ActiveFitness.ai

AI in Sports with Python course has recently been approved by ISTA Professional Sports Technology Institute for sports and technology professionals. This approval shows how important AI becomes for sports technology helping data scientists and sport professionals and changing the world of sports and athletic performance. Taking the course will give you access to many examples that you can use in your careers as data scientist or a sport practitioner working with health, fitness, sports or human movement analysis. If you are in the sport technology space, ISTA offers discounted access to my course for its members.

ISTA’s Professional Sports Technology Institute™ advances careers, improves athletic success and further matures the field of sports technology through recognized standards, certifications, accreditation, partner resources, tools, research, publications, courses and career development guides.

AI in Sports with Python is a comprehensive course and guide to using machine learning methods in sports. It’s for anyone interested in sports, artificial intelligence, and new applications for sports and fitness. Created by Kevin Ashley, Microsoft developer Hall of Fame engineer, a professional instructor and author of popular sports apps, this course complements his new book “Applied Machine Learning for Health and Fitness”, and has been reviewed by Olympic coaches and athletes, data scientists and engineers. Learn AI with Python with fun applications in many sports: tennis, surfing, skiing, snowboarding, skateboarding, football, gymnastics, basketball, javelin, weightlifting, track and field and much more! 

The course is accompanied by practical step-by-step Python code samples and Jupyter notebooks. These models and methods can be used to create solutions for AI enhanced coaching, judging, athletic performance improvement, movement analysis, simulations, in sports, health and fitness, motion capture, gaming, cinema production and more.

Check http://activefitness.ai and http://ai-learning.vhx.tv/

It’s been only 60 days after I launched a Kickstarter campaign on my video course AI in Sports with Python supplementing my book ‘Applied Machine Learning in Health and Fitness’ and it’s been an amazing success! The campaign has now closed with more than one hundred supporters backing my project, raising more than 450% of what was initially requested to fund it, providing valuable feedback and reviewing videos. I’ve been working with startups and ISVs in California and worldwide for almost 10 years, and this is a great proof that the entrepreneurial spirit drives this economy. No matter how big or small the project is, if it’s creative and provides value, it’ll fly! THANK YOU!

During these days of your support, I recorded 20 episodes out of about 25 episodes planned (about 80%), improving quality, light, cinematography, editing, learning along the way! I also listened to your feedback and comments, sending screeners to those of you who volunteered to review. The course is scheduled to launch together with the book by the end of August, but early episodes will be made available for those who supported the project during the campaign soon.

Now… that the Kickstarter caompaign is complete, here’s the link to the course. Ta-Da!

Release Dates and Availability:

• August 15 – Completed course episodes will be released to those with early access https://ai-learning.vhx.tv/
• September 1 – Public availability of the course https://ai-learning.vhx.tv/

Book Availability: The book availability is listed on the publisher’s web site, which says August 17 (https://www.apress.com/us/book/9781484257715). The ebook will be available a bit sooner (a few weeks) than the printed book. If you ordered the printed book, it’ll come separately once available. Those who requested signed rewards will receive once I get printed copies for you to sign and ship 🙂

I’m on track to record the last episodes (in italics):

Video Episodes Recording Schedule

• Module 1 – Getting Started
  Episode 1.1 – Getting Started
  Episode 1.2 – Data Scientist Toolbox
• Module 2 – Physics of Sports
  Episode 2.1 – Sensors
  Episode 2.2 – Kinematics
  Episode 2.3 – Figure Skating
• Module 3 – Neural Networks
  Episode 3.1 – Neural Networks
  Episode 3.2 – Simple Neural Network
  Episode 3.3 – Multi-Layer Neural Network
• Module 4 – Deep Vision
  Episode 4.1 – Computer Vision
  Episode 4.2 – Classification
  Episode 4.3 – Detection
  Episode 4.4 – Segmentation
  Episode 4.5 – Human Body Joints
• Module 5 – Human Body Pose Estimation
  Episode 5.1 Pose Estimation Methods: 2D and 3D
  Episode 5.2 Surfing
  Episode 5.3 Detecting Skill Level
  Episode 5.4 Multi-person pose estimation
• Module 6 – Video Action Recognition
  Episode 6.1 Action Recognition Methods, Datasets, Models
  Episode 6.2 Video Classification
• Module 7 – Reinforcement Learning
  Episode 7.1 – Reinforcement Learning
  Episode 7.2 – Skateboarding
  Episode 7.3 – Gymnastics
  Episode 7.4 – Human Models
• Module 8 – Machine Learning in the Cloud
  Episode 8.1 – Working with AI in the cloud
  Episode 8.2 – Training models in the cloud
  Episode 8.3 – Managing machine learning in the cloud
  Episode 8.4 – Taking your AI to production

Estimated length: 3 hours

It’s been a great pleasure working with a large team of enthusiastic engineers, data scientists and sport professionals at Microsoft Global Hackathon 2020, together with US Olympic Committee and USA Weightlifting team. One of the amazing opportunities at Microsoft is learning everyday something new and working with talented, diverse and enthusiastic team. This year we focused on AI in Sports, and many of us brought ideas, talent and passion towards the goal of helping athletes working out from home. Amazing stuff!

Great news, I just received my book proof from the publishers, APRESS: after reviewing it this week, the book goes to printing, earlier than scheduled! Huge thanks to everyone who contributed, reviewed it and backed my AI in Sports with Python video course that accompanies the book. There’s still one week left in the Kickstarter to back the video course to get a sweet discount, so you still have a chance to be among the supporters!

To thank my supporters, I added new episodes (the idea is to have the book and the video course released at the same time). Here’s what my recording schedule looks like in my home studio, I recorded 18 episodes, covering 5 modules of the course (those in progress are in italics). Once the course and the book is ready, it’ll go live on Vimeo:

Video Episodes Recording Schedule

• Module 1 – Getting Started
  Episode 1.1 – Getting Started
  Episode 1.2 – Data Scientist Toolbox
• Module 2 – Physics of Sports
  Episode 2.1 – Sensors
  Episode 2.2 – Kinematics
  Episode 2.3 – Figure Skating
• Module 3 – Neural Networks
  Episode 3.1 – Neural Networks
  Episode 3.2 – Simple Neural Network
  Episode 3.3 – Multi-Layer Neural Network
• Module 4 – Deep Vision
  Episode 4.1 – Computer Vision
  Episode 4.2 – Classification
  Episode 4.3 – Detection
  Episode 4.4 – Segmentation
  Episode 4.5 – Human Body Joints
• Module 5 – Human Body Pose Estimation
  Episode 5.1 Pose Estimation Methods: 2D and 3D
  Episode 5.2 Surfing
  Episode 5.3 Detecting Skill Level
  Episode 5.4 Multi-person pose estimation
  Episode 5.5 Pose Reconstruction in 3D
• Module 6 – Video Action Recognition
  Episode 6.1 Action Recognition Methods, Datasets, Models
  Episode 6.1 Video Classification
  Episode 6.1 Training Action Recognition Models
• Module 7 – Reinforcement Learning
  Episode 7.1 – Reinforcement Learning
  Episode 7.2 – Skateboarding
  Episode 7.3 – Gymnastics
  Episode 7.4 – Human Models
• Module 8 – Machine Learning in the Cloud
  Episode 8.1 – Working with AI in the cloud
  Episode 8.2 – Training models in the cloud
  Episode 8.3 – Managing machine learning in the cloud
  Episode 8.4 – Taking your AI to production

Stretch Modules: Weightlifting

Thanks to all my supporters, my AI in Sports with Python course progresses at a steady pace to the release this summer, along with the book! I now recorded about half of the course, 14 episodes, grouped into modules (those in progress are in italics). If you support this project on Kickstarter, you are getting a pretty sweet deal with early access and more than 50% discount *. I tested streaming video episodes (hosted on Vimeo), and sent some initial screeners to my reviewers. We are now backed at 200% and growing, so I as stretch goals, you’ll be getting more episodes and samples!

Video Episodes Recording Schedule

• Module 1 – Getting Started
  • Episode 1.1 – Getting Started
  • Episode 1.2 – Data Scientist Toolbox
• Module 2 – Physics of Sports
  • Episode 2.1 – Sensors
  • Episode 2.2 – Kinematics
  • Episode 2.3 – Figure Skating
• Module 3 – Neural Networks
  • Episode 3.1 – Neural Networks
  • Episode 3.2 – Simple Neural Network
  • Episode 3.3 – Multi-Layer Neural Network
• Module 4 – Deep Vision
  • Episode 4.1 – Computer Vision
  • Episode 4.2 – Classification
  • Episode 4.3 – Detection
  • Episode 4.4 – Segmentation
  • Episode 4.5 – Human Body Joints
• Module 5 – Human Body Pose Estimation
  • Episode 5.1 Pose Estimation Methods: 2D and 3D
  • Episode 5.2 Surfing
  • Episode 5.3 Detecting Skill Level
  • Episode 5.4 Multi-person pose estimation
  • Episode 5.5 Pose Reconstruction in 3D
• Module 6 – Video Action Recognition
  • Episode 6.1 Action Recognition Methods, Datasets, Models
  • Episode 6.1 Video Classification
  • Episode 6.1 Training Action Recognition Models
• Module 7 – Reinforcement Learning
  • Episode 7.1 – Reinforcement Learning
  • Episode 7.2 – Skateboarding
  • Episode 7.3 – Gymnastics
  • Episode 7.4 – Human Models
• Module 8 – Machine Learning in the Cloud
  • Episode 8.1 – Working with AI in the cloud
  • Episode 8.2 – Training models in the cloud
  • Episode 8.3 – Managing machine learning in the cloud
  • Episode 8.4 – Taking your AI to production

There’s still time to support this project and get the course at as huge discount as a reward for early support*!

• * included with some rewards

I learned a lot about home video recording, lighting and production, and believe that videos are becoming better with each episode!

As part of my daily work, I’m also working on new video series for Microsoft Azure! Also at the end of July I’ll be driving a team of Microsoft engineers at OneWeek Hackathon 2020, a yearly coding for good event helping US Olympics Team. Working on this course definitely helps me working on the next series for Microsoft Azure.

Thanks for your support of this hard work and innovation!

Kevin

Thanks to everyone supporting my video course project “AI in Sports with Python“, I already recorded 12 episodes (about 15 episodes to go!). You can support this effort here:

Back this project!

It’s a lot of work, but it’s very rewarding. Here’s a preview of my studio!

Thanks to all your support, my video course project that complements my book on “Applied Machine Learning for Health and Fitness” is now fully funded on Kickstarter! I added some stretch goals, to include more Python examples for other sports, please let me know your ideas! Everyone has a favorite sport.

I just launched a video course AI IN SPORTS WITH PYTHON. This is a fun new video course that complements and extends my book ‘Applied Machine Learning for Health and Fitness’  through videos and new examples. Learn AI with Python with fun applications in many sports: tennis, surfing, skiing, snowboarding, skateboarding, football, gymnastics, basketball, track and field and much more! 

Who is this course for?

This course is for anyone interested in AI, data science, gadgets and Python.  Whether you are new to machine learning or an experienced data scientist, the course contains practical examples for all levels: beginners and advanced. The goal is to make AI easy to learn. My book has been reviewed by Olympic level coaches, sport scientists and machine learning experts. Enjoy!

Since this is a video course, I want to take advantage of the video format and plan for some bonus stretch examples and exercises.

Content and format

The course will provide a recorded video guidance, adding video episodes for 12 sections, corresponding to the chapters of my book. Each section will be split into several segments of 5-10 minutes, with a total of about 2-3 hours of video:

• Machine Learning in Sports 101 (Supervised, unsupervised, reinforcement learning – Logic and machine learning – Tools – Neural networks – Deep vision – Classification – Detection – Semantic segmentation – Sensors – Reinforcement learning)
• Physics of Sports (Mechanics – Kinetics – Laws of motion – Inertia – Kinematics – Projectile motion – Using neural networks to predict a projectile range – Angular motion – Conservation laws – Energy, work and power – Physics and deep learning)
• Sports Scientist Toolbox (Data science tools – Python – Virtual environments – Packaging – Anaconda – Pip – Jupyter notebooks – Numpy – Pandas – Visualizations – Matplotlib – SciPy – Scikit-image – OpenCV – PyTorch – Keras – Tensorflow – OpenAI Gym – Pybullet)
• Neural Networks (Neurons – Activation – Perceptron – Training a perceptron in Python – Multi-layer networks – Backpropagation)
• Sensors (Deep Vision – Edge devices – Inertial movement sensors IMUs – Attitude and heading reference systems AHRS – Inertial and navigation systems GNSS – Range Imaging Sensors LIDAR – Pressure sensors – EMG sensors – Heart rate sensors )
• Deep Computer Vision (Neuroscience and deep learning – Computer vision – Visual datasets – Model zoo – Applying models – Classification – Classifying sport activity type – Detection – Segmentation –Semantic segmentation – Human body keypoint detection)
• 2D Pose Estimation (Methods – Neural networks – Datasets – Tools – Body pose estimation – Detecting athlete stance – Activity recognition – Detecting skill level – Multi-person pose estimation – Dealing with loss and occlusion)
• 3D Pose Estimation (Cameras and 3D – Camera Matrix – 3D Reconstruction – Using a single camera – Multi-view depth reconstruction – 3D reconstruction with sensors – Motion capture – 3D Datasets – 3D Machine learning methods – Sparse and dense reconstruction)
• Video Action Recognition (Video Data – Datasets – Models – Video classification – Action recognition – Loading videos for classifier training – Visualizing datasets – Video normalization – Training video recognition model)
• Reinforcement Learning (Tools – Applying reinforcement learning in sports – Action and observation spaces – Visualizing sample motion – Model zoo – Models – Reinforcement learning in gymnastics – Pendulum model – Humanoid models – Joints and action spaces – Human motion capture – Mocap – Reinforcement learning in humanoids)
• Machine Learning in the Cloud (Containers – Docker – Notebooks in the cloud – Data storage and datasets in the cloud – Loading and accessing datasets – Labeling data in the cloud – Training classification model – Preparing for training – Running experiments – Model management)
• Automating and Consuming Machine Learning (CI/CD – MLOps – Managing models – Creating a scoring script – Defining an environment – Deploying models – Consuming models – APIs – Machine learning pipelines)

Episodes

I created some pilot episodes, and have a good idea of the structure and the time it takes to record each episode. Because this course is based on sports, it’s really fun and visual! As a bonus to my backers, I’ll provide early access to episodes as I make them. This course is very practical, so for each episode I have a number of examples in Jupyter notebooks: in the beginning of each episode I start with goals and then walk the reader through the examples.

Why AI and Sports?

I worked with top Olympic athletes and professional sports organizations around the world, advising on technology and AI. Sports is fun! And we have a lot of data in sports, which makes it very valuable for anyone interested in AI. As a professional ski instructor, who also happened to be an engineer, I helped athletes with all abilities, and I believe using machine learning in sports can change the world! This project is a contribution to our understanding of sports with AI, machine learning and Python. AI can help in sports at any level: from any beginner learning a new sport to an elite Olympic level athlete. 

When I started working on my ‘Applied Machine Learning for Health and Fitness‘ book, the concept I had in my mind is making AI fun. Drawings, sketches and illustrations are part of why I like working on this book so much: it’s fun to make complex concepts easy to understand, and sketches definitely achieve the goal! A complex concept of CI/CD, automating ML and deploying machine learning models in the cloud becomes easy when you look at the sketch with a friendly Python in a container getting shipped to the cloud.

One thing about illustrating my book I discovered: towards the end of the book I really started to get better! My toolset for drawing also improved and now includes a lot more pencils and markers, as well as good sketchbooks with paper I enjoy drawing on.

I hope these illustrations will help you and make data science and concepts explained easier to understand. Another cool thing about this book, with the source code structured as Jupyter notebooks with live Python code, those illustrations really make reading it fun!

I absolutely love the new Microsoft Edge browser logo, that looks like a surf wave. I couldn’t resist and drew this surfer next to it (right after getting outside to ride some Pacifica waves). The new Microsoft browser is using Chromium open source engine, and looks really cool. While I still have some time before my book is published I’d love to use this picture in my book!