training course on ML as part of The Congruence Engine

[amsichani]

As part of The Congruence Engine training program, we are planning to host a course on Machine Learning for GLAM using this course - the session is now planned for the 3rd May, at the MakerSpace, SAS, London. We are excited to have @mark-bell-tna with us co-delivering the lesson. Hopefully we will be able to provide feedback both from participants and instructors.

[mark-bell-tna]

Thank you @amsichani . I have now merged the version with Episode 2 split into two parts. Looking forward to delivering the lesson next week!

[amsichani]

I am posting here a couple of comments and remarks made by our session's participants and instructors - thanks again @mark-bell-tna for this very useful training session and thanks to the entire lesson's team for putting this lesson together.
Please feel free to get in touch if something is not clear .
We can't wait to see the final version of it !

episode 2
what is machine learning ?

• structure a bit more clearly  the 4 types of ML , incl tasks (prediction , regression), to make more evident how they all fit together 
• activities: perhaps more GLAM-focused examples.
• estimated time 20-25min

episode 3 modelling the world 

• amend the scenarios' numbering
• conceptual and trained model: it would be great to have clearer definitions, and perhaps in more structured way (ie bullet points?)
• in the yellow cloud- note, the testing and training dataset bit should be moved into main text as its crucial information p
• erhaps use examples from the GLAM sector , also for the embeddings section
• estimated time 25 min

episode 4

• add data types such as audio, video etc 
• the distinction between Classic AI and Deep Learning is not clear - is it simply chronological (2012-->), technology-related?
  NLP
• it would be great to have more clear structure regarding the stages (perhaps better formatting would do the trick)
• clearer description of model words, word embeddings, model language
• estimated time 35-45 min

episode 5

• estimated time 30min

episodes 6-7
these are still under heavy development so we didn't actually use them - but we 'd love to have a part II for our training!