This talk was presented as part of JuliaCon 2021. Abstract: Learning from raw data input is one of the key components of many successful applications of machine learning methods. While machine learning problems are often formulated on data that naturally translate into a vector representation suitable for classifiers, there are data sources with a unifying hierarchical structure, such as JSON. This talk will describe Mill.jl and JsonGrinder.jl, which offers a theoretically justified approach to solve machine learning problems with these data sources. For more info on the Julia Programming Language, follow us on Twitter: https://twitter.com/JuliaLanguage and consider sponsoring us on GitHub: https://github.com/sponsors/JuliaLang Contents 00:00 Introduction and Motivation 01:07 Possible solutions to the device ID problem 02:52 HMIL approach 06:32 Comparison to SOTA 07:20 Theoretical guarantees 08:29 Mill.jl: node types 09:11 Mill.jl: representing leafs 10:58 Mill.jl: representing dicts 11:59 Mill.jl: representing arrays 13:19 Mill.jl: working with the whole sample 14:26 Mill.jl: training a model with JsonGrinder.jl and Flux.jl 19:58 Summary S/O to https://github.com/SimonMandlik for the video timestamps! Want to help add timestamps to our YouTube videos to help with discoverability? Find out more here: https://github.com/JuliaCommunity/YouTubeVideoTimestamps Interested in improving the auto generated captions? Get involved here: https://github.com/JuliaCommunity/YouTubeVideoSubtitles