Dynamic programming (DP) solves a variety of structured combinatorial problems by iteratively breaking them down into smaller subproblems. In spite of their versatility, DP algorithms are usually non-differentiable, which hampers their use as a layer in neural networks trained by backpropagation. To address this issue, we propose to smooth the max operator in the dynamic programming recursion, using a strongly convex regularizer. This allows to relax both the optimal value and solution of the original combinatorial problem, and turns a broad class of DP algorithms into differentiable operators. Theoretically, we provide a new probabilistic perspective on backpropagating through these DP operators, and relate them to inference in graphical models. We derive two particular instantiations of our framework, a smoothed Viterbi algorithm for sequence prediction and a smoothed DTW algorithm for time-series alignment. We showcase these instantiations on two structured prediction tasks and on structured and sparse attention for neural machine translation.

When an Iranian drone flew into Israeli airspace this month, it set off a rapid series of strikes and counterstrikes that deepened fears over whether a new, catastrophic war was brewing in the Middle East. That flare-up ended quickly, if violently, with the drone destroyed and an Israeli jet downed after bombing sites in Syria. But the day of fighting drew new attention to how deeply Iran has embedded itself in Syria, redrawing the strategic map of the region. Tactical advisers from Iran's Islamic Revolutionary Guards Corps are deployed at military bases across Syria. Its commanders regularly show up at the front lines to lead battles.

Multilingual recipe sharing website Cookpad has appointed computer science researcher Dr Danushka Bollegala to work on recipe and ingredient recognition. Dr Bollegala has worked with Cookpad on a range of projects looking at natural language processing (NLP) and machine learning. Having studied for 13 years in Japan, the senior lecturer at the University of Liverpool has been named as senior fellow at Cookpad, where he is working with the Bristol-based team on an analysis of the language associated with the many thousands of recipes and ingredients on the site and their complexity. Dr Bollegala has been researching NLP, machine learning and data mining for more than ten years. He is working with the Cookpad team on a number of projects, with the primary focus on name recognition for ingredients.

Paladion, a global cyber defence company that provides AI-Driven Managed Detection and Response Services, will participate in Artelligence Forum as a Silver Sponsor. The event, which is scheduled to take place on April 25 and 26, 2018 in Dubai, is presented by Khaleej Times and MIT Sloan Management Review GCC to bring government leaders, business leaders and technology innovators together to explore the applications of Artificial Intelligence across different industries. Paladion is consistently rated and recognised by independent advisory firms for its capabilities in Artifical Intelligence in Cyber, Security Automation, and next generation cyber security. Speaking about AI in cybersecurity and this initiative by Khaleej Times and MIT Sloan Management Review GCC, Amit Roy, EVP and Regional Head EMEA, Paladion, said, "Cyberattackers are leveraging AI to increase the speed, volume, and sophistication of attacks. Enterprises must evolve to combat such threats. Enterprises today must recognize the ineffectiveness of a rules and signature-based defence on modern attacks, and adopt AI driven cyber security to defend against AI based cyber threats. Paladion's AI-driven MDR Service offers the only way to collect, analyze, and process the massive volume of data produced by AI-driven cyberattacks. We are excited to showcase our AI-driven cyber security service at the Artelligence Forum."

Ladies and Gentlemen, I am happy to be here today at the inauguration of the Wadhwani Institute of Artificial Intelligence. Let me begin by congratulating Romesh Wadhwani ji and Sunil Wadhwani ji,The Government of Maharashtra, and Mumbai University for coming together to make this Institute a reality. This is a prime example of how the public sector and the private sector can combine with good intention to build a world-class institute aimed at benefiting the poor. I have had various interactions with Indian diaspora across the globe during the last three and a half years or so. I have felt an intense desire, to make a contribution to India.

Jonathan Kirk, Co-Founder & Director of AI Strategy at systemsGo, shares his thoughts. The last thirty years have seen great changes in the world, from the emergence of a globally connected internet, to the pervasive influence of social media. Looking at the advances of the last 10 years, 100 years, and 1000 years, the pace of change is clearly accelerating. But the impact of Artificial Intelligence (AI) has the potential to dwarf the scale of all human progress so far. At the very least, it is likely to bring changes to human society as big or even bigger than mankind's other great transitions, from feudalism to agrarian society, or from a farming culture to industrialized city-dwelling society.

Taipei, Feb. 18 (CNA) 2018 has been dubbed by many tech and business insiders the year of artificial intelligence (AI). As Taiwan seeks to innovate on multiple scientific, cultural and humanitarian fronts, AI might just be the answer it has been looking for to make a name for itself on the global stage. CNA sat down with Ethan Tu (杜奕瑾), founder of Taiwan AI Labs, in January to talk about the potential for AI development on the island. According to Tu, who leads a team of 30 at the labs, Taiwan is well-positioned to be what he calls the "AI island of the world." He explained that Taiwan has several advantages when it comes to software and big data, and can utilize them to put AI to good use. Using the health care system as an example, Tu said that Taiwan currently has the most comprehensive national health insurance system in the world and is a world leader in digitalizing hospital records.

The most exciting thing about smart speakers is that they are all so utterly imperfect. Each of the mainstream examples – Amazon's Echos, the Google Home, Sonos's wide range and the new Apple HomePod – packs within it stunning technology that would have been unimaginable just a couple of years ago. But they're also full of downsides, making choosing one a matter of deciding what you want, not simply settling on the best. Deciding is a matter of picking which things you want – and which things you definitely don't. Some sound good, but are terrible at talking back to you.

A new study from a team of economists at the University of Toronto has concluded that China is steadily gaining on the United States in the field of artificial intelligence. The 2017 Association for the Advancement of Artificial Intelligence (AAAI), a worldwide conference that presents the achievements of the world's AI leaders, indicated that 23 ...

A new artificial intelligence (AI) project from tech conglomerate Alibaba could alleviate some of the myriad problems facing Chinese farmers in the pork industry. China is the world's largest producer and consumer of pork, and keeping track of the nation's estimated 700 million animals is notoriously difficult for farmers. They need to pay careful attention to ensure that piglets aren't crushed to death by their mothers, sows aren't bred past their prime, and sick pigs don't pass their illnesses on to the rest of the population. Currently, farmers track pigs by clipping wireless radio frequency identification (RFID) tags to the animals' ears. These can be expensive, and farmers don't always have time to fit each pig with a tag and scan them. They also provide only very basic information about the pigs' locations -- they can't determine anything about the animals' health and wellbeing.