Long-Short-Term-Memory Recurrent Neural Network (LSTM RNN) is a state-of-the-art (SOTA) model for analyzing sequential data. Current implementations of LSTM RNN in machine learning frameworks usually either lack performance or flexibility. For example, default implementations in Tensorflow and MXNet invoke many tiny GPU kernels, leading to excessive overhead in launching GPU threads. Although cuDNN, NVIDIA's deep learning library, can accelerate performance by around 2x, it is closed-source and inflexible, hampering further research and performance improvements in frameworks, such as PyTorch, that use cuDNN as their backend. In this paper, we introduce a new RNN implementation called EcoRNN that is significantly faster than the SOTA open-source implementation in MXNet and is competitive with the closed-source cuDNN. We show that (1) fusing tiny GPU kernels and (2) applying data layout optimization can give us a maximum performance boost of 3x over MXNet default and 1.5x over cuDNN implementations. Our optimizations also apply to other RNN cell types such as LSTM variants and Gated Recurrent Units (GRUs). We integrate EcoRNN into MXNet Python library and open-source it to benefit machine learning practitioners.

This paper reconstructs the Freebase data dumps to understand the underlying ontology behind Google's semantic search feature. The Freebase knowledge base was a major Semantic Web and linked data technology that was acquired by Google in 2010 to support the Google Knowledge Graph, the backend for Google search results that include structured answers to queries instead of a series of links to external resources. After its shutdown in 2016, Freebase is contained in a data dump of 1.9 billion Resource Description Format (RDF) triples. A recomposition of the Freebase ontology will be analyzed in relation to concepts and insights from the literature on classification by Bowker and Star. This paper will explore how the Freebase ontology is shaped by many of the forces that also shape classification systems through a deep dive into the ontology and a small correlational study. These findings will provide a glimpse into the proprietary blackbox Knowledge Graph and what is meant by Google's mission to "organize the world's information and make it universally accessible and useful".

Speaking in Macclesfield, the Prime Minister will use a speech to challenge the NHS, Artificial Intelligence (AI) sector and health charities to use data and AI to transform the diagnosis of chronic diseases.

With both ARKit and ARCore available to the public, augmented reality is now enabled on over 500 million devices. There are over 2,000 AR apps available in iOS App Store and another 200-plus on Google Play. With few breakout hits, many are wondering what the killer use cases for AR will be. We can examine the growth of the mobile app ecosystem to better understand how mobile AR will evolve. The growth of the mobile ecosystem was driven, in part, by three use cases: Creative, Contextual and Connected apps.

Earlier this month, Google's Chief Executive Officer Sundar Pichai unveiled the new and improved Google News. According to Pichai, the app could now show "the news you care about from trusted sources while still giving you a full range of perspectives on events." The company is reportedly working on integrating artificial intelligence in its news app in an effort to eliminate disinformation and help people get more viewpoints that are beyond their so-called "filter bubble." The new initiative was said to be part of Google's goal to be at the center of online news. This also includes new efforts to help advertisers and publishers gain paid subscribers using the tech company's platform. Google's product chief Trystan Upstill also added that the Google News would make use of the "best artificial intelligence to find the best of human intelligence," which, according to Google, is the reporting done by local and international journalists.

The Basware Assistant uses natural language processing and artificial intelligence to create a new and simplified way for people to interact with Basware's e-procurement solution. They can communicate with the Basware Assistant like they would with a person, to search for orders and purchase requests using vendor and item names, as well as ID and document numbers. By giving people the ability to specify what they are looking, it eliminates the need for having to navigate a series of screens to reach their intended purchase. Through its natural language processing and AI capabilities, the virtual assistant improves system usability, taking another step in streamlining the overall procurement experience. Not only does it help people find purchase orders and order requests more quickly and save them time, it also reduces the training required for new people to buy with Basware.

It's time to revisit the discussion on recommendation engines. In this installment, I'm going to provide you a conceptual overview of the topic, and then, following that I'll show you how to build a recommendation engine in R. Ready? Before showing you how to build a recommendation engine in R, I need to get you up-to-speed on the concepts behind how recommendation engines work. In case you're totally new to marketing data science, let me illustrate the recommendation engine concept a little before proceeding. You know how, when you go buy something on Amazon, you see related products under the heading of'People who purchased this item also purchased…' (or something like that).

While the chief data science officer role is common at technology firms, it's not standard at accounting and consulting firms. "I'm excited for this opportunity and the commitment from Crowe to push forward into this area," said Bass, who joined Crowe in 2015 as the firm's only data scientist. "With a team of world-class data scientists, we've already set the groundwork and incorporated machine learning into a few solutions in a very short amount of time." Bass has extensive experience applying advanced analytic, automation and machine learning techniques to a variety of industries. He is a continual learner, as shown through five combined degrees in aerospace engineering, statistics and finance from the University of Michigan, Purdue University and the Pennsylvania State University.

Syncsort already offers high-performance data integration capabilities that include continuous streaming capabilities to make fresh data available on-demand from disparate, enterprise-wide data sources, both on-premise and in the cloud. To keep changes to source and target data in sync, the new CDC capabilities greatly extend supported data sources and targets. Now, in addition to IBM Db2/z, DMX Change Data Capture adds IBM i, IBM Informix, Oracle, Oracle RAC, Sybase, Db2 and MS SQL Server as both sources and targets, VSAM as a source, and HDFS, Hive, Impala, Teradata, MySQL, Azure SQL, PostgreSQL and Kafka as targets.

In recent months, edge computing has become a buzzword. But what exactly is edge computing? Let's time travel back 10 years to the era of personal computing. Back then, we typed our documents in Word, listened to music and played media from CDs we bought and stored our photos on the computer hard drive. Our one personal computer- a desktop or laptop, was the central hub of every computing activity we did back then.