In summer 2013, I interviewed for a lead role in the data science and analytics team at tech-for-good company JustGiving. During the interview, I said I planned to deliver batch machine learning, graph analytics and streaming analytics systems, both in-house and in the cloud. A few years later, my former boss Mike Bugembe and I were both presenting at international conferences, winning awards and becoming authors! Here is my story, and what I learnt on the journey -- plus my recommendations for you. I've always been interested in artificial intelligence (AI), machine learning (ML) and natural language processing (NLP).

Deep neural networks (DNNs) have been shown to outperform conventional machine learning algorithms across a wide range of applications, e.g., image recognition, object detection, robotics, and natural language processing. However, the high computational complexity of DNNs often necessitates extremely fast and efficient hardware. The problem gets worse as the size of neural networks grows exponentially. As a result, customized hardware accelerators have been developed to accelerate DNN processing without sacrificing model accuracy. However, previous accelerator design studies have not fully considered the characteristics of the target applications, which may lead to sub-optimal architecture designs. On the other hand, new DNN models have been developed for better accuracy, but their compatibility with the underlying hardware accelerator is often overlooked. In this article, we propose an application-driven framework for architectural design space exploration of DNN accelerators. This framework is based on a hardware analytical model of individual DNN operations. It models the accelerator design task as a multi-dimensional optimization problem. We demonstrate that it can be efficaciously used in application-driven accelerator architecture design. Given a target DNN, the framework can generate efficient accelerator design solutions with optimized performance and area. Furthermore, we explore the opportunity to use the framework for accelerator configuration optimization under simultaneous diverse DNN applications. The framework is also capable of improving neural network models to best fit the underlying hardware resources.

Privacy advocates have placed an unfair stigma on machine learning. Despite what you may have heard through the mass media, ML is not some fiendish tool for invading people's privacy. Regardless, now that European Union's General Data Protection Regulation has taken effect, there's an even stronger scrutiny of ML applications in target marketing, customer engagement, experience optimization and other use cases that touch personally identifiable information, or PII. But in fact, ML is becoming a key element in how organizations manage compliance with GDPR and other privacy mandates. The core of ML's role in GDPR compliance is in its use as a tool for discovering, organizing, curating and controlling enterprise PII assets across complex, distributed application environments.

Reuters reports the former Texas congressman once belonged to Cult of the Dead Cow, an influential group "jokingly named after an abandoned Texas slaughterhouse." While there's no evidence that O'Rourke really got his hands dirty with what we'd consider nefarious "hacking," he was a member, which might help explain some of the policies he could champion during a presidential run. Music royalties can be confusing, but this piece breaks down what's happening with Spotify, Google, Pandora and Amazon. And most importantly, why it matters. A profile of Demis Hassabis, a co-founder of DeepMind, examines the origins of the AI startup and asks how much longer it can retain its independence from Google.

University of Toronto Engineering researchers have built a set of magnetic'tweezers' that can position a nano-scale bead inside a human cell in three dimensions with unprecedented precision. The nano-bot has already been used to study the properties of cancer cells, and could point the way toward enhanced diagnosis and treatment. Professor Yu Sun and his team have been building robots that can manipulate individual cells for two decades. Their creations have the ability to manipulate and measure single cells--useful in procedures such as in vitro fertilization and personalized medicine. Their latest study, published today in Science Robotics, takes the technology one step further.

We're proud to add yet another AI award to our awards display case at the Conversica HQ! This time, our Conversica Conversational AI Assistant for Business was one of two technologies selected to win an AI award in the Self-Aware category of the 2019 Artificial Intelligence Excellence Awards. The other winner was ignio – a cognitive automation platform. It's flattering to keep company with other technologies pushing the boundaries of today's artificial intelligence for business. All in all, 20 other companies won awards in the 2019 Artificial Intelligence Excellence Awards, including other business-focused AI companies, products and people who are bringing Artificial Intelligence (AI) to life and applying it to solve real problems.

I love shopping at a Nordstrom store, but hate shopping on the Nordstrom website. Unless I know exactly what I want, like my favorite shade of lipstick, I have a time trying to find what I might want. Take a search for a blue dress. Nordstrom served up nearly 3,000 options and after narrowing it down to casual styles, I had a mere 1,000 to go through. Included in that search were jumpsuits (sorry, not a dress), plus sizes and maternity dresses.

The first five days or so of SXSW in Austin are always dedicated to the "interactive" portion of the festival. The city's downtown streets swell with lanyard-laden "entrepreneurs" and "founders" wearing that familiar uniform of T-shirts screen-printed with their company's clever logo, an outfit made professional by throwing a blazer over the ensemble. They bounce from panel to panel and branded "house" to branded "house" (this year, on scooters, so many scooters) hawking their new apps and software products, each promising to be more revolutionary and life-changing and utterly necessary than the next. For years, the unspoken question at the conference seemed to be which company will become SXSW famous, like Persicope, Foursquare, or, most memorably, Twitter? But this year, on the opening Friday of SXSW, Democratic presidential hopeful Elizabeth Warren unleashed a manifesto titled "Here's How We Can Break Up Big Tech," and a new question burst onto the scene: What do you think of Warren's proposal?

TANYA OTT: I'm Tanya Ott and this is the Press Room, where we talk about the issues that are or should be important to your business. In October, the Massachusetts Institute of Technology announced it's going to spend $1 billion dollars--that's with a big capital B--to create a new college focused on Artificial Intelligence.1 That is a huge investment. MIT says it's already raised two-thirds of the money and plans to start classes next fall. In announcing the school, MIT's president said he wants to "educate the bilinguals of the future."

Artificial intelligence (AI) is trending globally in commerce, science, health care, geopolitics, and more areas. Deep learning, a subset of machine learning, is the lever that launched the worldwide rush--an area of strategic interest for researchers, scientists, visionary CEOs, academics, geopolitical think tanks, pioneering entrepreneurs, astute venture capitalists, strategy consultants, and management executives from companies of all sizes. Yet in the midst of this AI renaissance, is a relatively fundamental unsolvable problem with machine learning that is not commonly known, nor frequently discussed outside of the small cadre of philosophers, and artificial intelligence experts. A global research team of researchers have recently demonstrated that machine learning has an unsolvable problem, and published their findings in Nature Machine Intelligence in January 2019. Researchers from Princeton University, the University of Waterloo, Technion-IIT, Tel Aviv University, and the Institute of Mathematics of the Academy of Sciences of the Czech Republic, proved that AI learnability cannot be proved nor refuted when using the standard axioms of mathematics.