Manifold learning offers nonlinear dimensionality reduction of high-dimensional datasets. In this paper, we bring geometry processing to bear on manifold learning by introducing a new approach based on metric connection for generating a quasi-isometric, low-dimensional mapping from a sparse and irregular sampling of an arbitrary manifold embedded in a high-dimensional space. Geodesic distances of discrete paths on the input pointset are evaluated through "parallel transport unfolding" (PTU) to offer robustness to poor sampling and arbitrary topology. Our new geometric procedure exhibits the same strong resilience to noise as one of the staples of manifold learning, the Isomap algorithm, as it also exploits all pairwise geodesic distances to compute a low-dimensional embedding. While Isomap is limited to geodesically-convex sampled domains, parallel transport unfolding does not suffer from this crippling limitation, resulting in an improved robustness to irregularity and voids in the sampling. Moreover, it involves only simple linear algebra, significantly improves the accuracy of all pairwise geodesic distance approximations, and has the same computational complexity as Isomap. Finally, we show that our connection-based distance estimation can be used for faster variants of Isomap such as L-Isomap.

We propose a novel online algorithm for training deep feedforward neural networks that employs alternating minimization (block-coordinate descent) between the weights and activation variables. It extends off-line alternating minimization approaches to online, continual learning, and improves over stochastic gradient descent (SGD) with backpropagation in several ways: it avoids the vanishing gradient issue, it allows for non-differentiable nonlinearities, and it permits parallel weight updates across the layers. Unlike SGD, our approach employs co-activation memory inspired by the online sparse coding algorithm of [Mairal et al, 2009]. Furthermore, local iterative optimization with explicit activation updates is a potentially more biologically plausible learning mechanism than backpropagation. We provide theoretical convergence analysis and promising empirical results on several datasets.

To be informative, an evaluation must measure how well systems generalize to realistic unseen data. We identify limitations of and propose improvements to current evaluations of text-to-SQL systems. First, we compare human-generated and automatically generated questions, characterizing properties of queries necessary for real-world applications. To facilitate evaluation on multiple datasets, we release standardized and improved versions of seven existing datasets and one new text-to-SQL dataset. Second, we show that the current division of data into training and test sets measures robustness to variations in the way questions are asked, but only partially tests how well systems generalize to new queries; therefore, we propose a complementary dataset split for evaluation of future work. Finally, we demonstrate how the common practice of anonymizing variables during evaluation removes an important challenge of the task. Our observations highlight key difficulties, and our methodology enables effective measurement of future development.

Training large neural networks requires distributing learning across multiple workers, where the cost of communicating gradients can be a significant bottleneck. signSGD alleviates this problem by transmitting just the sign of each minibatch stochastic gradient. We prove that it can get the best of both worlds: compressed gradients and SGD-level convergence rate. The relative $\ell_1/\ell_2$ geometry of gradients, noise and curvature informs whether signSGD or SGD is theoretically better suited to a particular problem. On the practical side we find that the momentum counterpart of signSGD is able to match the accuracy and convergence speed of Adam on deep Imagenet models. We extend our theory to the distributed setting, where the parameter server uses majority vote to aggregate gradient signs from each worker enabling 1-bit compression of worker-server communication in both directions. Using a theorem by Gauss we prove that majority vote can achieve the same reduction in variance as full precision distributed SGD. Thus, there is great promise for sign-based optimisation schemes to achieve fast communication and fast convergence. Code to reproduce experiments is to be found at https://github.com/jxbz/signSGD.

This article was written by Monica Rogati. Monica is an independent data science executive and advisor. She built key data products and teams at Jawbone and LinkedIn; she now helps companies make the most out of their data. Do a project you care about. Make it good and share it. A quick search yields a plethora of possible resources that could help -- MOOCs, blogs, Quora answers to this exact question, books, Master's programs, bootcamps, self-directed curricula, articles, forums and podcasts.

We are excited to announce Samsung NEXT Q Fund, an early-stage venture fund focused on AI startups. In ML, researchers want to maximize "Q," which is the quality of an action in noisy, partially observable environments. We are interested in startups tackling AI Grand Challenges. Problem spaces we are looking into include learning in simulation, scene understanding, intuitive physics, program learning programs, automl, robot control, human computer interaction, and meta learning, just to name a few. We prefer novel techniques over solutions that "import ai."

Now, some scientists are starting to explore if artificial intelligence (AI) could help find answers.

As the amount of data continues to grow at an almost incomprehensible rate, being able to understand and process data is becoming a key differentiator for IT professionals and data-scientists. The scikit-learn library is one of the most popular platforms for everyday Machine Learning and data science because it is built upon Python, a fully featured programming language. This comprehensive 2-in-1 course is a comprehensive, practical guide to master the basics and learn from real-life applications of machine learning. Learn how to build and evaluate the performance of efficient models using scikit-learn. This training program includes 2 complete courses, carefully chosen to give you the most comprehensive training possible.

Kendal Parmar's son went from being a sporty and sociable boy who loved school, to a child who would stay in his room and rarely go outside. The change in his personality was down to a gaming disorder that crept up on him at the age of 12, when he started secondary school. Three years later, Joseph is still struggling with the problem. Parmar says the biggest sign that something was wrong was the amount of arguing that would occur when she asked him to stop playing video games. "Eventually his habits developed and he was gaming all the time. He became too terrified to go to school and he was terrified of people," she says.

Within buildings 10 and 30 of the Siemens complex on the outskirts of Munich, the next generation of German workers are toiling over a range of test projects. The assignments are carefully chosen to impart the skills needed to continue the German miracle in automated manufacturing. In one room, a group of young men train to be automotive mechatronic engineers. They've just spent the past week feverishly programming a diminutive working model of an automated production line--complete with sensors, conveyor belts, and tools that work without human input. They're able to discuss their work in surprisingly good English, but what sets them apart from their peers in the US is that none of them attend a university.