Deep Learning is currently being used for a variety of different applications. But is frequently criticised for lacking a fundamental theory that can fully answer why does it work so well. It's only recently that the winner of the Test-of-Time award at the Conference on Neural Information Processing (NIPS) compared Deep Learning to Alchemy. Although Generalization Theory that explains why Deep Learning generalizes so well is an open problem, in this article we would discuss most recent theoretical and empirical advances in the field that attempt to explain it. An "apparent paradox" with Deep Learning is that it can generalize well in practice despite its large capacity, numerical instability, sharp minima, and nonrobustness. In a recent paper "Understanding deep learning requires rethinking generalization" it was shown that Deep Neural Networks (DNN) has enough capacity to memorize ImageNet and CIFAR10 datasets with random labels.

One of the greatest benefits of artificial intelligence (AI) to humankind is its influence on the medical field. This is according to Anton Jacobs, MD of Networks Unlimited, who says: "Powered by some of the most sophisticated technology, AI is assisting in improving medical diagnosis." From an AI doctor and chatbot to AI's powerful applications, machine learning and deep learning, a world that used to be all about coding, is transitioning into using computer programming to assist in life changing health issues such as early cancer detection. A massive advantage is that AI has the power to pool knowledge from the best specialists worldwide and provide it to patients anywhere geographically. "Imagine what this could mean to patients living in rural areas. They'd finally have the same access to knowledge as patients in top medical facilities," adds Jacobsz.

Machine learning (ML) is touted as the most critical skill of current times. Artificial intelligence (AI), an application of ML, is becoming pervasive. From autonomous vehicles to self-tuned databases, AI and ML are found everywhere. Industry analysts often refer to AI-driven automation as the job killer. Almost every domain and industry vertical are getting impacted by AI and ML.

Core ML makes it easy for iOS developers to add deep machine learning to their apps. In this post, I'll show you how you can train a Core ML model to derive intelligent insights. Machine learning has undoubtedly been one of the hottest topics over the past year, with companies of all kinds trying to make their products more intelligent to improve user experiences and differentiate their offerings. Google invested between $20B and $30B in artificial intelligence just last year alone, according to McKinsey's State Of Machine Learning And AI, 2017. AI is turning into a race for patents and intellectual property (IP) among the world's leading tech companies...The report cites many examples of internal development including Amazon's investments in robotics and speech recognition, and Salesforce on virtual agents and machine learning.

In this video on Deep Learning, Richard Socher of Salesforce provides insightful commentary of how research can work with engineering to help ensure artificial intelligence (AI) is optimized for product delivery and business growth. Deep Learning is a subfield of AI and Machine Learning concerned with algorithms inspired by the structure and function of the brain called artificial neural networks. It is a relatively recent field, first referenced in 1986, but made functional by recent developments in search algorithms and big data. Deep Learning is providing solutions in a wide range of areas now, including computer vision, feature recognition, anomaly detection and extraction of information from a sensed environment. It is directly relevant to robotics, search, knowledge extraction and the overall advancement of AI.

Martin Welker is the chief executive of Axonic. Even with the support of AI frameworks like TensorFlow or OpenAI, artificial intelligence still requires deep knowledge and understanding compared to a mainstream web developer. If you have built a working prototype, you are probably the smartest guy in the room. Congratulations, you are a member of a very exclusive club. With Kaggle you can even earn decent money by solving real world projects. All in all it is an excellent position to be in, but is it enough to build a business? You can not change market mechanics after all.

To be effective, state of the art machine learning technology needs large amounts of annotated data. There are numerous compelling applications in healthcare that can benefit from high performance automated decision support systems provided by deep learning technology, but they lack the comprehensive data resources required to apply sophisticated machine learning models. Further, for economic reasons, it is very difficult to justify the creation of large annotated corpora for these applications. Hence, automated annotation techniques become increasingly important. In this study, we investigated the effectiveness of using an active learning algorithm to automatically annotate a large EEG corpus. The algorithm is designed to annotate six types of EEG events. Two model training schemes, namely threshold-based and volume-based, are evaluated. In the threshold-based scheme the threshold of confidence scores is optimized in the initial training iteration, whereas for the volume-based scheme only a certain amount of data is preserved after each iteration. Recognition performance is improved 2% absolute and the system is capable of automatically annotating previously unlabeled data. Given that the interpretation of clinical EEG data is an exceedingly difficult task, this study provides some evidence that the proposed method is a viable alternative to expensive manual annotation.

Interpretation of electroencephalogram (EEG) signals can be complicated by obfuscating artifacts. Artifact detection plays an important role in the observation and analysis of EEG signals. Spatial information contained in the placement of the electrodes can be exploited to accurately detect artifacts. However, when fewer electrodes are used, less spatial information is available, making it harder to detect artifacts. In this study, we investigate the performance of a deep learning algorithm, CNN-LSTM, on several channel configurations. Each configuration was designed to minimize the amount of spatial information lost compared to a standard 22-channel EEG. Systems using a reduced number of channels ranging from 8 to 20 achieved sensitivities between 33% and 37% with false alarms in the range of [38, 50] per 24 hours. False alarms increased dramatically (e.g., over 300 per 24 hours) when the number of channels was further reduced. Baseline performance of a system that used all 22 channels was 39% sensitivity with 23 false alarms. Since the 22-channel system was the only system that included referential channels, the rapid increase in the false alarm rate as the number of channels was reduced underscores the importance of retaining referential channels for artifact reduction. This cautionary result is important because one of the biggest differences between various types of EEGs administered is the type of referential channel used.

Recurrent Neural Networks (RNNs) with sophisticated units that implement a gating mechanism have emerged as powerful technique for modeling sequential signals such as speech or electroencephalography (EEG). The latter is the focus on this paper. A significant big data resource, known as the TUH EEG Corpus (TUEEG), has recently become available for EEG research, creating a unique opportunity to evaluate these recurrent units on the task of seizure detection. In this study, we compare two types of recurrent units: long short-term memory units (LSTM) and gated recurrent units (GRU). These are evaluated using a state of the art hybrid architecture that integrates Convolutional Neural Networks (CNNs) with RNNs. We also investigate a variety of initialization methods and show that initialization is crucial since poorly initialized networks cannot be trained. Furthermore, we explore regularization of these convolutional gated recurrent networks to address the problem of overfitting. Our experiments revealed that convolutional LSTM networks can achieve significantly better performance than convolutional GRU networks. The convolutional LSTM architecture with proper initialization and regularization delivers 30% sensitivity at 6 false alarms per 24 hours.

Variational auto-encoder frameworks have demonstrated success in reducing complex nonlinear dynamics in molecular simulation to a single non-linear embedding. In this work, we illustrate how this non-linear latent embedding can be used as a collective variable for enhanced sampling, and present a simple modification that allows us to rapidly perform sampling in multiple related systems. We first demonstrate our method is able to describe the effects of force field changes in capped alanine dipeptide after learning a model using AMBER99. We further provide a simple extension to variational dynamics encoders that allows the model to be trained in a more efficient manner on larger systems by encoding the outputs of a linear transformation using time-structure based independent component analysis (tICA). Using this technique, we show how such a model trained for one protein, the WW domain, can efficiently be transferred to perform enhanced sampling on a related mutant protein, the GTT mutation. This method shows promise for its ability to rapidly sample related systems using a single transferable collective variable and is generally applicable to sets of related simulations, enabling us to probe the effects of variation in increasingly large systems of biophysical interest.