This week in AI & Machine Learning: Cattle tracking with computer vision & blockchain, fruit picking robots, GPT-Neo, simulation to real world transfer learning, building object detection models and more!
🤖 Artificial Intelligence News:
- Leading Technology Providers To Launch Industry’s First Cattle Tracking NFT At Live Event ← Holy Cow! Check out this post and event to learn about tracking cattle on the blockchain with facial recognition!
- Memo Outlines DOD Plans for Responsible Artificial Intelligence
- Ripe Robotics starts trials with fruit picking robot Eve ← This robot is ready to be hired to pick your fruit. You can see a video of it in action here.
- How Calm is using machine learning to keep us all mellow
- China’s gigantic multi-modal AI is no one-trick pony ← Trained on 1.75 trillion parameter vs the 175 billion parameters in GPT-3!
🛠️ Developer Tools & Education:
- The FLORES-101 data set: Helping build better translation systems around the world
- Toward Generalized Sim-to-Real Transfer for Robot Learning ← So cool seeing transfer learning from virtual environments to the real world improve. These examples and visuals are neat!
- PyTorch builds the future of AI and machine learning at Facebook
- Few-shot learning in practice: GPT-Neo and the Huggingface Accelerated Inference API ← This is a great post If you’ve been wanting to get started with GPT-Neo.
- Hyperparameter tuning for Deep Learning with scikit-learn, Keras, and TensorFlow
🎤 Interesting Podcasts & Interviews:
- Getting Clever with Kaolin: Researchers Accelerate 3D Deep Learning with New Tools
- The fastest way to build ML-powered apps
- Mike Schuster on whether AI can help hedge fund investors to beat the market?
📄 Notable Research Papers:
- DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification
- Not All Images are Worth 16×16 Words: Dynamic Vision Transformers with Adaptive Sequence Length
- SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers
- Feature extraction and evaluation for BioMedical Question Answering
- Decision Transformer: Reinforcement Learning via Sequence Modeling
- You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection