A monthly roundup of news about Artificial Intelligence, Machine Learning and Data Science. This is an eclectic collection of interesting blog posts, software announcements and data applications from Microsoft and elsewhere that I've noted over the past month or so.
Open Source AI, ML & Data Science News
torchdiffeq: A GPU-enabled PyTorch library for ordinary differential equation (ODE) solving.
Visual Studio Code adds support for Jupyter Notebooks, with the ability to import Notebooks into Code, and interact with cells via the iPython kernel.
Industry News
Amazon Comprehend, the AWS natural language processing service, can now be customized with user-defined entities and categories.
Google Cloud Platform introduces AI Hub, a place to publicly host and privately (within-enterprise) share and deploy pipelines, Jupyter notebooks, TensorFlow modules, and more.
A Chrome extension to open a Github-hosted Jupyter Notebook in Google Colab.
Facebook open-sources Horizon, a reinforcement learning platform that exports models in ONNX for deployment in large-scale production environments.
Google open-sources BERT (Bidirectional Encoder Representations), a system for pre-training natural language models with large corpuses of unannotated text.
Microsoft News
Microsoft has acquired XOXOCO, a startup focused on chatbots.
Microsoft Cognitive Services in Containers. Face detection and text analytics APIs in self-hosted containers that do not require private data to be sent to the cloud for analysis.
Named Entity Recognition comes to the Cognitive Services Text Analytics API, to identify persons, locations, organizations and other entities in unstructured text.
Azure Machine Learning Studio upgrades R support with the addition of the R 3.4 language engine.
New AI capabilities in preview for Power BI: AI model builder, Cognitive Services integration, and Azure Machine Learning interface.
The Microsoft Data Science Virtual Machine now includes Catboost, the open-source gradient boosting on decision trees library.
AzureR, a new suite of packages for managing Azure services from R.
Learning resources
Best Practices for Using Machine Learning in Businesses: Szilard Pafka cuts through the hype in this Twitter-summary of his keynote presentation.
A comparison of automatic machine learning services: Google AutoML, Microsoft Automated ML, AutoKeras and auto-sklearn.
Using Docker to deploy an R API with the plumber package.
Applications
Google uses acoustic analysis to identify humpback whales by their call.
Uber's approach to evaluating experiments: look at the entire distribution of outcomes, because focusing on just the mean or median can be misleading.
T-mobile uses R in production to route customer service queries with natural language AI.
Find previous editions of the monthly AI roundup here.
nice post,Most organizations grapple with quintillions of bytes of data every day, trying to figure out an information management strategy that could accelerate the flow of insights. This significantly complicates their big data solutions, increasing the cost of implementation and upkeep. However, big data engineering is not about using every bit of data originating from every source, it’s about making the smart decisions that accelerate business growth.for more info: Pridesys IT Ltd
Posted by: pridesys | November 27, 2018 at 21:30