This week in AI & Machine Learning: How AI influences your purchases, Alexa gets better, Pytorch updates, AI privacy, A/B testing for machine learning, and more!
Artificial Intelligence News:
You’re probably already somewhat aware of how artificial intelligence plays a role in influencing product consumption in the form of recommendation engines. Read this article to learn more about how brick and mortar and online retailers are using data to discover what you may want to purchase before you even know it.
Read about how Amazon is using machine learning to amp up Alexa’s conversational features to predict requests the user doesn’t explicitly ask for. This could be a big step in making AI voice assistants better at understanding what you’re actually asking for and reducing the number of times you hear the response, “I’m sorry, I don’t understand”.
Learn how some companies are using AI and “Digital Twin” technology to help engineers design better for physical environments and how one company is using it to disrupt the private aircraft manufacturing industry.
Developer Tools & Education:
Read a list of the updates here.
I thought The Private AI Series in partnership with Andrew Trask and OpenMined sounded really neat! Also, check out the new Pytorch mobile features for supporting GPU and the Android Neural Networks APIs (NNAPI).
Most datasets for computer vision today are 2D which makes it harder to research 3D capabilities in the industry. Google is making this problem easier to tackle by releasing the Objectron Dataset which contains 15000 annotated videos and 4M annotated images with 3D labels.
This week’s tutorial from Pyimagesearch explores how to perform super resolution on images with OpenCV and deep learning.
Pssst… Also check out our newly released web based computer vision data annotation tool!
Upcoming Online AI & Data Events:
Learn how to make better evaluations of machine learning models by using A/B testing methods.
Introduction to Computer Vision: How to build an object detection model with your own dataset | Nov 17 – 5:30pm PDT
Join me for this hands on workshop to learn the basics of building a deep learning object detector and how to label your own dataset in a practical way.
This talk will introduce recent breakthroughs in Natural Language Processing (NLP) using transfer learning and transformer architectures.
AWS re:invent is going virtual and is free this year!
Interesting Podcasts & Interviews:
Listen to this podcast featuring Sixgill’s own Elizabeth Spears talk about AI ecosystems, the future of AI, finding the signal in the noise, and much more!
Listen to Roland Memisevic of Twenty Billion Neurons explain how they train deep neural networks to understand exercise and physical movements.
Explore different ways AI can be used in public health to prevent disparities and what it takes to build out digital infrastructures in lower resource areas.
Learn how AI and computer vision can be used to increase accessibility to digital assets.
Notable Research Papers:
Some of the interesting machine learning papers published this week.
- Scaling Hidden Markov Language Models
- When Do You Need Billions of Words of Pretraining Data?
- Prediction problems inspired by animal learning
- GRCNN: Graph Recognition Convolutional Neural Network for Synthesizing Programs from Flow Charts
- Transformers for One-Shot Visual Imitation
- Real-Time Decentralized knowledge Transfer at the Edge
- Audrey: A Personalized Open-Domain Conversational Bot
- DeepI2I: Enabling Deep Hierarchical Image-to-Image Translation by Transferring from GANs
- Offline Learning of Counterfactual Perception as Prediction for Real-World Robotic Reinforcement Learning