Scikit-learn has evolved as a robust library for machine learning applications in Python with support for a wide range of supervised and unsupervised learning algorithms. TensorFlow is quickly becoming the technology of choice for deep learning, because of its ease to build powerful and sophisticated neural networks. To perform traditional machine learning tasks in supervised learning and unsupervised learning using cutting-edge techniques from deep learning, you need to be familiar with Python and basic machine learning concepts. This comprehensive 2-in-1 course teaches you how to perform your day-to-day machine learning tasks with Scikit-learn and TensorFlow. It's a perfect blend of concepts and practical examples which makes it easy to understand and implement.

Some of the trends analyzed during the 4th edition of Madrid.AI included the possibilities of eye-tracking systems, the possibilities and limitations of Artificial Intelligence and the applications of Deep Learning in the real world.

We're releasing an analysis showing that since 2012, the amount of compute used in the largest AI training runs has been increasing exponentially with a 3.5 month-doubling time (by comparison, Moore's Law had an 18-month doubling period). Since 2012, this metric has grown by more than 300,000x (an 18-month doubling period would yield only a 12x increase). Improvements in compute have been a key component of AI progress, so as long as this trend continues, it's worth preparing for the implications of systems far outside today's capabilities. The chart shows the total amount of compute, in petaflop/s-days, that was used to train selected results that are relatively well known, used a lot of compute for their time, and gave enough information to estimate the compute used. A petaflop/s-day (pfs-day) consists of performing 1015 neural net operations per second for one day, or a total of about 1020 operations.

Generative Text-to-Speech Synthesis, Heiga Zen, Research Scientist Abstract: Recent progress in deep generative models and its application to text-to-speech (TTS) synthesis has made a breakthrough in the naturalness of artificially generated speech.

Artificial intelligence is revolutionizing our world in many unimaginable ways. At the verge of the Fourth Industrial Revolution, humanity is currently witnessing the first steps made by machines in reinventing the world we live in. And while we keep debating about the potential drawbacks and benefits of substituting humans with intelligent, self-learning machines, there's one area where AI's positive impact will definitely improve the quality of our lives: the health care industry. Machine learning algorithms can process unimaginable amounts of info in the blink of an eye. And they can be much more precise than humans in spotting even the smallest detail in medical imaging reports such as mammograms and CT scans.

This paper introduces a new method to classify sentiment polarity for aspects in product reviews. We call it bitmask bidirectional long short term memory networks. It is based on long short term memory (LSTM) networks, which is a frequently mentioned model in natural language processing. Our proposed method uses a bitmask layer to keep attention on aspects. We evaluate it on reviews of restaurant and laptop domains from three popular contests: SemEval-2014 task 4, SemEval-2015 task 12, and SemEval-2016 task 5. It obtains competitive results with state-of-the-art methods based on LSTM networks. Furthermore, we demonstrate the benefit of using sentiment lexicons and word embeddings of a particular domain in aspect-based sentiment analysis.

Relation classification is an important task in natural language processing. Traditional relation classification techniques suffer from extensive use of linguistic features and external toolkits. In recent years, deep learning models that can automatically learn features from text are playing a more essential role in this area. In this paper we present a novel convolutional neural network (CNN) approach along shortest dependency paths (SDP) for Chinese relation classification. We first propose a baseline end-to-end model that only takes sentence-level features, and then improve its performance by joint use of pre-extracted linguistic features. The performance of the system is evaluated on the ACE 2005 Multilingual Training Corpus Chinese dataset. The baseline model achieved a 74.93% F-score on six general type relations and a 66.29% F-score on eighteen subtype relations, and the performance improved 10.71% and 13.60% respectively by incorporating linguistic features into the baseline system.

The paper presents the work that has been done as part of the Graph Poem project in developing metaphor classifiers, now by deep learning methods (after previously having developed rule-based and machine learning algorithms), and a web-based metaphor detection tool. After reviewing the existing work on metaphor in natural language processing (NLP), digital humanities (DH), and artificial intelligence (AI), we present our own research and argue in favor of adopting data-intensive approaches, developing NLP classifiers, and applying graph theory (and particularly networks of networks) in computational literary or poetry analysis, while also highlighting the relevance of such work to DH, NLP, and AI in general.

The problem of detecting insider threats i.e.\ authorized individuals who pose a threat to an organization is challenging. Since insiders have authorized access to and use sensitive data and systems on a day-to-day basis, the difference between an attack and benign normal behavior is small. We propose a method to address these issues by leveraging peer group metadata to build more robust models of normal behavior and investigate how to make use of multiple of these models and aggregate the results. Our experiments show that the use of peer group metadata improves performance over individual models trained using either hand-crafted features or event sequences.

Current solutions to multimedia modeling tasks feature sequential models and static tree-structured models. Sequential models, especially models based on Bidirectional LSTM (BLSTM) and Multilayer LSTM networks, have been widely applied on video, sound, music and text corpora. Despite their success in achieving state-of-the-art results on several multimedia processing tasks, sequential models always fail to emphasize short-term dependency relations, which are crucial in most sequential multimedia data. Tree-structured models are able to overcome this defect. The static tree-structured LSTM presented by Tai et al. forcingly breaks down the dependencies between elements in each semantic group and those outside the group, while preserves chain-dependencies among semantic groups and among nodes in the same group. Though the tree-LSTM network is able to better represent the dependency structure of multimedia data, it requires the dependency relations of the input data to be known before it is fed into the network. This is hard to achieve since for most types of multimedia data there exists no parsers which can detect the dependency structure of every input sequence accurately enough. In order to preserve dependency information while eliminating the necessity of a perfect parser, in this paper we present a novel neural network architecture which 1) is self-expandable and 2) maintains the layered dependency structure of incoming multimedia data. We call our new neural network architecture Seq2Tree network. A Seq2Tree model is applicable on classification, prediction and generation tasks with task-specific adjustments of the model. We prove by experiments that our Seq2Tree model performs well in all the three types of tasks.