This week in AI & Machine Learning: Teaching AI to forget, manufacturing growth, wildfire monitoring, Hugging Face Vision Transformers, Conceptual Understanding of Deep Learning, Machine Learning Production (MLOps) and more!
🤖 Artificial Intelligence News:
- Teaching AI how to forget at scale at Facebook “As a step toward achieving humanlike memory in machines, we’re announcing a novel method in deep learning”
- How Artificial Intelligence (AI) Is Helping Musicians Unlock Their Creativity
- This AI tool writes real estate descriptions without ever stepping inside a home ← Listing AI uses GPT-3 from openAI to generate real estate listings using only a small amount of data.
- AI is the Dominant Factor Causing the Growth of Manufacturing ← Artificial intelligence in manufacturing is booming! If you’re interested in finding out how you can start applying computer vision to manufacturing use cases, let us know.
- Intel is using machine learning to make GTA V look incredibly realistic
- Artificial intelligence aids in spotting wildfires ← As fire season approaches on the west coast it’s important to improve monitoring to catch fires as soon as possible. Alertwildfire is using machine learning to detect the difference between clouds, fog and smoke plumes.
🛠️ Developer Tools & Education:
- Hugging Face Vision Transformers ← I’m excited that Hugging face now offers computer vision transformer options!
- Pytorch Lighting implementation of FastFace
- DeepLearning.AI Launches Machine Learning Production (MLOps) Specialization ← this course looks like it may fill a lot of gaps in other machine learning courses!
- OpenCV Eigenfaces for Face Recognition ← I always find it amusing that faces can be recognized fairly well using eigenvectors!
- Using a TensorFlow Python MIRNet model in Node.js
- AI Learned To Perform A Cartwheel…With Style! ← this 2-minute paper explains the “Learning and Exploring Motor Skills with Spacetime Bounds” paper with great visuals!
🎤 Interesting Podcasts & Interviews:
- Probabilistic Numeric CNNs with Roberto Bondesan
- 25 years of speech technology innovation
- What the Human Brain Can Tell Us About NLP Models with Allyson Ettinger
- Learning Deep Learning Together
- Marc Segura talks about the millions of robots already changing industries around the world
📄 Notable Research Papers:
- Diffusion Models Beat GANs on Image Synthesis
- Dynamic View Synthesis from Dynamic Monocular Video
- Contrastive Learning of Image Representations with Cross-Video Cycle-Consistency
- Self-Supervised Collision Handling via Generative 3D Garment Models for Virtual Try-On
- 3D Spatial Recognition without Spatially Labeled 3D
- High-Resolution Complex Scene Synthesis with Transformers
- Using Self-Supervised Co-Training to Improve Facial Representation
- Self-Supervised Learning with Swin Transformers