This week in AI & Machine Learning: Are humanoid robots the future, recreating voices, AI movie directors, self-supervised anomaly detection, infinite memory transformers, and more!
My Top AI Highlight:
Agility Robotics’ Digit gets a warehouse gig
Humanoid robots are not a new concept, but after the “Tesla bot” was announced at Tesla’s AI Day, there has been a debate around the efficiency of a humanoid design.
Boston Dynamics, who arguably has the most advanced humanoid robot, has even opted to explore other designs in their commercially available warehouse robot and robotic dog spot.
Agility Robotics is going a different route with Digit, their commercially available humanoid robot designed to work in an existing environment made for humans.
“Imagine robots thoughtfully designed to be helpful and approachable, constructed to work beside people, in spaces built for people.”
The future of robotics, humanoid or any other form factor is super exciting!
🤖 Artificial Intelligence News:
- Recreating Natural Voices for People with Speech Impairments
- Should retail robots go humanoid?
- For Patients to Trust Medical AI, They Need to Understand It
- Can AI Direct Movies? This One Just Did
- AI Future: Why The University Of Florida Added 100 AI Faculty And The 22nd Fastest Supercomputer In The World
- AI computers can’t patent their own inventions — yet — a US judge rules
🛠️ Developer Tools & Education:
- Detecting Abnormal Chest X-rays using Deep Learning
- Introducing Omnimattes: A New Approach to Matte Generation using Layered Neural Rendering
- Discovering Anomalous Data with Self-Supervised Learning
- Detecting and OCR’ing Digits with Tesseract and Python
🎤 Interesting Podcasts & Interviews:
- Wojciech Zaremba: OpenAI Codex, GPT-3, Robotics, and the Future of AI
- Advancing Robotic Brains and Bodies with Daniela Rus
- Neural Synthesis of Binaural Speech From Mono Audio with Alexander Richard
- Stellar inference speed via AutoNAS
- Statistical Programming with Friends
📄 Notable Research Papers:
- ∞-former: Infinite Memory Transformer
- Whole Brain Vessel Graphs: A Dataset and Benchmark for Graph Learning and Neuroscience (VesselGraph)
- Digging into Uncertainty in Self-supervised Multi-view Stereo
- Seeing Implicit Neural Representations as Fourier Series
- Eyes Tell All: Irregular Pupil Shapes Reveal GAN-generated Faces