50+ days of Machine Learning

 Hello :) Today is Day 103! A quick summary of today: continued with XCS224W: ML with Graphs  and covered 2/3rds of module 5: Knowledge Graphs 1st part is about ML with Heterogeneous graphs, and 2nd part is about Knowledge Graph embeddings. ML with Heterogeneous graphs Covered topics: heterogeneous graph,s Relational GCNs, benchmarks for heterogeneous graphs, Heterogeneous Graph Transformers, designing, heterogeneous GNNs Knowledge Graph embeddings Covered topics: why do we need KGs, KG completion task, connectivity patterns in KGs, TransE, TransR, DistMult, ComplEx, KG embeddings in practice That is all for today! See you tomorrow :)

 Hello :) Today is Day 102! A quick summary of today: continued with XCS224W: ML with Graphs  and covered the next part - Label propagation Before I share my notes, I saw that all the 3 major cloud providers - AWS, GCP and Azure have beginner cloud certification. But I am yet to decide in which one to invest more time and eventually money for the exam. One the one hand, yes it is basic intro to the respective company's cloud platform, but it also can show some initiative and show that I know at least something about the cloud to potential employers. I made post on 2 subredits (learnmachinelearning and deeplearning) and asked on 2 discord servers, but so far not many people voted, but AWS seems to be winning. I will keep this certification in mind. Besides that, continued with the XCS224W: ML with graphs course and covered the next module which is about label propagation, and below are my notes. Covered topics: correlations in networks (homophily and influence), how do we lever...

 Hello :) Today is Day 101! A quick summary of today: did GCP's Digital Leader Learning Path After yesterday's attempt and failure to deploy an LLM to production, I wanted to learn a bit more about the cloud platforms used. The top 3 are GCP, Azure and AWS. From past experiences (in my previous job at Lloyds Banking Group) I heard many people using GCP (some Azure), but also that AWS is very popular. I decided to check out what kind of learning paths GCP and AWS have, and I found the GCP's Digital leader learning path and AWS' Cloud Practitioner: Foundamental  and both have certification.  GCP's Digital Leader Learning Path covers the below topics: Digital Transformation with Google Cloud Exploring Data Transformation with Google Cloud Innovating with Google Cloud Artificial Intelligence Modernize Infrastructure and Applications with Google Cloud Trust and Security with Google Cloud Scaling with Google Cloud Operations and below are my notes. The cloud The cloud ref...

 Hello :) Today is Day 100! A quick summary of today: finished Vicki Boykis' 'What are embeddings?' read 2 papers Graph Convolutional Neural Networks for Web-Scale Recommender Systems (Ying et al., 2018) TwHIN: Embedding the Twitter Heterogeneous Information Network for Personalized Recommendation (El-Kishky et al., 2022) tried to deploy a PDF chat app using streamlit (spoiler: failed) My full notes from Vicki Boykis' What are embeddings? book The research papers below, I saw them in Vicki Boykis' vook and read them because they referenced GNNs. Graph Convolutional Neural Networks for Web-Scale Recommender Systems (Ying et al., 2018) The paper talks about a a large-scale deep recommendation engine developed and deployed at Pinterest called PinSage which tackles the challenge of scaling deep neural networks for graph-structured data to web-scale recommendation tasks with huge amounts of users and items (pictures in pinterest's case). PinSage utilizes an efficient...

 Hello :)  Today is Day 99! A quick summary of today: covered the next part of XCS224W: ML with Graphs Instead of doing the 3rd homework which is related to createing and training a GNN, I decided to continue with the next theory part, which funny enough is called 'theory of GNNs'.  Covered topics: Expressiveness of GNNs, computational graphs, problems with GCN and GraphSAGE's mean and max-pooling, designing the most expressive GNN, the Universal Approximation Theorem, Graph Isomorphism Network (GIN) That is all for today! See you tomorrow :)