Collaborating Authors

AlphaFold advances protein folding research


The grand challenge of protein folding hit the news this week when it was announced that the latest version of DeepMind's AlphaFold system had predicted protein structures with very high accuracy in CASP's 2020 experiment. Proteins are large, complex molecules, and the shape of a particular protein is closely linked to the function it performs. The ability to accurately predict protein structures would enable scientists to gain a greater understanding of how they work and what they do. This new version of AlphaFold builds on the initial system, which you can read about in this paper. The associated code is available here.

Faster, Smaller and More Accurate Edge AI Using Deeplite and Andes Technology Software + Hardware


MONTREAL, CANADA and HSINCHU, TAIWAN – December 3, 2020 – The push for low-power and low-latency deep learning models, computing hardware, and systems for artificial intelligence (AI) inference on edge devices continues to create exciting new opportunities. There has been unprecedented interest from industry stakeholders in the development of hardware and software solutions for on-device deep learning, also called Edge AI. This has already begun to yield progress on hallmark applications such as keyword spotting in audio classification, anomaly detection and, in this case, person detection in computer vision applications. Specifically, tinyML, the branch of machine learning tailored to ultra-low power systems, holds tremendous promise. The efficiency of proposed solutions (milliwatt or even microwatt power consumption) and vast applicability and deployment of such devices in real-world settings will lead to over 100 billion IoT sensors and devices expected to ship over the next 5 years 1.

Top Deep Learning Based Time Series Methods


The components of time-series can be as complex and sophisticated as the data itself. With every passing second, the data obtained multiplies and modelling becomes tricky. For instance, social media platforms, the data handling chores get worse with their increasing popularity. Twitter stores 1.5 petabytes of logical time series data and handles 25K query requests per minute. There are more critical applications of time series modelling, such as IoT and on various edge devices.

Welcome to the Artificial Intelligence Incident Database


Intelligent systems are currently prone to unforeseen and often dangerous failures when they are deployed to the real world. Much like the transportation sector before it (e.g., FAA and FARS) and more recently computer systems, intelligent systems require a repository of problems experienced in the real world so that future researchers and developers may mitigate or avoid repeated bad outcomes. The initial set of more than 1,000 incident reports have been intentionally broad in nature. You are invited to explore the incidents collected to date, view the complete listing, and submit additional incident reports. The database is a constantly evolving data product and collection of applications.

6 things that could help advance AI adoption in radiology - Watson Health Perspectives


Radiology has always been on the cutting edge of technology in medicine, and artificial intelligence (AI) has made some progress within this specialty. But AI is at an inflection point. To make use of AI as a more accessible, trusted reality, radiology will need six critical components: 1) Reassurance that AI will never replace a physician With the growing body of evidence that shows the extraordinary potential of AI in healthcare, I find it interesting that widespread adoption remains slow. Although I hear multiple reasons for the reluctance of radiologists to embrace AI, one of the most common that I hear is the misperception that AI is going to put us radiologists out of business. This simply isn’t realistic. At the end of the day, algorithms address technology; radiologists take care of people, with all the inherent responsibilities, nuances and complexities. Algorithms aren’t designed to do that. However, It’s also increasingly unrealistic to expect radiologists to manage…

Smart, AI-powered virtual assistants help drive sales and convert leads


Poor customer experiences are the primary cause for low conversion rates, lost leads and customer churn. In fact, a customer who suffered a third-rate experience will tell five times as many people about what happened than if they had a great experience. These bad impressions tend to stick with people and influence their future decisions for months or even years later. As a result, frustrated Sales Managers and Marketers often blame these sour interactions on substandard or poorly deployed software. Similarly, many marketers prefer today to keep things as hands-on as possible.

Why Intel believes confidential computing will boost AI and machine learning


Companies are collecting increasing amounts of data, a trend that is driving the development of better analytical tools and tougher security. Analysis and security are now converging as confidential computing prepares to deliver a critical boost to artificial intelligence. Intel has been investing heavily in confidential computing as a way to expand the amount and types of data companies will manage through cloud services. According to Intel Fellow Ron Perez, who works on security architecture with the Intel Data Center Group, the company believes the emerging security standard will allow enterprises and large organizations to explore new ways to share the data needed to fuel AI and machine learning. "We see this as a long-term effort," Perez said.

A Gentle Introduction to the Rectified Linear Unit (ReLU)


In a neural network, the activation function is responsible for transforming the summed weighted input from the node into the activation of the node or output for that input. The rectified linear activation function or ReLU for short is a piecewise linear function that will output the input directly if it is positive, otherwise, it will output zero. It has become the default activation function for many types of neural networks because a model that uses it is easier to train and often achieves better performance. In this tutorial, you will discover the rectified linear activation function for deep learning neural networks. A Gentle Introduction to the Rectified Linear Activation Function for Deep Learning Neural Networks Photo by Bureau of Land Management, some rights reserved. A neural network is comprised of layers of nodes and learns to map examples of inputs to outputs. For a given node, the inputs are multiplied by the weights in a node and summed together.

Using AI to achieve environmental, social and governance goals


Environmental, social and governance (ESG) investments have fast become an important area of interest. It was estimated that sustainable investments amounted to some $30 trillion in 2018, up by 34 per cent from 2016. Indeed, investors (and our societies in general) are increasingly keen to understand whether and by what means businesses are being environmentally and socially responsible and governed. Simultaneously, boards and managements have become cognisant that ESG is crucial to the long-term survival of their companies. Small wonder, then, that some 90 per cent of investors globally already have in place, or have plans to develop, specific ESG investment policies.

Warp Speed, Atomic Structure, and the Future of Vaccines


If the first vaccines against Covid-19 really do start coming online in a couple of weeks, that'll be a blazingly fast scientific achievement--from new virus to new vaccine in just about 12 months, faster than ever before, and using a new vaccine technology, too. And also only sort of true, because the path of the two vaccines likeliest to become available first, one from the pharmaceutical companies Pfizer and BioNTech and one from Moderna, began long before people started getting sick in Wuhan in December 2019. Like all scientific discoveries, that path has many trailheads. One of them is the lab of John Mascola, director of the Vaccine Research Center at the National Institute of Allergy and Infectious Diseases. He didn't come up with the idea of using genetic material to make vaccines, but he and collaborators around the US spent years trying to direct those efforts against coronaviruses, the family that includes SARS-CoV-2, the cause of Covid-19.