exciting natural language challenges

You’re in a field. In front of you, there’s a white house. The door is boarded shut. The immediate challenge—investigate the house. The game—Zork I: The Great Underground Empire, a treasure-seeking adventure in which you’ll encounter monsters, a thief, and other interesting characters along the way. As a player of this text-based game, you string together simple commands of only several words, like “walk to the house.” Once there, you type a series of commands,…

January 2020

Microsoft Research Blog

animation of reinforcement learning agents beating human competitors in Atari

Finding the best learning targets automatically: Fully Parameterized Quantile Function for distributional RL

Reinforcement learning has achieved great success in game scenarios, with RL agents beating human competitors in such games as Go and poker. Distributional reinforcement learning, in particular, has proven to be an effective approach for training an agent to maximize reward, producing state-of-the-art results on Atari games, which are widely used as benchmarks for testing RL algorithms. Because of the intrinsic randomness of game environments—with the roll of the dice in Monopoly, for example, you…

December 2019

Microsoft Research Blog

FastSpeech: New text-to-speech model improves on speed, accuracy, and controllability

FastSpeech: New text-to-speech model improves on speed, accuracy, and controllability

Text to speech (TTS) has attracted a lot of attention recently due to advancements in deep learning. Neural network-based TTS models (such as Tacotron 2, DeepVoice 3 and Transformer TTS) have outperformed conventional concatenative and statistical parametric approaches in terms of speech quality. Neural network-based TTS models usually first generate a mel-scale spectrogram (or mel-spectrogram) autoregressively from text input and then synthesize speech from the mel-spectrogram using a vocoder. (Note: the Mel scale is used…

December 2019

Microsoft Research Blog

illustrated palm tree on an island

Provable guarantees come to the rescue to break attack-defense cycle in adversarial machine learning

Artificial intelligence has evolved to become a revolutionary technology. It is rapidly changing the economy, both by creating new opportunities (it’s the backbone of the gig economy) and by bringing venerable institutions, like transportation, into the 21st century. Yet deep at its core something is amiss, and more and more experts are worried: the technology seems to be extremely brittle, a phenomenon epitomized by adversarial examples. Adversarial examples exploit weaknesses in modern AI. Today, most…

December 2019

Microsoft Research Blog

Game of Drones simulation

Game of Drones at NeurIPS 2019: Simulation-based drone-racing competition built on AirSim

Drone racing has transformed from a niche activity sparked by enthusiastic hobbyists to an internationally televised sport. In parallel, computer vision and machine learning are making rapid progress, along with advances in agile trajectory planning, control, and state estimation for quadcopters. These advances enable increased autonomy and reliability for drones. More recently, the unmanned aerial vehicle (UAV) research community has begun to tackle the drone-racing problem. This has given rise to competitions, with the goal…

December 2019

Microsoft Research Blog

Metalearned Neural Memory: Teaching neural networks how to remember

Memory is an important part of human intelligence and the human experience. It grounds us in the current moment, helping us understand where we are and, consequently, what we should do next. Consider the simple example of reading a book. The ultimate goal is to understand the story, and memory is the reason we’re able to do so. Memory allows us to efficiently store the information we encounter and later recall the details we’ve previously…

December 2019

Microsoft Research Blog

Optimistic Actor Critic avoids the pitfalls of greedy exploration in reinforcement learning

One of the core directions of Project Malmo is to develop AI capable of rich interactions. Whether that means learning new skills to apply to challenging problems, understanding complex environments, or knowing when to enlist the help of humans, reinforcement learning (RL) is a core enabling technology for building these types of AI. In order to perform RL well, agents need to do exploration efficiently, which means understanding when to try new things out and…

November 2019

Microsoft Research Blog