This week in AI & Machine Learning: Wildfire detection, detecting deepfakes, transforming manufacturing, AugLy data augmentation, inclusive people annotations, and more!
🤖 Artificial Intelligence News:
- AI Could Spot Wildfires Faster Than Humans ←A new fire prediction and monitoring system is being tested as wildfire season is looming closer in the U.S.
- AI-driven robot Mayflower begins Atlantic voyage
- Facebook Researchers Say They Can Detect Deepfakes And Where They Came From ← This could be a big deal in combating deepfakes and fake news!
- Supply chain monitoring: Audi uses artificial intelligence (AI) for sustainability
- Deep Reinforcement Learning will Transform Manufacturing as We Know It
🛠️ Developer Tools & Education:
- AugLy: A new data augmentation library to help build more robust AI models “We are open-sourcing AugLy, a new Python library that will help AI researchers use data augmentations to evaluate and improve the robustness of their machine learning models.”
- A Step Toward More Inclusive People Annotations in the Open Images Extended Dataset
- Learning an Accurate Physics Simulator via Adversarial Reinforcement Learning
- A gentle introduction to tf.data with TensorFlow ← Learn how to build faster, more efficient deep learning data pipelines with tf.data!
- Modeling Pipeline Optimization With scikit-learn
- Intro to graph neural networks (ML Tech Talks) “In this session of Machine Learning Tech Talks, Senior Research Scientist at DeepMind, Petar Veličković, will give an introductory presentation and Colab exercise on graph neural networks (GNNs).”
🎤 Interesting Podcasts & Interviews:
- Waste Not, Want Not: AI Startup Opseyes Revolutionizes Wastewater Analysis
- AI and Society: Past, Present and Future with Eric Horvitz
- Agile Applied AI Research with Parvez Ahammad
- Multi-GPU training is hard (without PyTorch Lightning)
- Charles Isbell makes the case for more ethical AI
📄 Notable Research Papers:
- SimSwap: An Efficient Framework For High Fidelity Face Swapping
- GANs N’ Roses: Stable, Controllable, Diverse Image to Image Translation (works for videos too!)
- Automating Augmentation Through Random Unidimensional Search
- Federated Semi-supervised Medical Image Classification via Inter-client Relation Matching
- Efficient Deep Learning: A Survey on Making Deep Learning Models Smaller, Faster, and Better
- Anomaly Detection in Video Sequences: A Benchmark and Computational Model
- End-to-End Learning of Keypoint Representations for Continuous Control from Images
- ResDepth: A Deep Prior For 3D Reconstruction From High-resolution Satellite Images