This week in AI & Machine Learning: computer vision for livestock monitoring, AI generated NFTs, scikit-learn 1.0 released, WIT: a Wikipedia based image-text dataset, and more!
My Top AI Highlight:
One of my favorite things about computer vision is seeing all the different industries using it in practical applications!
This post outlines the steps used for tracking complex objects for counting or monitoring, such as the sheep pictured below.
Check out more vision AI agriculture use cases here!
🤖 Artificial Intelligence News:
- An NFT art world where machine learning AI is holding the paintbrush
- Deep learning helps predict new drug combinations to fight Covid-19
- Honeywell Introduces New Robotic Technology To Help Warehouses Boost Productivity, Reduce Injuries
- New machine learning method to analyze complex scientific data of proteins
🛠️ Developer Tools & Education:
- Scikit-learn 1.0 is released!
- New Captum version features more ways to build AI responsibly
- Announcing WIT: A Wikipedia-Based Image-Text Dataset
- Pathdreamer: A World Model for Indoor Navigation
- How to Recognize a Person Using their Veins/Vascular Data
- High-Quality, Robust and Responsible Direct Speech-to-Speech Translation
- Using Machine Learning to Help Track Bolides (Really Bright Meteors)
🎤 Interesting Podcasts & Interviews:
- Travis Oliphant: NumPy, SciPy, Anaconda, Python & Scientific Programming
- Jay McClelland: Neural Networks and the Emergence of Cognition
- Wild Things: NVIDIA’s Sifei Liu Talks 3D Reconstructions of Endangered Species
- AI’s Legal and Ethical Implications with Sandra Wachter
- Compositional ML and the Future of Software Development with Dillon Erb
📄 Notable Research Papers:
- Primer: Searching for Efficient Transformers for Language Modeling
- Relation-Guided Pre-Training for Open-Domain Question Answering
- Pix2seq: A Language Modeling Framework for Object Detection
- Neural forecasting at scale
- Trust Your Robots! Predictive Uncertainty Estimation of Neural Networks with Sparse Gaussian Processes
- Towards robustness under occlusion for face recognition