AILabby: Bridging Classical and Modern AI
Having been living with GenAI and thinking about the transformer and MoE, which BTW was amazingly origined in this 1991 paper Adaptive Mixtures of Local Experts , the year another dinasour got his master’s degree in classical AI, I suddenly had the urge this Saturday morning to ask AI this question: “give me examples in the deep learning and genai field where one seemingly small change or adoption…
Another 5 years have passed. Here’s a summary of what happened after I wrote in my last post “my mission statement” as “Building True AI with NLU and Common Sense”: I ported all my iOS and Android AI apps built with TensorFlow Mobile to PyTorch Mobile around 2021 - both platforms are long gone now but I had fun and learned a bunch on how to run all kinds AI models - CV, NLP, audio, video - on…
It’s been over 2 years since my last post on building mobile AI apps with TensorFlow. A quick recap of what I’ve been up to during the time, to justify a little bit my mission statement “Building True AI with NLU and Common Sense”: Started learning French by myself in June 2018, for about one hour a day; Did hard thinking on which field of AI I should focus on, and by late 2018, with the release…
Just developed and released a Bilingual French Classics app, as a convergence of my passions for language, AI and mobile, to kill the pain I was having while learning French one hour a day for about two years. The app offers a fun, addictive and effective way to learn French, by letting you read timeless French classics in the bilingual format to keep you in the flow. Improving your vocabulary…
Update June 12, 2018: The book has been officially published and the complete source code is available here . I’m honored to have Pete Warden, the lead of Google TensorFlow mobile team as one technical reviewer of the book, and Aurélien Géron, the best-seller author of Hands-On Machine Learning with Scikit-Learn and TensorFlow, as the Foreword author who kindly writes about my book - “This is…
Late in May, I decided to learn more about CNN via participating in a Kaggle competition called Sealion Population Count . First, I came across this nice Notebook by Radu Stoicescu: Use keras to classify Sea Lions: 0.91 accuracy , and there’s a statement in it that says “This is the state of the art (object detection) method at the moment: https://cs.stanford.edu/people/karpathy/rcnn/”. So I…
Update March 8, 2018: I’m busy writing a book on building iOS and Android apps with TensorFlow and one of the chapters I have completed writing has updated info on this model as well as a detailed tutorial of using the TensorFlow multiple styled model (stylize_quantized.pb from the official TensorFlow 1.4/1.5/1.6 Android example apps) on iOS. The TensorFlow 1.4/1.5/1.6 stylize_quantized.pb model…
If you’re thinking of buying or building your own deep learning machine with some nice GPU for training, you may have come across Andrej Karpathy’s tweet about his deep learning rig build , which was a little outdated, being published in Sep. 2015, or more recently Lukas Biewald’s Build a super fast deep learning machine for under $1,000 , published on Feb. 1, 2017. I happened to start exploring…
Different people may learn in different ways. Some prefer to have a teacher, a mentor, a supervisor, guiding them on each step in their learning process; others are more like self-learners. Supervised learning no doubt is very important, but I found reinforcement learning (RL) fascinating because of its self learning flavor - a program called TD-Gammon developed in 1992 learned how to play…
Even before I had my first dog, a Labrador Retriever, in June 2015, while walking and seeing a dog I often wondered what kind of breed it is. About a year ago, I found the Stanford Dogs Dataset and then asked a friend who had a Ph.D. in computer vision from CMU if it’s possible to use it and some machine learning algorithm on my iPhone and reach a recognition precision of about 80% or 90%, and…