In order to interact intelligently with objects in the world, animals must first transform neural population responses into estimates of the dynamic, unknown stimuli which caused them. The Bayesian solution to this problem is known as a Bayes filter, which applies Bayes' rule to combine population responses with the predictions of an internal model. In this paper we present a method for learning to approximate a Bayes filter when the stimulus dynamics are unknown. To do this we use the inferential properties of probabilistic population codes to compute Bayes' rule, and train a neural network to compute approximate predictions by the method of maximum likelihood. In particular, we perform stochastic gradient descent on the negative log-likelihood with a novel approximation of the gradient. We demonstrate our methods on a finite-state, a linear, and a nonlinear filtering problem, and show how the hidden layer of the neural network develops tuning curves which are consistent with findings in experimental neuroscience.

Multi-Task Learning (MTL) can enhance a classifier's generalization performance by learning multiple related tasks simultaneously. Conventional MTL works under the offline or batch setting, and suffers from expensive training cost and poor scalability. To address such inefficiency issues, online learning techniques have been applied to solve MTL problems. However, most existing algorithms of online MTL constrain task relatedness into a presumed structure via a single weight matrix, which is a strict restriction that does not always hold in practice. In this paper, we propose a robust online MTL framework that overcomes this restriction by decomposing the weight matrix into two components: the first one captures the low-rank common structure among tasks via a nuclear norm and the second one identifies the personalized patterns of outlier tasks via a group lasso. Theoretical analysis shows the proposed algorithm can achieve a sub-linear regret with respect to the best linear model in hindsight. Even though the above framework achieves good performance, the nuclear norm that simply adds all nonzero singular values together may not be a good low-rank approximation. To improve the results, we use a log-determinant function as a non-convex rank approximation. The gradient scheme is applied to optimize log-determinant function and can obtain a closed-form solution for this refined problem. Experimental results on a number of real-world applications verify the efficacy of our method.

Recommender systems are widely used to predict personalized preferences of goods or services using users' past activities, such as item ratings or purchase histories. If collections of such personal activities were made publicly available, they could be used to personalize a diverse range of services, including targeted advertisement or recommendations. However, there would be an accompanying risk of privacy violations. The pioneering work of Narayanan et al.\ demonstrated that even if the identifiers are eliminated, the public release of user ratings can allow for the identification of users by those who have only a small amount of data on the users' past ratings. In this paper, we assume the following setting. A collector collects user ratings, then anonymizes and distributes them. A recommender constructs a recommender system based on the anonymized ratings provided by the collector. Based on this setting, we exhaustively list the models of recommender systems that use anonymized ratings. For each model, we then present an item-based collaborative filtering algorithm for making recommendations based on anonymized ratings. Our experimental results show that an item-based collaborative filtering based on anonymized ratings can perform better than collaborative filterings based on 5--10 non-anonymized ratings. This surprising result indicates that, in some settings, privacy protection does not necessarily reduce the usefulness of recommendations. From the experimental analysis of this counterintuitive result, we observed that the sparsity of the ratings can be reduced by anonymization and the variance of the prediction can be reduced if $k$, the anonymization parameter, is appropriately tuned. In this way, the predictive performance of recommendations based on anonymized ratings can be improved in some settings.

Deep neural networks require a large amount of labeled training data during supervised learning. However, collecting and labeling so much data might be infeasible in many cases. In this paper, we introduce a source-target selective joint fine-tuning scheme for improving the performance of deep learning tasks with insufficient training data. In this scheme, a target learning task with insufficient training data is carried out simultaneously with another source learning task with abundant training data. However, the source learning task does not use all existing training data. Our core idea is to identify and use a subset of training images from the original source learning task whose low-level characteristics are similar to those from the target learning task, and jointly fine-tune shared convolutional layers for both tasks. Specifically, we compute descriptors from linear or nonlinear filter bank responses on training images from both tasks, and use such descriptors to search for a desired subset of training samples for the source learning task. Experiments demonstrate that our selective joint fine-tuning scheme achieves state-of-the-art performance on multiple visual classification tasks with insufficient training data for deep learning. Such tasks include Caltech 256, MIT Indoor 67, Oxford Flowers 102 and Stanford Dogs 120. In comparison to fine-tuning without a source domain, the proposed method can improve the classification accuracy by 2% - 10% using a single model.

Its founder knows his demographic well: residents of the religiously conservative Gaza Strip, with its culture of resistance. Some 1,400 men have been killed in the three wars with Israel since 2008, leaving many widows who would like to remarry. Tradition, however, can make it difficult for them to wed single men. Mr. Abu Mustafa, 34, a math teacher, said he had no specific reasons to get married again, but said he did wish to give "dignity" to a widow. Ms. Abu Mustafa's first husband died during the conflict between Hamas and Israel in 2012.

Silvia Chiappa is a Senior Research Scientist at DeepMind, working at the intersection of probabilistic modeling and deep learning. Prior to DeepMind, she worked at Microsoft Research Cambridge, at the Statistical Laboratory University of Cambridge and the Max-Planck Institute for Biological Cybernetics. I spoke with Silvia to learn about her career in science, how we can overcome barriers for women in tech, and more. How did you begin your work in science and technology? At the age of 12 I started to appreciate the elegance of maths when learning about trigonometry.

Apple has revealed iOS 11, the future of the iPhone. The software comes with a range of new features, including changes to Messages and other iOS apps. But it also added a range of "technologies" – new tweaks intended to make the apps run much better. That was in keeping with a focus in Apple on improving the experience of its software, turning up and tweaking things rather than adding a host of new features, something it also did in macOS. The I.F.O. is fuelled by eight electric engines, which is able to push the flying object to an estimated top speed of about 120mph.

Parents often despair at the amount of time their children spend playing video games, but new research suggests it could be doing them more good than harm. Playing games like Minecraft or Lara Croft can make students smarter, according to a new study. This is because video games help young people develop good communication skills and mental versatility, the researchers claim. Gaming was found to teach students to be resourceful, and the study's lead researcher suggests that video games have a role to play in the future of education. Modern video games often require players to be adaptable and resourceful.

To meet the growing demand for AI expertise, companies are offering online education courses to prepare the workforce for the future. Increasingly, computers and devices learn and act on their own using software algorithms, the building blocks for artificial intelligence (AI) and machine learning (ML). Getting smartphones to understand voice commands, smart home sprinkler systems to change with the weather and online services to predict what people want requires programmers skilled in AI and ML. Demand for these coding skills is skyrocketing. Making devices smart and proactive remains controversial to anyone who fears that automation will lead to human job loss.

That's a broad topic which has been treated many times. Much of what has been written on this topic is good, much is bad. But I find that the stats vs. machine learning argument, at that level, tends to focus on the forest at the cost of completely overlooking the trees. Shah's definitions, which I believe are reflective of many approaches, tend to focus on different ends of the respective spectrums of each of these concepts, treating machine learning as a practical activity and statistics as a theoretical abstraction (and, yes, I'm lumping "statistical modeling" together with "statistics" in this case... at least, for now). The relationship between statistics and machine learning is actually a highly complex one, and merely defining the 2 concepts is not helpful in dissecting this connection.