Chintu, the robot, slowly sat down on the floor, with both hands resting on its knees. Then, on command, it stood up, using one hand for support. The 58-centimetre-tall robot, manufactured by Softbank Robotics of France and owned by Maharashtra Institute of Technology (MIT), Pune, was one of the attractions of IBM Cloud Forum, a jamboree of companies using IBM's cloud and machine learning (ML) solutions in the last week of May in Mumbai. Alongside Chintu were its guardians -- Astitva Shah and Krishnamohan M, final-year engineering students from MIT, Pune. The duo have been working on a project to develop Chintu as an assistant for elderly people who are living alone.

Oracle are poised to help customers make the leap from on premise, up into the cloud. At Infosecurity Europe 2017, you couldn't escape the buzz of what will may soon become the next great battle on the cybersecurity frontier – automation. However, the belief that automation is a cyber security silver bullet is not one that is well believed it seems, with Oracle's Rohit Gupta telling CBR that not all situations can be solely monitored by technology. "There are certain conditions where automation will never be accepted. As an example, let's say the system discovers something suspicious going on in an executive's credentials, the CFO's credentials. You don't want to turn off the access between the CFO and his or her system, that could be career suicide, you never want that to happen. "In that scenario you have policy in place that says for these specific types of roles, or these specific types of entitlements, I want to have human intervention – somebody to look over this detected issue ...

By tracking cursor movement, lie detection becomes a game of cat and mouse. Every year, millions of people have their identities stolen. There's no foolproof way to pinpoint fakers, but thanks to Italian researchers, investigators may soon have another tool at their disposal--a way to suss out frauds and other liars online with just a few clicks of a mouse. Traditional methods of lie detection include face-to-face interviews and polygraphs that measure heart rate and skin conductance. But they can't be done remotely, or with large numbers of people.

Facts matter: but is that statement a fact? Suppose it could be proven or refuted – what then? What's the best way to tell someone they've made an error? How do you nudge the world in the direction of truth? These are the questions professional factcheckers wrestle with on a daily basis.

Police have released images of the fake explosives belts worn by the men who carried out the London Bridge attack. Khuram Butt, Rachid Redouane and Youssef Zaghba wore the belts as they stabbed people in Borough Market. The officer leading the investigation said use of the belts was a "tactic" he had not seen before in the UK and one that created "maximum fear". The belts were still on the attackers, who murdered eight people, when they were shot dead by police. Each belt had three disposable water bottles covered in masking tape attached to them.

We consider the problem of quickest change-point detection in data streams. Classical change-point detection procedures, such as CUSUM, Shiryaev-Roberts and Posterior Probability statistics, are optimal only if the change-point model is known, which is an unrealistic assumption in typical applied problems. Instead we propose a new method for change-point detection based on Inductive Conformal Martingales, which requires only the independence and identical distribution of observations. We compare the proposed approach to standard methods, as well as to change-point detection oracles, which model a typical practical situation when we have only imprecise (albeit parametric) information about pre- and post-change data distributions. Results of comparison provide evidence that change-point detection based on Inductive Conformal Martingales is an efficient tool, capable to work under quite general conditions unlike traditional approaches.

Anomalies in time-series data give essential and often actionable information in many applications. In this paper we consider a model-free anomaly detection method for univariate time-series which adapts to non-stationarity in the data stream and provides probabilistic abnormality scores based on the conformal prediction paradigm. Despite its simplicity the method performs on par with complex prediction-based models on the Numenta Anomaly Detection benchmark and the Yahoo!

The success of automated driving deployment is highly depending on the ability to develop an efficient and safe driving policy. The problem is well formulated under the framework of optimal control as a cost optimization problem. Model based solutions using traditional planning are efficient, but require the knowledge of the environment model. On the other hand, model free solutions suffer sample inefficiency and require too many interactions with the environment, which is infeasible in practice. Methods under the Reinforcement Learning framework usually require the notion of a reward function, which is not available in the real world. Imitation learning helps in improving sample efficiency by introducing prior knowledge obtained from the demonstrated behavior, on the risk of exact behavior cloning without generalizing to unseen environments. In this paper we propose a Meta learning framework, based on data set aggregation, to improve generalization of imitation learning algorithms. Under the proposed framework, we propose MetaDAgger, a novel algorithm which tackles the generalization issues in traditional imitation learning. We use The Open Race Car Simulator (TORCS) to test our algorithm. Results on unseen test tracks show significant improvement over traditional imitation learning algorithms, improving the learning time and sample efficiency in the same time. The results are also supported by visualization of the learnt features to prove generalization of the captured details.

Teaching a computer to "think" the way the human brain does means feeding it huge amounts of real-world data so that it can learn, analyze, predict, and solve problems. But in this brave new world of artificial intelligence (AI) and machine learning, there are no ethical guidelines, no regulations, and no parameters to govern how this data is collected and used. Artificial intelligence is a computer science branch that aims to develop computers that can learn and solve problems, much as a human brain does. When BM's AI supercomputer Watson is enlisted to help doctors tailor therapy to breast cancer patients, it needs to consume high volumes of medical data before it can provide better insights into personalized treatment options and their outcomes. Every time you pick a Netflix movie or ask your digital voice assistant to call a friend, you are dealing with AI.

The age of touch could soon come to an end. From smartphones and smartwatches, to home devices, in-car systems, touch is no longer the primary user interface. During this talk, Christophe, Principal Consultant at GraphAware will walk you through the design of building Conversational Bots. To this end, he used Amazon Alexa and combined it with a Natural Language Processing stack backed by a Neo4j Graph Database. You will discover the basics of an Amazon Alexa skill and how the user experience with voice devices can be enhanced with graph based algorithms such as recommendations.