This week in AI & Machine Learning: How ML can save lives, Torch for R, Explainable AI, TF2 Object Detection, GANs for good, and more!

Artificial Intelligence News:

Create Safe Environments and Save Lives with Machine Learning

This article does an amazing job at showing some ways companies can utilize AI to reduce workplace accidents, mitigate natural disasters, and making it safer to return back to work environments during the current pandemic. 

These solutions may sound futuristic, but the technology to build and deploy them is available today! If you would like to see a demo of how you can use the Sixgill platform to solve these problems, request a demo.

Google’s Tone Transfer Attempts to Make Music from Any Sound 

Tone Transfer by is pretty neat. You can use it to change a 15 sec audio clip into machine learning generated music. As with most generative AI methods, there are always some fun results. Try it out yourself here.

Sophia Genetics Raises $110 Million

Further showing there is a lot of interest and potential demand around using AI in healthcare, SOPHiA GENETICS raises $110 Million in a series F round. The company uses machine learning to analyze medical images and offers a service around detecting and classifying genomic variants.


Developer Tools & Education

Torch for R Programming

If you’ve ever dreamed of having PyTorch functionality in the R programming language, it’s your lucky day! Torch is now an R package that lets you use PyTorch like functionality without any python installation!

Better TPU Support for PyTorch

Google’s TensorFlow team and Facebook’s PyTorch team collaborated to help bring better support for TPUs (Tensor Processing Units) to PyTorch and the PyTorch Lightning team. 

Deep Learning AI Launches GAN specialization

Many people are familiar with Andrew Ng and his work in building approachable AI courses covering machine learning, deep learning, NLP (Natural Language Processing), and more. This week at the GANs for Good panel announced their latest specialization: Generative Adversarial Networks (GANs)

TensorFlow 2 Object Detection API Updated

Ever since ITensorflow 2.x launched, I personally have been waiting for updates to better support the super useful Tensorflow Object Detection API, and it looks like we’ve gotten it! 

Interesting Podcasts & Interviews:

Open Source at Qualcomm AI Research with Jeff Gehlhaar and Zahra Koochak

Learn how Qualcomm Technologies is utilizing AI in research in this episode of “This Week in Machine Learning”.

Visualizing Climate Impact with GANs w/ Sasha Luccioni

Continuing with the theme of GANs everywhere this week! Listen how GANs can be used to help visualize climate impact in this episode of “This Week in Machine Learning”.

How to Use Calculus to Design Learning Machines

Learn why matrix calculus can be useful for the analysis and design machine learning models with lots of examples In this episode of “Learning Machines 101”.

Notable Papers:

Some of the interesting machine learning papers published this week.

EvolGAN: Evolutionary Generative Adversarial Networks

Learning Category- and Instance-Aware Pixel Embedding for Fast Panoptic Segmentation

Bag of Tricks for Adversarial Training

Dynamic Facial Asset and Rig Generation from a Single Scan

Mastering Textbook Questions with Pre-trained Transformers and Bottom-Up and Top-Down Attention

Towards a Mathematical Understanding of Neural Network-Based Machine Learning: what we know and what we don’t

Few-shot Learning for Time-series Forecasting

About Plainsight

Plainsight provides the unique combination of AI strategy, a vision AI platform, and deep learning expertise to develop, implement, and oversee transformative computer vision solutions for enterprises. Through the widest breadth of managed services and a vision AI platform for centralized processes and standardized pipelines, Plainsight makes computer vision repeatable and accountable across all enterprise vision AI initiatives. Plainsight solves problems where others have failed and empowers businesses across industries to realize the full potential of their visual data with the lowest barriers to production, fastest value generation, and monitoring for long-term success. For more information, visit


 View All Blogs