Duran Parsi headed to Pepperdine's law school three years ago with a mission: By the end, he'd either practice law or commit to his fledgling e-sports business. With graduation near, Parsi might need to grant himself an extension. Collegiate Star League, the 30-person e-sports operation run from his Studio City apartment, has essentially become the NCAA for video games. The company organized tournaments that 30,000 college students in the U.S. and Canada participated in this school year. Sponsorship sales tripled from last school year, and enough cash remained for Parsi, 29, to live off his business instead of student loans.

In March, robot programming students took part in the 2016-17 FIRST LEGO LEAGUE competition (FLL) with the theme'ANIMAL ALLIES'. FIRST LEGO League (FLL) is a program that supports children and youngsters and introduces them to science and technology in a sporty atmosphere. The theme of this year's FLL competition was to explore the relationship between people and animals. Today, it's time to think about how we can help each other. What might become possible when we work together with our ANIMAL ALLIES?

Google research scientist Kevin Murphy dropped a knowledge bomb on a crowd full of data scientists at the RE.WORK Deep Learning Summit Tuesday. While working on Google's image recognition programs -- algorithms that can analyze a photo and precisely identify items -- he thought of a unique application: Counting calories by analyzing photographs of food. Murphy explained that food photos are the most common type of photos on the web after pictures of actual people. With this developing software, a Google computer can look at an image of, say, eggs, pancakes and bacon, and identify each food. But Murphy's vision goes one step farther: He wants to create an app that looks at the photo, identifies the foods, analyzes the size of the foods, matches the nutritional information for each food, and then spits back the caloric intake of your meal. Snap a pic of your dinner (some of us already do it for Instagram), give it to Google and learn the caloric count of what you're putting into your body.

In debates over the future of artificial intelligence, many experts think of the new systems as coldly logical and objectively rational. But in a new study, researchers have demonstrated how machines can be reflections of us, their creators, in potentially problematic ways. Common machine learning programs, when trained with ordinary human language available online, can acquire cultural biases embedded in the patterns of wording, the researchers found. These biases range from the morally neutral, like a preference for flowers over insects, to the objectionable views of race and gender. Identifying and addressing possible bias in machine learning will be critically important as we increasingly turn to computers for processing the natural language humans use to communicate, for instance in doing online text searches, image categorization and automated translations.

We prove a \emph{query complexity} lower bound on rank-one principal component analysis (PCA). We consider an oracle model where, given a symmetric matrix $M \in \mathbb{R}^{d \times d}$, an algorithm is allowed to make $T$ \emph{exact} queries of the form $w^{(i)} = Mv^{(i)}$ for $i \in \{1,\dots,T\}$, where $v^{(i)}$ is drawn from a distribution which depends arbitrarily on the past queries and measurements $\{v^{(j)},w^{(j)}\}_{1 \le j \le i-1}$. We show that for a small constant $\epsilon$, any adaptive, randomized algorithm which can find a unit vector $\widehat{v}$ for which $\widehat{v}^{\top}M\widehat{v} \ge (1-\epsilon)\|M\|$, with even small probability, must make $T = \Omega(\log d)$ queries. In addition to settling a widely-held folk conjecture, this bound demonstrates a fundamental gap between convex optimization and "strict-saddle" non-convex optimization of which PCA is a canonical example: in the former, first-order methods can have dimension-free iteration complexity, whereas in PCA, the iteration complexity of gradient-based methods must necessarily grow with the dimension. Our argument proceeds via a reduction to estimating the rank-one spike in a deformed Wigner model. We establish lower bounds for this model by developing a "truncated" analogue of the $\chi^2$ Bayes-risk lower bound of Chen et al.

BENGALURU: The humongous amount of digital data being generated, and companies' need to glean insights and make predictions from them have made skills in data visualisation, data science, and machine learning among the most valued for technology recruiters today. This is reflected in the number of working professionals signing up for specialised courses in these spaces. Candidates who complete the courses tend to get between 20% and 50% increase in salaries. Kashyap Dalal, chief business officer at online learning platform Simplilearn, says that big data and analytics courses were the big growth drivers in the past three years. While data science continues to remain popular, accounting for 30% of all learners, courses on visualisation tools and machine learning have become very attractive over the past six months, he said.

In this paper, we consider a new low-quality label learning problem: learning time series detection models from temporally imprecise labels. In this problem, the data consist of a set of input time series, and supervision is provided by a sequence of noisy time stamps corresponding to the occurrence of positive class events. Such temporally imprecise labels commonly occur in areas like mobile health research where human annotators are tasked with labeling the occurrence of very short duration events. We propose a general learning framework for this problem that can accommodate different base classifiers and noise models. We present results on real mobile health data showing that the proposed framework significantly outperforms a number of alternatives including assuming that the label time stamps are noise-free, transforming the problem into the multiple instance learning framework, and learning on labels that were manually re-aligned.

We develop a novel family of algorithms for the online learning setting with regret against any data sequence bounded by the empirical Rademacher complexity of that sequence. To develop a general theory of when this type of adaptive regret bound is achievable we establish a connection to the theory of decoupling inequalities for martingales in Banach spaces. When the hypothesis class is a set of linear functions bounded in some norm, such a regret bound is achievable if and only if the norm satisfies certain decoupling inequalities for martingales. Donald Burkholder's celebrated geometric characterization of decoupling inequalities (1984) states that such an inequality holds if and only if there exists a special function called a Burkholder function satisfying certain restricted concavity properties. Our online learning algorithms are efficient in terms of queries to this function. We realize our general theory by giving novel efficient algorithms for classes including lp norms, Schatten p-norms, group norms, and reproducing kernel Hilbert spaces. The empirical Rademacher complexity regret bound implies --- when used in the i.i.d. setting --- a data-dependent complexity bound for excess risk after online-to-batch conversion. To showcase the power of the empirical Rademacher complexity regret bound, we derive improved rates for a supervised learning generalization of the online learning with low rank experts task and for the online matrix prediction task. In addition to obtaining tight data-dependent regret bounds, our algorithms enjoy improved efficiency over previous techniques based on Rademacher complexity, automatically work in the infinite horizon setting, and are scale-free. To obtain such adaptive methods, we introduce novel machinery, and the resulting algorithms are not based on the standard tools of online convex optimization.

KOLKATA: Interface between artificial intelligence and humans will have to be meaningfully leveraged to generate jobs in future, according to Alice P Gast, president of Imperial College, London. In Kolkata to deliver a keynote address at a CII meet on Monday, Gast - an acclaimed researcher, scholar and academic leader - pointed out that computers will soon have to be turned into collaborators for work rather than mere tools to carry out tasks. Graduates must now be trained for'jobs that don't yet exist', she felt. "Artificial intelligence and robotics have already started replacing people, particularly in manufacturing type jobs. We should now look into new areas where there could be an interaction between the two. Human-machine interface will have to be seriously explored and university courses should be now designed to meet that challenge. Jobs in future will perhaps be more analytical and collaborative. Computers will be something you won't just be working with, but collaborating with," said Gast.

Artificial intelligence (AI), in particular the field of machine learning, is touted as the "next big thing" according to Silicon Valley pundits. AI could be the key technology used to create cars that drive themselves or computers that diagnose some diseases better than doctors can. Backing up, though, let's take a moment to look at what exactly machine learning is. To put it simply, it's a computer that can program itself, usually by looking at feedback from its results. How is this important for education?