We study the problem of self-interested planning under uncertainty in settings shared with more than a thousand other agents, each of which plans at its own individual level. We refer to such large numbers of agents as an agent population. The decision-theoretic formalism of interactive partially observable Markov decision process (I-POMDP) is used to model the agent's self-interested planning. The first contribution of this article is a method for drastically scaling the finitely-nested I-POMDP to certain agent populations for the first time. Our method exploits two types of structure that is often exhibited by agent populations -- anonymity and context-specific independence. We present a variant called the many-agent I-POMDP that models both these types of structure to plan efficiently under uncertainty in multiagent settings. In particular, the complexity of the belief update and solution in the many-agent I-POMDP is polynomial in the number of agents compared with the exponential growth that challenges the original framework. While exploiting structure helps mitigate the curse of many agents, the well-known curse of history that afflicts I-POMDPs continues to challenge scalability in terms of the planning horizon. The second contribution of this article is an application of the branch-and-bound scheme to reduce the exponential growth of the search tree for look ahead. For this, we introduce new fast-computing upper and lower bounds for the exact value function of the many-agent I-POMDP. This speeds up the look-ahead computations without trading off optimality, and reduces both memory and run time complexity. The third contribution is a comprehensive empirical evaluation of the methods on three new problems domains -- policing large protests, controlling traffic congestion at a busy intersection, and improving the AI for the popular Clash of Clans multiplayer game. We demonstrate the feasibility of exact self-interested planning in these large problems, and that our methods for speeding up the planning are effective. Altogether, these contributions represent a principled and significant advance toward moving self-interested planning under uncertainty to real-world applications.

Deep Recurrent Neural Networks (RNNs) achieve state-of-the-art results in many sequence-to-sequence tasks. However, deep RNNs are difficult to train and suffer from overfitting. We introduce a training method that trains the network gradually, and treats each layer individually, to achieve improved results in language modelling tasks. Training deep LSTM with Gradual Learning (GL) obtains perplexity of 61.7 on the Penn Treebank (PTB) corpus. As far as we know (as for the 20.05.2017), GL improves the best state-of-the-art performance by a single LSTM/RHN model on the word-level PTB dataset.

Classification and clustering algorithms have been proved to be successful individually in different contexts. Both of them have their own advantages and limitations. For instance, although classification algorithms are more powerful than clustering methods in predicting class labels of objects, they do not perform well when there is a lack of sufficient manually labeled reliable data. On the other hand, although clustering algorithms do not produce label information for objects, they provide supplementary constraints (e.g., if two objects are clustered together, it is more likely that the same label is assigned to both of them) that one can leverage for label prediction of a set of unknown objects. Therefore, systematic utilization of both these types of algorithms together can lead to better prediction performance. In this paper, We propose a novel algorithm, called EC3 that merges classification and clustering together in order to support both binary and multi-class classification. EC3 is based on a principled combination of multiple classification and multiple clustering methods using an optimization function. We theoretically show the convexity and optimality of the problem and solve it by block coordinate descent method. We additionally propose iEC3, a variant of EC3 that handles imbalanced training data. We perform an extensive experimental analysis by comparing EC3 and iEC3 with 14 baseline methods (7 well-known standalone classifiers, 5 ensemble classifiers, and 2 existing methods that merge classification and clustering) on 13 standard benchmark datasets. We show that our methods outperform other baselines for every single dataset, achieving at most 10% higher AUC. Moreover our methods are faster (1.21 times faster than the best baseline), more resilient to noise and class imbalance than the best baseline method.

PUNE: California-headquartered data analytics firm Saama Technologies is setting up a deep learning centre in Pune as it attempts to build on its expertise in the data analytics space. Ken Coleman, chairman, Saama Technologies, told ET that the company was in the process of hiring people for the lab, which would work in collaboration with its existing engineering centre in the city. "Pune will play an increasingly important role in our success. We are not in India for the low costs but to tap into the brain power," said Coleman, who is also an advisor to venture capital firm Andreessen Horowitz. Saama would also be increasing its headcount by 15-20% by the end of this year, he said.

Typically, students work with robots that have been pre-programmed or program robots to undertake simple tasks for which the outcome is known. But a research project in Israel came up with a way for high schoolers and first-year engineering students to learn robot intelligence technologies by engaging them in teaching robots -- both physical and digital -- to learn. In a paper recently published by the International Journal of Online Engineering, three researchers described how students taught their robots to acquire skills by implementing a "reinforcement learning (RL) process" that used simulation modeling and cloud communication. The idea of RL is to use trial and error rather than direct instructions to help the robot determine appropriate performance criteria -- in this case, what angle it should situate itself in to lift varying weights. The project followed three phases.

Passengers' baggage is collected by robots, they relax in a luxurious waiting area complete with an indoor garden before getting a face scan and swiftly passing through security and immigration -- this could be the airport of the future. It's a vision that planners hope will become reality as new technology is rolled out, transforming the exhausting experience of getting stuck in lengthy queues in ageing, overcrowded terminals into something far more pleasant. The Asia-Pacific has been leading the way but faces fierce competition from the Middle East as major hubs compete to attract the growing number of long-haul travellers who can choose how to route their journey. The Asia-Pacific has been leading the way toward the airports of the future. The regions'are the two leading pockets of technology growth because they are really competing to be the global hubs for air transportation,' Seth Young, director of the Center for Aviation Studies at Ohio State University, told AFP. 'If I'm going to fly from New York to Bangalore, do I transfer through Abu Dhabi or Dubai or do I transfer through Hong Kong?

This story was originally published by Data-Smart City Solutions. It was excerpted from a paper "Artificial Intelligence for Citizen Services and Government" written by Harvard Ash Center Technology and Democracy Fellow Hila Mehr. From online services like Netflix and Facebook, to chatbots on our phones and in our homes like Siri and Alexa, we are beginning to interact with artificial intelligence (AI) on a near daily basis. AI is the programming or training of a computer to do tasks typically reserved for human intelligence, whether it is recommending which movie to watch next or answering technical questions. From small cities in the US to countries like Japan, government agencies are looking to AI to improve citizen services.

Home-grown technology and Internet of Things (IoT) company Smartron on Thursday unveiled'tronX' -- an Artificial Intelligence (AI)-powered IoT platform that would help make users' daily life easier and smarter. Terming it as one of the first global technologies being developed in India, the company said'tronX' is an intelligent ecosystem that helps connect a range of devices. Built on the world of'Internet of Trons', the ecosystem allows instant access to profile, data, content, services, Cloud, care, community and other IoT devices whether you are at home, in the car or at the office. "Smartron has been working for more than two years on creating a new connected ecosystem fuelled by AI-powered'IoT' and'tronX' is at the core of this brave new world," Mahesh Lingareddy, Founder and Chairman, Smarton, told IANS here. Lingareddy added that the company was excited to unveil'tronX' that is a kind of next-generation operating system designed to run seamlessly across devices.

The same technology that helped a computer beat a human in the ancient Chinese strategic game Go could dramatically change the way wars are fought, a report suggests. Artificially intelligent systems today can competently identify images and rapidly carry out repetitive virtual tasks. The technology is far from consciousness or replacing human warfighters but even in its current state, it will force the U.S. military to reassess its major strategies, analysts argue. As adversaries, namely China, invest in AI applications, the Defense Department should preempt any AI-driven attacks by conducting AI-themed war games, and by investing more in its own use of the technology, according to a report written on behalf of the Intelligence Advanced Research Projects Activity, the intelligence community's research and development unit. Harvard's Belfer Center for Science and International Affairs published the study.

China faces an acute shortage of doctors. Even in China's first-tier cities - Beijing, Shanghai, Chongqing, Shenzhen and Tianjin - many Chinese must wait in long lines at hospitals and clinics to receive examinations, diagnosis and treatment. The World Health Organization issued a report in 2016 disclosing that in China, there's a ratio: 1.5 doctors for every 1,000 people, while in the United States, it's 2.4 per 1,000 and in the United Kingdom, 2.8 per 1,000. Apparently, new solutions are required to help Chinese doctors reduce workloads. Hence, Chinese developers of Artificial Intelligence (AI) may have found the answer.