Determinantal point processes (DPPs) have attracted significant attention as an elegant model that is able to capture the balance between quality and diversity within sets. DPPs are parameterized by a positive semi-definite kernel matrix. While DPPs have substantial expressive power, they are fundamentally limited by the parameterization of the kernel matrix and their inability to capture nonlinear interactions between items within sets. We present the deep DPP model as way to address these limitations, by using a deep feed-forward neural network to learn the kernel matrix. In addition to allowing us to capture nonlinear item interactions, the deep DPP also allows easy incorporation of item metadata into DPP learning. We show experimentally that the deep DPP can provide a considerable improvement in the predictive performance of DPPs.

Recent advances in employing neural networks on graph domains helped push the state of the art in link prediction tasks, particularly in recommendation services. However, the use of temporal contextual information, often modeled as dynamic graphs that encode the evolution of user-item relationships over time, has been overlooked in link prediction problems. In this paper, we consider the hypothesis that leveraging such information enables models to make better predictions, proposing a new neural network approach for this. Our experiments, performed on the widely used ML-100k and ML-1M datasets, show that our approach produces better predictions in scenarios where the pattern of user-item relationships change over time. In addition, they suggest that existing approaches are significantly impacted by those changes.

We tackle the problem of protein secondary structure prediction using a common task framework. This lead to the introduction of multiple ideas for neural architectures based on state of the art building blocks, used in this task for the first time. We take a principled machine learning approach, which provides genuine, unbiased performance measures, correcting longstanding errors in the application domain. We focus on the Q8 resolution of secondary structure, an active area for continuously improving methods. We use an ensemble of strong predictors to achieve accuracy of 70.7% (on the CB513 test set using the CB6133filtered training set). These results are statistically indistinguishable from those of the top existing predictors. In the spirit of reproducible research we make our data, models and code available, aiming to set a gold standard for purity of training and testing sets. Such good practices lower entry barriers to this domain and facilitate reproducible, extendable research.

The primary objective of domain adaptation methods is to transfer knowledge from a source domain to a target domain that has similar but different data distributions. Thus, in order to correctly classify the unlabeled target domain samples, the standard approach is to learn a common representation for both source and target domain, thereby indirectly addressing the problem of learning a classifier in the target domain. However, such an approach does not address the task of classification in the target domain directly. In contrast, we propose an approach that directly addresses the problem of learning a classifier in the unlabeled target domain. In particular, we train a classifier to correctly classify the training samples while simultaneously classifying the samples in the target domain in an unsupervised manner. The corresponding model is referred to as Discriminative Encoding for Domain Adaptation (DEDA). We show that this simple approach for performing unsupervised domain adaptation is indeed quite powerful. Our method achieves state of the art results in unsupervised adaptation tasks on various image classification benchmarks. We also obtained state of the art performance on domain adaptation in Amazon reviews sentiment classification dataset. We perform additional experiments when the source data has less labeled examples and also on zero-shot domain adaptation task where no target domain samples are used for training.

Reliable uncertainty quantification is a first step towards building explainable, transparent, and accountable artificial intelligent systems. Recent progress in Bayesian deep learning has made such quantification realizable. In this paper, we propose novel methods to study the benefits of characterizing model and data uncertainties for natural language processing (NLP) tasks. With empirical experiments on sentiment analysis, named entity recognition, and language modeling using convolutional and recurrent neural network models, we show that explicitly modeling uncertainties is not only necessary to measure output confidence levels, but also useful at enhancing model performances in various NLP tasks.

In this paper we introduce a novel pattern match neural network architecture that uses neighbor similarity scores as features, eliminating the need for feature engineering in a disfluency detection task. We evaluate the approach in disfluency detection for four different speech genres, showing that the approach is as effective as hand-engineered pattern match features when used on in-domain data and achieves superior performance in cross-domain scenarios.

We start with a function to get the decision tree. The user will call this to get the decision tree object, and then pass that result into subsequent functions to make predictions. Here we will construct it again, but for more complex examples we could train the model offline and save the trained model (using "pickle" for example) and then load that instead of recreating it. Creating the model in Excel gives us the possibility to play with some of the training parameters, which is fun. The code above is the same as the code we saw previously for training the decision tree, except for the @xl_func decorator.

While the brain's inner workings may remain elusive, it is only a matter of time before researchers develop a theory of deep learning. The difference is that when studying computers, researchers have access to every connection and pattern of activity in the network. The pace of progress is rapid, with research papers appearing daily on arXiv. Surprising advances are eagerly anticipated this December at the Neural Information Processing Systems conference in Montreal, which sold out 8,000 tickets in 11 minutes, leaving 9,000 hopeful registrants on the waiting list.

Editor's note: This is the second of two parts about Johnson & Johnson's technology operations. You can find the first part about early cloud adoption and the role of the CIO here. In technology, there is a delicate balance of when to adopt cutting-edge offerings and when to lay in wait, cautiously optimistic about capabilities but hesitant to waste resources. Chief information officers are key in the enterprise to understanding where technology fits, advising where to trial emerging technologies. Successful CIOs balance modernizing the technology core with working on moonshots to drive future business.

Alphabet's London-based AI lab DeepMind made history in 2016 when its AlphaGo software defeated a champion at the complex board game Go. On Tuesday, the company said it was handing off a seemingly much simpler software challenge: a health care app for hospital staff called Streams being tested by UK hospitals. That project and its staff will be transferred to DeepMind's much larger sister, Google. The announcement prompted an outcry from privacy researchers, which, along with legal constraints on the move, illustrating the challenges Google faces expanding its data-hungry operating style into the more sensitive business of health care. Last week, Google hired health industry veteran David Feinberg, who previously led the Pennsylvania health system Geisinger, to unify its scattered projects in the field.