Identifying novel drug–protein interactions is crucial for drug discovery. For this purpose, many machine learning-based methods have been developed based on drug descriptors and one-dimensional protein sequences. However, protein sequences cannot accurately reflect the interactions in three-dimensional space. However, direct input of three-dimensional structure is of low efficiency due to the sparse three-dimensional matrix, and is also prevented by the limited number of co-crystal structures available for training. Here we propose an end-to-end deep learning framework to predict the interactions by representing proteins with a two-dimensional distance map from monomer structures (Image) and drugs with molecular linear notation (String), following the visual question answering mode. For efficient training of the system, we introduce a dynamic attentive convolutional neural network to learn fixed-size representations from the variable-length distance maps and a self-attentional sequential model to automatically extract semantic features from the linear notations.

Imagine you're a fossil hunter. You spend months in the heat of Arizona digging up bones only to find that what you've uncovered is from a previously discovered dinosaur. That's how the search for antibiotics has panned out recently. The relatively few antibiotic hunters out there keep finding the same types of antibiotics. With the rapid rise in drug resistance in many pathogens, new antibiotics are desperately needed.

Please join us this evening as we discuss how to make your own AI and AI apps. Following is a brief agenda for the evening: 6:00 - 6:30 PM: Doors open for networking and pizza 6:30 - 7:00 PM: Presentation by Suren and Mohamed 7:00 - 7:30 PM: Presentation by Cody 7:30 - 8:00 PM - Q&A and Networking Forward-thinking businesses are using AI to gain greater insights, make more accurate predictions and use AI to create new and compelling products and services. Across industries, AI is having a positive impact including drug discovery in pharmaceutical to cure diseases, sepsis detection in healthcare to save lives, supply chain transportation optimization to take costs out in manufacturing, fraud detection in banking to save money, subscriber churn predictions in telecoms to retain customers and personalized policies in insurance to delight consumers. Every industry stands to benefit from AI in their business. In this meetup, we will discuss how domain experts and advanced data scientists can make their own AI by writing their own recipes and seamlessly extending H2O Driverless AI with their favorite tools from the rich ecosystem of open-source data science and machine learning libraries.

Alan Gibson, VP EMEA at Alteryx explains that these technologies have real business-altering potential and that's why Gartner's Enter the Age of Analytics report predicts that by 2023, AI and deep-learning techniques will be the two most common approaches for new applications of data science. But despite the promise, few companies have been able to successfully implement and deploy this technology as part of their overall data and analytics strategy--according to Gartner, 46 percent of CIOs have developed plans to deploy AI but just 4 percent have made the concept a reality. The truth is that it will take years before many organisations realise the true potential of AI and ML, but it is never too early to lay the groundwork now for an AI-driven future. In fact, if an organisation is not already thinking about what an AI strategy looks like, its competition is likely one step ahead. There's no time to waste, so here are five important points to consider when getting started with AI and ML.

This paper presents Neural Lyapunov MPC, an algorithm to alternately train a Lyapunov neural network and a stabilising constrained Model Predictive Controller (MPC), given a neural network model of the system dynamics. This extends recent works on Lyapunov networks to be able to train solely from expert demonstrations of one-step transitions. The learned Lyapunov network is used as the value function for the MPC in order to guarantee stability and extend the stable region. Formal results are presented on the existence of a set of MPC parameters, such as discount factors, that guarantees stability with a horizon as short as one. Robustness margins are also discussed and existing performance bounds on value function MPC are extended to the case of imperfect models. The approach is tested on unstable non-linear continuous control tasks with hard constraints. Results demonstrate that, when a neural network trained on short sequences is used for predictions, a one-step horizon Neural Lyapunov MPC can successfully reproduce the expert behaviour and significantly outperform longer horizon MPCs.

Here we present a general framework for learning simulation, and provide a single model implementation that yields state-of-the-art performance across a variety of challenging physical domains, involving fluids, rigid solids, and deformable materials interacting with one another. Our framework---which we term "Graph Network-based Simulators" (GNS)---represents the state of a physical system with particles, expressed as nodes in a graph, and computes dynamics via learned message-passing. Our results show that our model can generalize from single-timestep predictions with thousands of particles during training, to different initial conditions, thousands of timesteps, and at least an order of magnitude more particles at test time. Our model was robust to hyperparameter choices across various evaluation metrics: the main determinants of long-term performance were the number of message-passing steps, and mitigating the accumulation of error by corrupting the training data with noise. Our GNS framework is the most accurate general-purpose learned physics simulator to date, and holds promise for solving a wide range of complex forward and inverse problems.

Many state-of-the-art technologies developed in recent years have been influenced by machine learning to some extent. Most popular at the time of this writing are artificial intelligence methodologies that fall under the umbrella of deep learning. Deep learning has been shown across many applications to be extremely powerful and capable of handling problems that possess great complexity and difficulty. In this work, we introduce a new layer to deep learning: the fuzzy layer. Traditionally, the network architecture of neural networks is composed of an input layer, some combination of hidden layers, and an output layer. We propose the introduction of fuzzy layers into the deep learning architecture to exploit the powerful aggregation properties expressed through fuzzy methodologies, such as the Choquet and Sugueno fuzzy integrals. To date, fuzzy approaches taken to deep learning have been through the application of various fusion strategies at the decision level to aggregate outputs from state-of-the-art pre-trained models, e.g., AlexNet, VGG16, GoogLeNet, Inception-v3, ResNet-18, etc. While these strategies have been shown to improve accuracy performance for image classification tasks, none have explored the use of fuzzified intermediate, or hidden, layers. Herein, we present a new deep learning strategy that incorporates fuzzy strategies into the deep learning architecture focused on the application of semantic segmentation using per-pixel classification. Experiments are conducted on a benchmark data set as well as a data set collected via an unmanned aerial system at a U.S. Army test site for the task of automatic road segmentation, and preliminary results are promising.

In this paper we demonstrate speech synthesis using different electroencephalography (EEG) feature sets recently introduced in [1]. We make use of a recurrent neural network (RNN) regression model to predict acoustic features directly from EEG features. We demonstrate our results using EEG features recorded in parallel with spoken speech as well as using EEG recorded in parallel with listening utterances. We provide EEG based speech synthesis results for four subjects in this paper and our results demonstrate the feasibility of synthesizing speech directly from EEG features.

Music that is generated by recurrent neural networks often lacks a sense of direction and coherence. We therefore propose a two-stage LSTM-based model for lead sheet generation, in which the harmonic and rhythmic templates of the song are produced first, after which, in a second stage, a sequence of melody notes is generated conditioned on these templates. A subjective listening test shows that our approach outperforms the baselines and increases perceived musical coherence.

The vulnerability of deep neural networks to small, adversarially designed perturbations can be attributed to their "excessive linearity." In this paper, we propose a bottom-up strategy for attenuating adversarial perturbations using a nonlinear front end which polarizes and quantizes the data. We observe that ideal polarization can be utilized to completely eliminate perturbations, develop algorithms to learn approximately polarizing bases for data, and investigate the effectiveness of the proposed strategy on the MNIST and Fashion MNIST datasets.