This week in AI & Machine Learning: Ego4D teaches AI to perceive the world, detecting colon cancer, robots used in nursing homes, Dermatologist uses for TensorFlow.js, responsible AI and more!
I have two upcoming workshops this month! You can always see future events at plainsight.ai/events
- Faster Data Annotation for Computer Vision with AI-Powered Labeling | October 20th at 5:30 pm PDT
- Intro to Computer Vision: Building Object Detection Models and Datasets | October 27th at 5:30 pm PDT
My Top AI Highlight:
Facebook AI Announces the Ego4D dataset for Teaching AI to Perceive the World from a First-person Perspective
Creating computer vision applications to understand, navigate and interact with the world from a first-person perspective has been a difficult task, but one that’s becoming increasingly important for new applications of augmented reality, virtual reality, and robotics.
Facebook AI is hoping to make research into these tasks easier with Ego4D, the world’s largest first-person dataset. It contains over 2200 hours of unscripted activities captured by 700 people in 9 different countries.
The dataset hopes to set new benchmarks for five complex interaction tasks: Episodic memories, object manipulation, audio visual conversation transcription, and social interaction. Learn more about how Ego4D can help accomplish this in the video below or the official Facebook AI blog post.
🤖 Artificial Intelligence News:
- Northwest Health now using AI to detect colon cancer in ‘first for Indiana’
- New robots to begin work in Minnesota nursing care center
- AI and maths to play bigger role in global diplomacy, says expert
- The data economy: How AI helps us understand and utilize our data
- Building better startups with responsible AI
- AI fake-face generators can be rewound to reveal the real faces they trained on
🛠️ Developer Tools & Education:
- How DermAssist uses TensorFlow.js for on-device image quality checks
- Teaching AI to perceive the world through your eyes
- Azure AI empowers organizations to serve users in more than 100 languages
- An ML-Based Framework for COVID-19 Epidemiology
- Self-Supervised Learning Advances Medical Image Classification
- State of Data Science and Machine Learning 2021
📅 Upcoming Online AI & Data Science Events:
- Faster Data Annotation for Computer Vision with AI-Powered Labeling (Oct 20 5:30 pm PDT)
- Google AI/ML Workshop — MLOps and Vertex AI (Oct 21 9:00 am PDT)
- Building and Scaling Robust Zero-code IoT Streaming Data Pipelines (Oct 26 10:00 am PDT)
- Intro to Computer Vision: Building Object Detection Models and Datasets (Oct 27 5:30 pm PDT)
- Semantic Search and Neural Information Retrieval (Nov 9 12:00 pm PDT)
🎤 Interesting Podcasts & Interviews:
- Building Technical Communities at Stack Overflow with Prashanth Chandrasekar
- Advancing Deep Reinforcement Learning with NetHack, w/ Tim Rocktäschel
- SDS 513: Transformers for Natural Language Processing — with Denis Rothman
- Federated Learning
📄 Notable Research Papers:
- A Few More Examples May Be Worth Billions of Parameters
- Self-supervised Learning is More Robust to Dataset Imbalance
- The Dawn of Quantum Natural Language Processing
- HUMAN4D: A Human-Centric Multimodal Dataset for Motions and Immersive Media
🐘 About the Author & Plainsight:
- Sage Elliott is a Developer Evangelist at Plainsight & passionate about making AI more approachable. Connect with Sage on Twitter or LinkedIn.
- Plainsight’s vision AI platform streamlines the end-to-end machine learning process. From data annotation through deployment, customers quickly create and successfully operationalize their own vision AI applications to solve highly diverse business challenges. Join the conversation on Slack.