This week in AI & Machine Learning: AI designing AI chips faster than humans, Kroger drone delivery, government data for AI research, NetHack RL challenge, TensorFlow Quantum, and more!
🤖 Artificial Intelligence News:
- Google is using AI to design its next generation of AI chips more quickly than humans can ← AI creating computer chips for AI is an interesting cycle to think about!
- These creepy fake humans herald a new age in AI
- DeepMind says reinforcement learning is ‘enough’ to reach general AI ← Google DeepMind seems to be doubling down on using reward maximization for Artificial Intelligence creation.
- U.S. Launches Task Force to Study Opening Government Data for AI Research
- Kroger drone delivery takes off in Centerville ← Kroger marks its entry into the drone delivery market by dropping off groceries in Centerville Ohio.
🛠️ Developer Tools & Education:
- Launching the NetHack Challenge at NeurIPS 2021 ← How far do you think reinforcement learning agents can make it in the difficult game NetHack?
- FRILL: On-Device Speech Representations using TensorFlow-Lite
- AI can now emulate text style in images in one shot — using just a single word ← Potentially a awesome data generation technique!
- Leveraging Machine Learning for Unstructured Data Processing at Pixie
- Easy Hyperparameter Tuning with Keras Tuner and TensorFlow
- Training with Multiple Workers using TensorFlow Quantum
🎤 Interesting Podcasts & Interviews:
- Haptic Intelligence with Katherine J. Kuchenbecker
- Learning to learn deep learning
- Data Science on AWS with Chris Fregly and Antje Barth
- Alison Gopnik on the different (and similar) ways robots and children learn
- Clément Delangue, CEO of Hugging Face, on the power of the open source community
📄 Notable Research Papers:
- DETReg: Unsupervised Pretraining with Region Priors for Object Detection
- ImaginE: An Imagination-Based Automatic Evaluation Metric for Natural Language Generation
- Space-time Mixing Attention for Video Transformer
- Learning by Watching
- Adversarial Motion Modelling helps Semi-supervised Hand Pose Estimation