Celebrated \emph{Sequence to Sequence learning (Seq2Seq)} and its fruitful variants are powerful models to achieve excellent performance on the tasks that map sequences to sequences. However, these are many machine learning tasks with inputs naturally represented in a form of graphs, which imposes significant challenges to existing Seq2Seq models for lossless conversion from its graph form to the sequence. In this work, we present a general end-to-end approach to map the input graph to a sequence of vectors, and then another attention-based LSTM to decode the target sequence from these vectors. Specifically, to address inevitable information loss for data conversion, we introduce a novel graph-to-sequence neural network model that follows the encoder-decoder architecture. Our method first uses an improved graph-based neural network to generate the node and graph embeddings by a novel aggregation strategy to incorporate the edge direction information into the node embeddings. We also propose an attention based mechanism that aligns node embeddings and decoding sequence to better cope with large graphs. Experimental results on bAbI task, Shortest Path Task, and Natural Language Generation Task demonstrate that our model achieves the state-of-the-art performance and significantly outperforms other baselines. We also show that with the proposed aggregation strategy, our proposed model is able to quickly converge to good performance.

For data scientists looking to apply Apache Spark's advanced analytics techniques and deep learning models at scale, Databricks is happy to provide The Data Scientist's Guide to Apache Spark . This eBook features excerpts from the larger Definitive Guide to Apache Spark that will be published later this year.

Using deep learning to generate music: Part 1 and Part 2 on The AI Show on Channel 9.

The Linux Foundation launched a deep learning initiative this week designed to create a "neutral space for harmonization and acceleration" of AI, machine learning and deep learning technologies. The initial focus will be an AI standardization effort. The open source project called LF Deep Learning Foundation seeks to make the emerging AI technologies widely available to data scientists and developers, the group said during this week's Open Networking Summit in Los Angeles. Founding members include several Chinese technology giants that are pouring huge sums into AI research as part of a national strategy to dominate AI. They are: China's search giant Baidu; networking giant Huawei (SHE: 002502); Tencent, often referred to as China's Facebook; and telecommunications equipment vendor ZTE (SHE: 000063).

In the following sections, we review results for training time for one type of CNN model, feature extraction on a pre-trained ResNet50 model, and training time for one type RNN model.

While the prospects and relevance of text mining have provoked a lot of attention, there has been the plenty debate about its definition as investigators are yet to come to a unanimous decision on what text mining is all about. Given the aforesaid, it becomes clear that additional inquiries would have to be carried out before we are able to reach a unanimous decision. While there are still uncertainties about a singular definition for Text mining, it should be noted that its relevance and applicability is incontestable. Given the fact that the world is becoming increasingly swamped with an almost infinite amount of digital content, it is only expected that we can organize this seemingly boundless content in a way that supports improved accessibility and enhanced analytics. With an infinite amount of content that is continually put online on a daily basis, we are able to analyze the data to extract relevant and applicable info.

This is a implementation of Face Aligment with simple cnn in Keras, which is the second step of my FaceID system. Today there are lots of excellent face alignment algorithms, but they are somehow too complex to implement, and most of methods based on deep learning don't meet the requirement of real-time, here I introduce an efficient method based on simple convolution neural network, which can realize real-time face feature points detection. It costs me about only 10 minutes with cpu to train a model on a training set containing 7049 images. It's really fast, and the testing time is about 60ms per face. You can easily improve the accuray using different methods, such as make the convnet structure deeper or make a data augmentation and so on.

Deep Learning, a prominent topic in Artificial Intelligence domain, has been in the spotlight for quite some time now. It is especially known for its breakthroughs in fields like Computer Vision and Game playing (Alpha GO), surpassing human ability. Since the last survey, there has been a drastic increase in the trends. If you are interested in the topic here's an excellent non-technical introduction. If you are interested to know the recent trends, here's a great compilation.

This article comes from GitHub. A curated list of resources dedicated to bayesian deep learning. To read the original article, click here.

There were new announcements from Nvidia's GPU Technology Conference and Google's TensorFlow Dev Summit while France announced its own national AI strategy in a report and vowed €1.5bn in public funding. TensorFlow conference - Google hosted the TensorFlow Dev Summit on Friday in Mountain View, California, and announced a range of updates for its code, which remains the most popular software framework for AI and machine learning. Here's a list of the main highlights; all these tools are now going to be rolled out to researchers: If you're a TF nerd, there are more details on the blog and the Youtube channel. Also, here is the download link to TensorFlow 1.7.0 for you to play around with over the Easter weekend. La AI stratégie de France - France's President Emmanuel Macron acknowledged the importance of AI and the need to catch up to the the United States and China.