X is some sort of an object, e.g. an email text, an image, or a document. Y is either a single class label from a finite set of labels, like spam/no spam, detected object or a cluster name for this document, or some number, like salary, in the next month or stock price. While such tasks can be daunting to solve (like sentiment analysis or predicting stock prices in real-time), they require rather clear steps to achieve good levels of mapping accuracy. Again, I'm not discussing situations with lack of training data to cover the modeled phenomenon or poor feature selection. In contrast, somewhat less straightforward areas of AI are the tasks that present you with a challenge of predicting as fuzzy structures as words, sentences, or complete texts.

NVIDIA has emerged as one of the leading organizations in the machine learning and deep learning space. We have previously seen some breakthrough software from them in this field – from a robot that can copy and execute human actions to an open source Python library that makes anyone an artist. And now they have announced a slew of machine learning tools at the Computer Vision and Pattern Recognition Conference (CVPR) in Utah. CVPR is an annual machine learning conference which sees the top minds in the ML and DL industry come together to discuss and present the latest tools and research to the community. These latest tools by NVIDIA include TensorRT 4, Apex, NVIDIA DALI (data loading library) and Kubernetes on NVIDIA's GPUs.

AI can detect signs of nuclear weapons testing banned under the Comprehensive Nuclear-Test-Ban Treaty, according to research from the US Pacific Northwest National Laboratory. A group of scientists have built a neural network to sniff out any unusual nuclear activity. Researchers from the Pacific Northwest National Laboratory (PNNL), one of the United States Department of Energy national laboratories, decided to see if they could use deep learning to sort through the different nuclear decay events to identify any suspicious behavior. The lab, buried beneath 81 feet of concrete, rock and earth, is blocked out from energy from cosmic rays, electronics and other sources. It means that the data collected is less noisy, making it easier to pinpoint unusual activity.

Google has perfected its ability to execute web search results for its users all over the world. In the early days of the Internet, the search engine was primarily suited for displaying search results for English users. Non-English-speaking users have complained that search results are often displayed in the wrong language entirely. However, Google is becoming more proficient at providing search results in other languages as well. A lot of factors can play a role, but one of the biggest is its use of deep learning to understand semantic references--enter semantic indexing. This can now be accomplished in any language that Google serves.

At some point in the not too distant future, the answer will seem self-evident. Like a lot of things, time changes perspective; the essential advancements we take for granted now were once deemed insurmountable. I believe we'll look back at the introduction of a 2 petaFLOPS deep learning system as essential to the evolution of AI in the enterprise. Single GPU systems once offered a seemingly limitless playground for researchers and developers on which to innovate. As deep learning model complexity and datasets grew to address increasingly exotic (but important) use cases, the standard currency of deep learning compute grew in response.

The task of drug-target interaction prediction holds significant importance in pharmacology and therapeutic drug design. In this paper, we present FRnet-DTI, an auto encoder and a convolutional classifier for feature manipulation and drug target interaction prediction. Two convolutional neural neworks are proposed where one model is used for feature manipulation and the other one for classification. Using the first method FRnet-1, we generate 4096 features for each of the instances in each of the datasets and use the second method, FRnet-2, to identify interaction probability employing those features. We have tested our method on four gold standard datasets exhaustively used by other researchers. Experimental results shows that our method significantly improves over the state-of-the-art method on three of the four drug-target interaction gold standard datasets on both area under curve for Receiver Operating Characteristic(auROC) and area under Precision Recall curve(auPR) metric. We also introduce twenty new potential drug-target pairs for interaction based on high prediction scores. Codes Available: https: // github. com/ farshidrayhanuiu/ FRnet-DTI/ Web Implementation: http: // farshidrayhan. pythonanywhere. com/ FRnet-DTI/

It is well known that over-parametrized deep neural networks (DNNs) are an overly expressive class of functions that can memorize even random data with $100\%$ training accuracy. This raises the question why they do not easily overfit real data. To answer this question, we study deep networks using Fourier analysis. We show that deep networks with finite weights (or trained for finite number of steps) are inherently biased towards representing smooth functions over the input space. Specifically, the magnitude of a particular frequency component ($k$) of deep ReLU network function decays at least as fast as $\mathcal{O}(k^{-2})$, with width and depth helping polynomially and exponentially (respectively) in modeling higher frequencies. This shows for instance why DNNs cannot perfectly \textit{memorize} peaky delta-like functions. We also show that DNNs can exploit the geometry of low dimensional data manifolds to approximate complex functions that exist along the manifold with simple functions when seen with respect to the input space. As a consequence, we find that all samples (including adversarial samples) classified by a network to belong to a certain class are connected by a path such that the prediction of the network along that path does not change. Finally we find that DNN parameters corresponding to functions with higher frequency components occupy a smaller volume in the parameter.

We investigate deep generative models, which allow us to use training data from one domain to build a model for another domain. We consider domains to have similar structure (texts, images). We propose the Variational Bi-domain Triplet Autoencoder (VBTA) that learns a joint distribution of objects from different domains. There are many cases when obtaining any supervision (e.g. paired data) is difficult or ambiguous. For such cases we can seek a method that is able to the information about data relation and structure from the latent space. We extend the VBTAs objective function by the relative constraints or triplets that sampled from the shared latent space across domains. In other words, we combine the deep generative model with a metric learning ideas in order to improve the final objective with the triplets information. We demonstrate the performance of the VBTA model on different tasks: bi-directional image generation, image-to-image translation, even on unpaired data. We also provide the qualitative analysis. We show that VBTA model is comparable and outperforms some of the existing generative models.

Recurrent neural networks (RNNs) have been drawing much attention with great success in many applications like speech recognition and neural machine translation. Long short-term memory (LSTM) is one of the most popular RNN units in deep learning applications. LSTM transforms the input and the previous hidden states to the next states with the affine transformation, multiplication operations and a nonlinear activation function, which makes a good data representation for a given task. The affine transformation includes rotation and reflection, which change the semantic or syntactic information of dimensions in the hidden states. However, considering that a model interprets the output sequence of LSTM over the whole input sequence, the dimensions of the states need to keep the same type of semantic or syntactic information regardless of the location in the sequence. In this paper, we propose a simple variant of the LSTM unit, persistent recurrent unit (PRU), where each dimension of hidden states keeps persistent information across time, so that the space keeps the same meaning over the whole sequence. In addition, to improve the nonlinear transformation power, we add a feedforward layer in the PRU structure. In the experiment, we evaluate our proposed methods with three different tasks, and the results confirm that our methods have better performance than the conventional LSTM.

This paper attempts to provide a state of the art in trend prediction using news headlines. We present the research done on predicting DJIA trends using Natural Language Processing. We will explain the different algorithms we have used as well as the various embedding techniques attempted. We rely on statistical and deep learning models in order to extract information from the corpuses.