It makes me really want to start collecting all sorts data to analyze.
The list of Labs is as follows and covers OCR, Language Translation, NLP, Face and Landmark detection, Sentiment Analysis, Image Entity Classification (using K8s), implementing a ChatBot and finally training your own custom model with TensorFlow and then uploading it to Google’s “Cloud Machine Learning Engine”:
The courses were light on the theory and really easy to follow along with. For someone not wanting to spend any money on Qwiklabs, a lot of the content/examples you can just do yourself as Google makes it publicly available, for example:
NOTE: You do need a Google Compute account to work through the examples, though there might be a “free tier”
I would highly recommend them to anyone that keeps getting bombarded by the AI/ML hype and wants to actually see what it’s all about and how to use it.
Also, yesterday Google did an announcement for something they’re calling “AutoML” (see: https://www.blog.google/topics/google-cloud/cloud-automl-making-ai-accessible-every-business/) which just shows what a moving target the whole ML/AI industry is at the moment.
The announcement potentially has the impact to make it drastically simpler to create/train your own custom models and might in some sense make the last course listed there somewhat redundant.
Now for the hard part! Figuring out a “real world” use case for the API’s above. Personally I’d be very much interested in the OCR API for digitizing documents and the classification for auto-labeling them.
Also, using sentiment analysis to determine news articles’ about companies and using that information to augment buy/sell stock decisions (I know, it’s being done already… but the purpose is to learn as much as to make any money).