As human beings, we use our voices for conversation. When we interact with voice interfaces, therefore, our natural instinct is to apply the same rules that we would to a human conversation. We expect to be understood, but more than this, we expect the entity we're conversing with to remember the history of our conversation and understand the context of any following remarks. For some time, major search companies like Google and Bing have worked to teach their search engines to understand queries in natural language. Natural language search queries are queries that sound natural spoken aloud, such as, "How high is the Empire State building?" They often begin with question words ("When…?" "How…?" "Why…?"), contain stop words ("a", "the", "of", "for") and full sentences.

I'm attending the Microsoft Imagine Cup, a global competition that empowers young computer science students to team up and use their creativity, passion and knowledge of technology and quantitative skills to create applications that improve the world in which we live. This year's event is the 16th annual event that includes 49 teams from around the world. Finalists will be vying for the chance to win up to $100K plus a mentoring session with Microsoft CEO Satya Nadella. You can watch the winner being introduced live on July 25th at 9:00 am PT! I had the opportunity to talk with the finalists, watch their demonstrations and ask questions to better understand their knowledge of the space in which they are working.

NEW DELHI – India's first training center for teachers of Japanese was officially opened Monday in the capital, New Delhi. The inauguration ceremony for the center, which is a joint project involving the Indian Ministry of External Affairs and the Japanese Embassy with the support of the Japan Foundation, was attended by Japanese Ambassador to India Kenji Hiramatsu. The ambassador noted in a speech that demand for learning Japanese is growing significantly as the Japan-India relationship flourishes, leading to an increasing number of employment opportunities in Japanese companies in the country. "The number of Japanese companies is increasing every year and is now about four times the number seen 10 years ago. These companies require Indians who can speak the Japanese language, in order to act as bridges between their Indian subsidiaries and headquarters in Japan," he said.

IN DECEMBER 1989 Guido van Rossum, a Dutch computer scientist, set himself a Christmas project. Irked by shortcomings in other programming languages, he wanted to build his own. First, it should be easy to read. Rather than sprawling over line-endings and being broken up by a tangle of curly braces, each chunk would be surrounded with indented white space. Second, it should let users create their own packages of special-purpose coding modules, which could then be made available to others to form the basis of new programs.

An Irish video games company has snapped up a Japanese business as it continues an ambitious acquisition spree. The Dublin-based Keywords Studios bought Yokozuna Data from Silicon Studio Corporation, a Japan-based games engine developer. Yokozuna Data makes artificial intelligence and machine learning technologies. Keywords acquired the business and assets of Yokozuna Data for $1.5 million (€1.3 million). Andrew Day, chief executive of Keywords, said that it had been looking for opportunities to establish a foothold in the area of video games analytics "for a few years".

In spite of the fact the EdTech market is predicted to reach $252 Billion by 2020 and that over $1 billion was invested in the sector in the US this year alone, many promising initiatives in the sector fail to make any real lasting impact. This is something that Vikas Pota is looking to change by putting education at the heart of the technology discussion. After eight years as CEO of the Varkey Foundation, Pota has now been appointed as Group Chief Executive of Tmrw Digital, Sunny Varkey's new holding company which will oversee all his EdTech businesses and investments, as well as set up a research institute dedicated to finding ways of leveraging technology to democratize access to education. My role at the Foundation taught me how incredibly complex EdTech is. You have this fundamental conflict between a really innovative group of technology entrepreneurs and a schooling system that dates back to the first industrial revolution." Tmrw Digital is a business but its mission will still help to advance the broader societal impact of the Varkey Foundation. It has inherited a portfolio of EdTech investments, with five dedicated funds in Silicon Valley having funded between 400-500 EdTech start-ups so far. You don't make a quick buck in EdTech. Everything we do has to be built on a foundation of sound educational research. Educating a child is a process that takes place over many years, so why do we expect EdTech to deliver learning outcomes instantly?"

Recently we saw a new "Master algorithm" that could be used to create the first generation of super intelligent machines, and now a team of researchers from Maryland, USA, announced this week that they've invented a general Artificial Intelligence (AI) way for machines to identify and process 3D images that doesn't require humans to go through the tedium of inputting specific information that accounts for each and every instance, scenario, difference, change and category that could crop up, and they claim it's a world first, even though it follows on from a not too dissimilar breakthrough from Google DeepMind whose own platform, Alpha Zero, recently taught itself a mix of board games including chess to a grand master level, in just four hours. All that said though it's important to say here that while doing away with huge volumes of training data used to train today's AI's is a staggering breakthrough of potential epic proportions this announcement doesn't yet herald the emergence of fabled Artificial General Intelligence (AGI) which is still a little way off and which Google Deepmind published an architecture for last year. However, this new development, albeit a distant cousin of AGI for now, will no doubt have an influence on the speed at which the first AGI's now emerge. This is actually a huge deal for the technology sector, and it's a massive step for the case of general AI versus specific AI, and once fine tuned, this development will have the power to shape and change how everyone, from police and intelligence officials, to retail marketers and medical professionals, go about their daily business. Currently, neural networks, defined as those computing systems that are aimed at mimicking how humans think and make decisions, are only as good as the information that's inputted.

Technology experts have stressed the need for schools to prepare students for tomorrow's jobs by increasingly teaching them artificial intelligence (AI) skills to an adequate level. Necip Ozyucel, Cloud and Enterprise Business Group Lead, Microsoft Gulf, said the latest World Economic Forum's'Future of Jobs' report, released in January this year showed that two in three children starting school this year are destined for professional roles yet to be created. "There is an urgent need to recognise that many of the jobs referred to by the World Economic Forum, are supported by artificial intelligence," Ozyucel told Khaleej Times in an exclusive interview. A recent survey by McKinsey & Company on specialist education practice - where 2,000 students, 2,000 teachers and 70 thought leaders across the Americas, Europe and Asia including the Gulf region were interviewed - found that students would be better prepared for'future jobs' if they shored up their soft skills, particularly the social and emotional. And only 42 per cent of employers consider today's graduates as having developed those attributes to an adequate level, according to Ozyucel, "When trying to determine the best approaches to bridge these soft-skills gaps, there is evidence in both studies that students and teachers alike favour collaborative learning scenarios. Mentor-led class discussions or group-learning scored consistently high as the most effective methods of learning," said Ozyucel.

One day, on my way past the outskirts of Kabukichō – Tokyo's red-light district, infamously depicted in the Yakuza games – I spot a curious advertisement. At first glance, it looks like nothing out of the ordinary: a woman cheerfully donning a VR headset, with kanji lettering welcoming passersby to come in and try the technology for themselves. As my eyes wander to the logo in the corner, I realise that the poster is promoting Soft On Demand – one of Japan's biggest porn, or "AV" (adult video), companies. A stone's throw away is Bandai Namco's massive VR Zone complex, an indoor, 38,000 sq ft all-VR theme park that opened just over a year ago. And further south, on the artificial island of Odaiba, Sega recently cleared out a massive room in its Joypolis amusement park to make space for Zero Latency VR, a "warehouse scale, free-roam, multiplayer virtual reality entertainment" where a team of zombie hunters are equipped with "military-grade" motion-tracking backpacks and let loose on the undead with an arsenal of plastic firearms.

In typical machine learning tasks and applications, it is necessary to obtain or create large labeled datasets in order to to achieve high performance. Unfortunately, large labeled datasets are not always available and can be expensive to source, creating a bottleneck towards more widely applicable machine learning. The paradigm of weak supervision offers an alternative that allows for integration of domain-specific knowledge by enforcing constraints that a correct solution to the learning problem will obey over the output space. In this work, we explore the application of this paradigm to 2-D physical systems governed by non-linear differential equations. We demonstrate that knowledge of the partial differential equations governing a system can be encoded into the loss function of a neural network via an appropriately chosen convolutional kernel. We demonstrate this by showing that the steady-state solution to the 2-D heat equation can be learned directly from initial conditions by a convolutional neural network, in the absence of labeled training data. We also extend recent work in the progressive growing of fully convolutional networks to achieve high accuracy (< 1.5% error) at multiple scales of the heat-flow problem, including at the very large scale (1024x1024). Finally, we demonstrate that this method can be used to speed up exact calculation of the solution to the differential equations via finite difference.