Google's hardware engineering team that designed and developed the TensorFlow Processor Unit detailed the architecture and benchmarking experiment earlier this month. This is a follow up post on the initial announcement of the TPU from this time last year.
Over the past five years, Google searches for Machine Learning have gone up five times. “Fo anything that has machine learning in it or blockchain in it, the valuation goes up, 2, 3, 4, 5x”, Andy Stewart pointed out. Zachary Lipton claimed a "misinformation epidemic" in the field in a recent blog post. In this article we present the technical perspective of ML and how it can be presented.
Google recently announced the Cloud Machine Learning API updates at the Google Cloud Next Conference. This includes a set of APIs in the areas of vision, video intelligence, speech, natural language, translation and job search.
The research team at IBM recently announced they've reached a new industry record at 5.5%, using the SWITCHBOARD linguistic corpus. This brings us closer to what's considered to be the human error rate, 5.1%. They used deep learning technologies and acoustic models to accomplish this milestone.
Cloud enthusiasts from around the world attended Google Cloud Next to hear an update from the search giant. Three broad themes emerged from the many keynotes and 200+ sessions: service scale and maturity, usable machine learning, and enterprise-friendliness.
Google recently announced TensorFlow version 1.0. Python API is now stable and experimental APIs for Java and Go have been added. XLA delivers significant performance increase. Keras can also be integrated with TensorFlow using a build-in module. tf.transform, tf.layers, tf.metrics, and tf.losses all add new features to the framework..
Recently released Developer Preview 2 (DP2) for Android Things makes it easier to use TensorFlow for machine learning and computer vision on IoT devices. Additionally, it extends USB audio for several IoT platforms, adds Intel Joule support, and enables direct use of native drivers through a new Native PIO API.
MindMeld, a conversational AI company, has published The Conversational AI Playbook, a guide outlining the challenges and the steps to be made to create conversational applications.
Google’s Multilingual Neural Machine Translation System creates an interlingua and translates between language pairs and phrases with no previous direct translation available, dubbed Zero-Shot translation.
Intel open-sources BigDL, a distributed deep learning library that runs on Apache Spark. It leverages existing Spark clusters to run deep learning computations and simplifies the data loading from big datasets stored in Hadoop.
Instacart is an online delivery service for groceries under one hour. Customers order the items on the website or using the mobile app, and a group of Instacart’s shoppers go to local stores, purchase the items and deliver them to the customer. InfoQ interviewed Mathieu Ripert, data scientist at Instacart, to find out how machine learning is leveraged to guarantee a better customer experience.
“Fast and Probably Good Seedings for k-Means” by Olivier Bachem et al. was presented on 2016’s Neural Information Processing Systems (NIPS) conference and describes AFK-MC2, an alternative method to generate initial seedings for k-Means clustering algorithm that is several orders of magnitude faster than the state of art method k-Means++.
Using Neural Networks for sequence prediction is a well-known Computer Science problem with a vast array of applications in speech recognition, machine translation, language modeling and other fields. FB AI Research scientists designed adaptive softmax, an approximation algorithm tailored for GPUs which can be used to efficiently train neural networks over vocabularies of a billion words & beyond.
Amazon's Werner Vogels announces MXNet as the deep learning toolkit of choice for internal adoption, and extends AWS commitment to open-source MXNet ecosystem development.
Logz.io offers a hosted service which performs intelligent log analysis by using machine learning to derive insights from human interactions with log data that includes discussions on tech forums and public code repositories.