Traditional statistical analysis requires that the analysis process and data are independent. By contrast, the new field of adaptive data analysis hopes to understand and provide algorithms and accuracy guarantees for research as it is commonly performed in practice, as an iterative process of interacting repeatedly with the same data set, such as repeated tests against a holdout set. Previous work has defined a model with a rather strong lower bound on sample complexity in terms of the number of queries, $n\sim\sqrt q$, arguing that adaptive data analysis is much harder than static data analysis, where $n\sim\log q$ is possible. Instead, we argue that those strong lower bounds point to a limitation of the previous model in that it must consider wildly asymmetric scenarios which do not hold in typical applications. To better understand other difficulties of adaptivity, we propose a new Bayesian version of the problem that mandates symmetry. Since the other lower bound techniques are ruled out, we can more effectively see difficulties that might otherwise be overshadowed. As a first contribution to this model, we produce a new problem using error-correcting codes on which a large family of methods, including all previously proposed algorithms, require roughly $n\sim\sqrt[4]q$. These early results illustrate new difficulties in adaptive data analysis regarding slightly correlated queries on problems with concentrated uncertainty.

We introduce the value iteration network (VIN): a fully differentiable neural network with a `planning module' embedded within. VINs can learn to plan, and are suitable for predicting outcomes that involve planning-based reasoning, such as policies for reinforcement learning. Key to our approach is a novel differentiable approximation of the value-iteration algorithm, which can be represented as a convolutional neural network, and trained end-to-end using standard backpropagation. We evaluate VIN based policies on discrete and continuous path-planning domains, and on a natural-language based search task. We show that by learning an explicit planning computation, VIN policies generalize better to new, unseen domains.

This story was delivered to BI Intelligence "Fintech Briefing" subscribers. To learn more and subscribe, please click here. Insurtech continues to be an active segment of fintech, with startups emerging in all areas of the industry, including within the life insurance space. One such insurtech, Israeli firm Atidot, is now bolstering its capabilities, announcing the launch of its cloud-based, predictive data analytics platform for incumbent life insurers on Wednesday, according to a press release seen by BI Intelligence. Atidot says its platform can help life insurers better leverage their data to more accurately predict their policyholders' behavior, and adjust their business strategies more effectively.

This was a question from a university researcher in the Pacific northwest. The good news is our Azure Machine Learning or Azure ML is a powerful native Platform as a Service (PaaS) offering which has been around since 2015. When you think ML many different things fall into this like AI, neural networks, predictive outcomes. Past data is often huge – the quantity of data is doubling about every 18 months and that's only increasing from here. Computers can consider far more variables than a human making the same decision.

Will progress in Artificial Intelligence provide humanity with a boost of unprecedented strength to realize a better future, or could it present a threat to the very basis of human civilization? The future of artificial intelligence is up for debate, and the Origins Project is bringing together a distinguished panel of experts, intellectuals and public figures to discuss who's in control. Eric Horvitz, Jaan Tallinn, Kathleen Fisher and Subbarao Kambhampati join Origins Project director Lawrence Krauss. Recorded Saturday, February 25th, 2017 Eric Horvitz is managing director of Microsoft Research's main Redmond Lab, an American computer scientist, and technical fellow at Microsoft. Horvitz received his PhD and MD degrees at Stanford University, and has continued his research and work in areas that span theoretical and practical challenges of machine learning and inference, human-computer interaction, artificial intelligence, and more.

We pick the cell size to be bounded by (r)/( (n)), so that each grid cell will contain at most one sample, and thus the grid can be implemented as a simple n-dimensional array of integers: the default 1 indicates no sample, a non-negative integer gives the index of the sample located in a cell. Step 1. Select the initial sample, x0, randomly chosen uniformly from the domain.

HEN shown photographic images under a free-viewing (i.e task-free) paradigm, human eyes preferentially fixate on image locations which are visually salient. Multiple studies [1]-[5] have demonstrated that this fixation mechanism is bottom-up, predominantly driven by image content and richness of detail (color, texture etc.). This explanation, while satisfactory for photographic images, seems inadequate for certain categories of images such as line drawings. In particular, one class of line drawings - hand-drawn sketches - are sparse and largely devoid of detailed content. In addition, they are typically binary images containing virtually no color-based information (see Figure 1). Even so, multiple studies have demonstrated a "fixations-intonothing" phenomenon [6]-[9], wherein the eye fixations on the same stimulus by multiple subjects fall on empty regions, yet exhibit enough regularity to make gaze-based inferences. One possible explanation is that the first eye fixation conveys all there is to know ('Gestalt') about the underlying scene semantics [10] and the regularity in rest of the fixations is a statistical anomaly. However, a more intriguing explanation is that these empty region fixations aim to implicitly verify the overall consistency of the scene content depicted in the sketch [11], [12]. Which of these explanations is correct?

We design, conduct and present the results of a highly personalized baseline emotion recognition experiment, which aims to set reliable ground-truth estimates for the subject's emotional state for real-life prediction under similar conditions using a small number of physiological sensors. We also propose an adaptive stimuli-selection mechanism that would use the user's feedback as guide for future stimuli selection in the controlled-setup experiment and generate optimal ground-truth personalized sessions systematically. Initial results are very promising (85% accuracy) and variable importance analysis shows that only a few features, which are easy-to-implement in portable devices, would suffice to predict the subject's emotional state.

This paper analyses properties of localized kernel based, nonparametric statistical machine learning methods, in particular of support vector machines (SVMs) and methods close to them. Caused by the enormous research activities there is abundance of general introductions to this field of computer science and statistics. Beside many publications in international journals there are summarizing textbooks like for example Cristianini & Shawe-Taylor (2000), Schölkopf & Smola (2001), Steinwart & Christmann (2008) or Cucker & Zhou (2007) from a mathematical or statistical point of view. Nevertheless, we want to give a short overview over the analyzed topic. Support vector machines were initially introduced by Boser, Guyon & Vapnik (1992) und Cortes & Vapnik (1995), based on earlier work like the Russian original of Vapnik, Chervonenkis & Červonenkis (1979).

The probability that a user will click a search result depends both on its relevance and its position on the results page. The position based model explains this behavior by ascribing to every item an attraction probability, and to every position an examination probability. To be clicked, a result must be both attractive and examined. The probabilities of an item-position pair being clicked thus form the entries of a rank-$1$ matrix. We propose the learning problem of a Bernoulli rank-$1$ bandit where at each step, the learning agent chooses a pair of row and column arms, and receives the product of their Bernoulli-distributed values as a reward. This is a special case of the stochastic rank-$1$ bandit problem considered in recent work that proposed an elimination based algorithm Rank1Elim, and showed that Rank1Elim's regret scales linearly with the number of rows and columns on "benign" instances. These are the instances where the minimum of the average row and column rewards $\mu$ is bounded away from zero. The issue with Rank1Elim is that it fails to be competitive with straightforward bandit strategies as $\mu \rightarrow 0$. In this paper we propose Rank1ElimKL which simply replaces the (crude) confidence intervals of Rank1Elim with confidence intervals based on Kullback-Leibler (KL) divergences, and with the help of a novel result concerning the scaling of KL divergences we prove that with this change, our algorithm will be competitive no matter the value of $\mu$. Experiments with synthetic data confirm that on benign instances the performance of Rank1ElimKL is significantly better than that of even Rank1Elim, while experiments with models derived from real data confirm that the improvements are significant across the board, regardless of whether the data is benign or not.