How to evaluate model’s performance
Evaluating the performance of machine learning models is a crucial step to understand how well your model is generalizing to new, unseen data. There are various evaluation metrics and techniques…
KNN, Kmeans, mean shift
KNN (K-Nearest Neighbors), K-means, and Mean Shift are all machine learning techniques used for different types of tasks, primarily in clustering and pattern recognition. Let’s break down the differences between…
Different between boost, bagging , xgboost , random forest in ML
Boosting and bagging are both ensemble learning techniques in machine learning that aim to improve the performance of individual models by combining the predictions of multiple base models. They work…
Entropy and cross-entropy
why entropy is defined as H(X)=−∑p(x)log(p(x)) Entropy, as defined by the formula H(X) = -∑p(x)log(p(x)), might seem complex at first, but it has a deep and intuitive connection to information…
ML workflow and pipeline orchestration
Kale – Aims at simplifying the Data Science experience of deploying Kubeflow Pipelines workflows. Flyte – Easy to create concurrent, scalable, and maintainable workflows for machine learning. MLRun – Generic mechanism for data…
responsible AI
General intro https://github.com/alexandrainst/responsible-ai https://ai.google/responsibilities/responsible-ai-practices/ https://www.tensorflow.org/responsible_ai Open source implementation https://github.com/microsoft/responsible-ai-toolbox https://www.tensorflow.org/responsible_ai/api_docs https://opendatascience.com/15-open-source-responsible-ai-toolkits-and-projects-to-use-today/ Responsible AI Toolkits for AI Ethics & Privacy TensorFlow Privacy TensorFlow Privacy is a Python library that includes implementations…
Bias vs Variance in ML
Somehow even wiki https://en.wikipedia.org/wiki/Bias%E2%80%93variance_tradeoff seems not clearly explained what it is due to un-cleary math denotion. Here we refer to: model/algorithm’s bias and variance ( not the data itself ).…
optimization theory, lagrange, dual function, dual problem, KKT
Theory background min: f0(x) ==> p * s.t.: fi(x) <= 0 i = 1, … m hi(x) = 0 , j= 1 … m define lagrange function:…
webassembly and web audio worklet
Some notes: web assembly: a cool technology allows you to compile c/c++/rust and other languages into wasm, expose API to javascript world. Web audio worklet/worker: allows developer to intercept audio…
