Discover Deep Crossing model that pioneered deep learning for CTR prediction. Learn its four-layer architecture — Embedding, Stacking, Residual Units, and Scoring — and why it was a revolutionary milestone.
Learn the fundamentals of feature crossing in recommender systems. Explore Simpson's Paradox, POLY2 brute-force crossing, and the latent vector breakthroughs in Factorization Machines (FM) and Field-aware Factorization Machines (FFM).
Explore Logistic Regression — the foundational model that transformed recommender systems from pure collaborative filtering to multi-feature fusion. Learn its mathematical formulation, training process, strengths, and limitations.
Explore the evolution of Collaborative Filtering from UserCF and ItemCF to Matrix Factorization. This post explains how recommender systems leverage user similarity, item relationships, and latent factors to model user preferences.
Migrate personal blog to Astro: Dev Containers for isolated dev environments, Claude Code for vibe coding, GitHub Actions for continuous deployment.
Ch 3 notes for Kubernetes in Action. Learn pod lifecycle basics (create, run, stop), resource organization via labels and namespaces, executing bulk operations on labeled pods, and scheduling workloads onto specific types of cluster worker nodes.
Ch 2 notes for Kubernetes in Action. Key topics include: creating and sharing Docker container images, running a local single-node cluster via Minikube, configuring the kubectl CLI, and deploying/horizontally scaling a Kubernetes application.
Reading Note for *Chapter 01 Introducing Kubernetes* of *Kubernetes In Action*: the history of software developing, isolation by containers,how containers and Docker are used by Kubernetes and how to simplify works by Kubernetes