A number of machine learning algorithms are using a metric, or a distance, in order to compare individuals. The Euclidean distance is usually employed, but it may be more efficient to learn a parametric distance such as Mahalanobis metric. Learning such a metric is a hot topic since more than ten years now, and a number of methods have been proposed to efficiently learn it. However, the nature of the problem makes it quite difficult for large scale data, as well as data for which classes overlap. This paper presents a simple way of improving accuracy and scalability of any iterative metric learning algorithm, where constraints are obtained prior to the algorithm. The proposed approach relies on a loss-dependent weighted selection of constraints that are used for learning the metric. Using the corresponding dedicated loss function, the method clearly allows to obtain better results than state-of-the-art methods, both in terms of accuracy and time complexity. Some experimental results on real world, and potentially large, datasets are demonstrating the effectiveness of our proposition.

Learning from the crowd has become increasingly popular in the Web and social media. There is a wide variety of crowdlearning sites in which, on the one hand, users learn from the knowledge that other users contribute to the site, and, on the other hand, knowledge is reviewed and curated by the same users using assessment measures such as upvotes or likes. In this paper, we present a probabilistic modeling framework of crowdlearning, which uncovers the evolution of a user's expertise over time by leveraging other users' assessments of her contributions. The model allows for both off-site and on-site learning and captures forgetting of knowledge. We then develop a scalable estimation method to fit the model parameters from millions of recorded learning and contributing events. We show the effectiveness of our model by tracing activity of ~25 thousand users in Stack Overflow over a 4.5 year period. We find that answers with high knowledge value are rare. Newbies and experts tend to acquire less knowledge than users in the middle range. Prolific learners tend to be also proficient contributors that post answers with high knowledge value.

T ensors have a rich history, stretching over almost a century, and touching upon numerous disciplines; but they have only recently become ubiquitous in signal and data analytics at the confluence of signal processing, statistics, data mining and machine learning. This overview article aims to provide a good starting point for researchers and practitioners interested in learning about and working with tensors. As such, it focuses on fundamentals and motivation (using various application examples), aiming to strike an appropriate balance of breadth and depth that will enable someone having taken first graduate courses in matrix algebra and probability to get started doing research and/or developing tensor algorithms and software. Some background in applied optimization is useful but not strictly required. The material covered includes tensor rank and rank decomposition; basic tensor factorization models and their relationships and properties (including fairly good coverage of identifiability); broad coverage of algorithms ranging from alternating optimization to stochastic gradient; statistical performance analysis; and applications ranging from source separation to collaborative filtering, mixture and topic modeling, classification, and multilinear subspace learning. Index Terms --T ensor decomposition, tensor factorization, rank, canonical polyadic decomposition (CPD), parallel factor analysis (PARAF AC), T ucker model, higher-order singular value decomposition (HOSVD), multilinear singular value decomposition (MLSVD), uniqueness, NPhard problems, alternating optimization, alternating direction method of multipliers, gradient descent, Gauss-Newton, stochastic gradient, Cram er-Rao bound, communications, source separation, harmonic retrieval, speech separation, collaborative filtering, mixture modeling, topic modeling, classification, subspace learning. N.D. Sidiropoulos, X. Fu, and K. Huang are with the ECE Department, University of Minnesota, Minneapolis, USA; email: (nikos,xfu,huang663)@umn.edu .

In this article, we revisit the most fundamental statistics theorem, talking in layman terms. We investigate a special but interesting and useful case, which is not discussed in textbooks, data camps, or data science classes. This article is part of a series about off-the-beaten-path data science and mathematics, offering a fresh, original and simple perspective on a number of topics. Previous articles in this series can be found here and also here. The theorem discussed here is the central limit theorem.

The year Jojo and I started eighth grade, the power plant officially cut electricity to two hours a day. We'd already been through years of brownouts, of flickering lights, blinking monitors, older ag drones without artificial neural networks rebooting in their stations and randomly launching to spray the fields again or overfeed the chickens. So when Public Works & Electric issued a message to all our devices telling us about its irregular hours of operation, no one was surprised. The message was full of obfuscating language, but anyone with a tide chart could spot the correlation. Anyone driving down the causeway to the airport, past the power plant, could see through its chain-link fence the turbines standing silent, tense as raised shoulders; the grounds swamped in seawater, the ebbing tide dragging out an iridescent Rorschach of petroleum.

A. had a recurring dream. He dreamed it almost every night, but in the morning, when Goodman or one of the instructors woke him and asked if he remembered what he had dreamed, he was always quick to say no. That wasn't because the dream was scary or embarrassing, it was just a stupid dream in which he was standing on the top of a grassy hill beside an easel, painting the pastoral landscape in water colors. The landscape in the dream was breathtaking, and since A. had come to the institution as a baby, the grassy hill was probably an imaginary place he had created or a real place he had seen in a picture or short film in one of his classes. The only thing that kept the dream from being completely pleasant was a huge cow with human eyes that was always grazing right next to A.'s easel. There was something infuriating about that cow: the spittle dripping from its mouth, the sad look it gave A., and the black spots on its back, which looked less like spots and more like a map of the world. Every time A. had that dream, it aroused the same feelings in him--calm that turned into frustration that turned into anger that immediately turned into compassion. He never touched the cow in the dream, never, but he always wanted to.

Udacity is positioned perfectly to benefit from the rush on talent in a number of growing areas of interest among tech companies and startups. The online education platform has added 14 new hiring partners across its Artificial Intelligence Engineer, Self-Driving Car Engineer and Virtual Reality Developer Nanodegree programs, as well as in its Predictive Analytics Nanodegree, including standouts like Bosch, Harma, Slack, Intel, Amazon Alexa and Samsung. That brings the total number of hiring partners for Udacity to over 30, which means a lot of potential soft landings for graduates of its nanodegree programs. The nanodegree offered by Udacity is its own original form of accreditation, which is based on a truncated field of study that spans months, rather than years, and allows students to direct the pace of their own learning. It also all takes place online, so students can potentially learn from anywhere. For Udacity, hiring partners help prove the value of their program to potential students, as they're effectively votes of confidence made by exactly the kinds of companies where students are looking to get jobs.

We are assisting a top international company currently building a Machine Learning and Data Scientist team in Dublin source a number of Software Engineers with proven experience implementing and applying Machine Learning techniques and methodologies in a commercial environment. This is a fantastic opportunity for Software Engineers with expertise in Machine Learning and Cognitive Computing technologies join a new operation with huge expansion plans for 2016/17 and beyond. Bachelor's Degree in Computer Science, Engineering or related field Proven experience in service development (REST, API, Microservices) Familiarity (preferably experience) with some or all of the following technologies: IBM Watson, SAS, Amazon ML, Google Cloud ML, Azure ML Studio Ability to create prototypes in SAS, R, Python, Scala, Java, C or similar stack Data retrieval and manipulation experience: SQL, Hive QL, Python, Hadoop, R, Unstructured Data Strong software design and architecture skills with an eye toward avoiding and reducing technical debt Bachelor's Degree in Computer Science, Engineering or related field This is a fantastic opportunity to join a brand new Software Engineering team within a Global company which is focused on developing and implementing Machine Learning solutions for their business.

Amazon's Echo has made tangible the promise of an artificially intelligent personal assistant in every home. Those who own the voice-activated gadget (known colloquially as Alexa, after its female interlocutor) are prone to proselytizing "her" charms, applauding Alexa's ability to call an Uber, order pizza or check a 10th-grader's math homework. The company says more than 5,000 people a day profess their love for Alexa. On the other hand, Alexa devotees also know that unless you speak to her very clearly . . . I hate her, I love her," one customer wrote on Amazon's website, while still awarding Alexa five stars. "You will very quickly learn how to talk to her in a way that she will understand and it's not unlike speaking to a small frustrating toddler." Voice recognition has come a long way in the past few years. But it's still not good enough to popularize the technology for everyday use and usher in a new era of human-machine interaction, allowing us to talk with all our ...

Unfortunately, the new training platform created by OpenAI, a San Francisco-based nonprofit, is only available to AI -- so if you're human, you're out of luck. The new'college' is, in actuality, a training platform called Universe, whereby AI can interact with games, web browsers, protein folding software, and "transfer learning," which allows them to take what they've learned in one application and apply it to another. The AI engages via Virtual Network Computing, or VNC, which involves them sending simulated mouse and keyboard strokes. The Universe digital suite's home is in the OpenAI artificial intelligence learning center in San Francisco, where developers will begin "measuring and training AI agents." OpenAI is the non-profit brainchild of entrepreneurs Elon Musk and Peter Thiel, who have made no secret of their ambitions to greatly accelerate the research and development of transhumanist technologies.