3 min read

Microsoft releases new data science tools, Tensorflow publishes an implementation of SPINN, Machinelabs now supports private labs, and more in today’s top stories around machine learning, deep learning,and data science news.

1. Microsoft Releases DataScience Tools for Interactive Data Exploration and Data Modeling

Microsoft has introduced the early preview release of the Data Science Utilities developed by Team Data Science Process (TDSP). At present, the Data Science Utilities are released in the GitHub repository and these include:

  • Interactive Data Exploration, Analysis, and Reporting (IDEAR) in R, MRS, and in Python are tools developed for data scientists to interactively explore, visualize, and analyze data sets prior to building modeling tasks.The Python version of IDEAR is delivered through Jupyter Notebooks which runs on both Jupyter Notebook Server available and any notebook services in Python 2.7 or 3.5 kernel, as long as the required Python libraries are installed on the notebook server.
  • Automated Modeling and Reporting in R (AMAR in R) tool creates an automated workflow for generating and comparing multiple modeling approaches on a data-set.

One can easily run these utilities on sample data in the Data/Common directory. To read more on this, visit the GitHub repo.

2. Tensorflow publishes an implementation of SPINN written with Eager execution

Tensorflow recently published an implementation of SPINN written with Eager execution. Stack-Augmented Parser-Interpreter Neural Network (SPINN), is a recursive neural network that utilizes syntactic parse information for natural language understanding. It was originally described in the paper, A Fast Unified Model for Parsing and Sentence Understanding. The Tensorflow implementation is based on Jek bradbury’s PyTorch implementation. It includes model definition and training routines, a pipeline for loading and preprocessing the SNLI data and GloVe word embedding, written using the tf.data API, saving and loading checkpoints, TensorBoard summaries for monitoring and visualization, etc. More information can be found at the Github repo.

3. Machinelabs now supports private labs

MachineLabs announces the support for Private Labs.

MachineLabs is an open platform for sharing machine learning experiments with others. This means Labs can be viewed by and shared with everyone one, even via a browser. If a user wants to work in secrecy, or use it for company’s internal tasks, or may want to do a trial and error and don’t want to create a clutter of trails and errors in the public labs, they can now use private labs. To set labs as private is as easy as setting a flag private in one’s lab settings.

Once private, the lab will be only visible to the user, including its executions. One can also find all the public and private labs on their own profile page. Also, these private labs can be recognized by the little “private” badge.

Know more about these Private labs on MachineLab’s blog post.

4. Aureum 5.3 to power predictive analytics for  Data-Driven Industrial World

Peaxy announced the release of Aureum 5.3, a data access solution that provides a foundation for industrial digital twins and predictive analytics. Manuel Terranova, CEO, Peaxy, says,  “Aureum has evolved since 2012 from an advanced distributed data platform to an incredibly useful infrastructure component in complex analytical solutions and predictive applications. Our team of engineers are experts in supporting predictive analytics solutions to difficult industrial problems at enterprise scale.”

Aureum 5.3 is being used by Fortune 100 companies in the aviation, power generation and oil & gas industries as an essential data staging area for analytics that solve real-world business problems.

Know more on the website.

5. Dodge Data & Analytics launches Dodge Construction Central

Dodge Data & Analytics announced the launch of Dodge Construction Central, a single unified hub where all construction industry and project stakeholders can discover, share and access new and unique insights from across the entire construction ecosystem and along the full project lifecycle to make timely, data-driven decisions.

Dodge Construction Central

  • Delivers deep intelligence to project stakeholders from the most-comprehensive industry data cloud.
  • It empowers project stakeholders to collaborate with project teams and integrate insights directly into their business processes by leveraging artificial intelligence, advanced analytics, collaboration and workflow automation technologies.

To know about the new capabilities offered by Dodge Construction Central, you can visit this website.

 

LEAVE A REPLY

Please enter your comment!
Please enter your name here