AVBytes: AI & ML Developments this week - Microsoft's NLP AI, Build your Own Face Emoji, Google's Music Making Model, Reuter's AI Redefining Journalism, etc.

Pranav Dar — Published On March 19, 2018

The job of a data scientist is one of the most challenging in today’s world. You have to keep yourself up to date almost on a daily basis regarding the developments and ongoings in the machine learning universe. Slack off, and you might well find yourself lagging behind the competition.

Let AVBytes be your go-to guide! The articles under the AVBytes umbrella include official research papers, open source code, GitHub links, among various other things to help you, as a data science practitioner and enthusiast, keep yourself updated. You can download the code and replicate the model on your own machine, browse through the deep learning research papers, or just read through the article as a source of information.

In the past week, we saw quite a few breakthroughs – Microsoft’s NLP AI has reached human levels of quality and accuracy, a new deep learning technique to build your own 3D digital avatar, Google’s algorithm to generate new music from scratch, Reuters using AI to redefine how news stories are curated, among others.

Scroll down to view the articles. Also, Subscribe here to get AVBytes delivered directly to your inbox daily!

Below is a round-up of all the happenings in the last week. Click on each title to read the full article.

Microsoft’s Language Translation AI has Reached Human Levels of Accuracy: Another major breakthrough in NLP – Microsoft’s NLP AI translated sentences with the quality and accuracy of humans! We have provided links to the official research paper and the dataset in this article.

You can now Build your own 3D Digital Face Emoji using Deep Learning: Have you wanted to build your own animoji? Well, here’s deep learning with the solution again. Researchers have built a deep learning model that needs just an image to recreate facial and hair structures of a person in digital 3D.

Google has Released the Deep Learning Model used to Create Pixel 2’s Portrait Mode: Interested in learning the code behind Pixel 2’s portrait mode? Google has open sourced the code behind the technology. It uses TensorFlow and other libraries to build upon CNNs. Read more about it inside.

Google is Making Music with Machine Learning – and has released the code on GitHub: Want to generate new music on your machine using deep learning? Google has released the code for it’s open sourced project ‘NSynth Super’ that let’s you create your own musical instrument from scratch. Videos and GitHub links inside.

Supported by Andrew Ng, Woebot is a Mental Health Chatbot to help with Depression: Supported by Andrew Ng, Woebot is an application that uses Natural Language Processing to help people deal with depression, free of cost and without having to see a therapist.

Google is Using Deep Learning to Make Computers Better with Age!: Our devices slow down and wear out as they age, it can’t be helped. Or can it? Google might have a solution! Using deep learning, they’re aiming to make our machines actually improve with age! Check out the details + the official research paper link inside.

Source: Pexels

Reuters is Using AI to Spot Patterns and Redefine Journalism: AI writing and presenting news stories? That fantasy is close to becoming reality. Reuters is rolling out an AI tool that will trawl through massive datasets and present a quick summary to the journalist.