Machine Learning

Various Bounds in Optimization

by ComputingNotes
Regret Bound:https://www.emergentmind.com/topics/cumulative-regret-analysishttp://sbubeck.com/LecturesALL_Bubeck.pdfhttps://www.jmlr.org/papers/volume11/jaksch10a/jaksch10a.pdfhttps://datascience.stackexchange.com/questions/62141/what-are-regret-boundshttps://arxiv.org/pdf/1403.5556 Lecture 1:Introduction to regret analysisS´ebastien BubeckMachine Learning and Optimization group, MSR AI

Assumptions and Their Meaning in Optimization Problems

by ComputingNotes
1. Lipschitz Continuity: The “Speed Limit”This assumption prevents the function from changing too rapidly over a certain distance. 2. 𝐿-Smoothness: The Curvature CeilingSmoothness ensures the gradient (slope) of the function doesn’t change abruptly. A function is 𝐿-smooth if its gradient is Lipschitz continuous. 3. 𝜇-Strong Convexity : The Curvature Floor – Bowl ShapeStrong convexity guarantees that ... Read More

Federated Learning Algorithm Averaging Methods FedAvgM, FedProx, FedAvg

by ComputingNotes
FedAvg: https://arxiv.org/abs/1602.05629 __________________________________________________________________________________________________________ FedAvgM: https://arxiv.org/pdf/1909.06335 stands for Federated Averaging with Server Momentum. It is an upgrade to the original FedAvg algorithm designed specifically to solve the “Client Drift” problemwhere the model gets confused because different users have very different data (Non-IID). The Formula: How it works mathematically The FedAvgM update happens in three steps on ... Read More

Deep Reinforcement Learning Parameters

by ComputingNotes
Epsilon (ϵepsilon 𝜖): This is a hyperparameter (a value between 0 and 1) that defines the probability of an agent taking a random action (exploration) instead of the action believed to have the highest reward (exploitation). Epsilon Decay: This is the process of gradually reducing the value of 𝜖 over time or a number of ... Read More

All Machine Learning MOdels

by ComputingNotes
https://www.ibm.com/think/topics/machine-learning-types Machine learning models are grouped in one of 5 types, supervised, unsupervised, self-supervised, semi-supervised and reinforcement learning. Machine learning libraries, useful.

MNIST PCA

by ComputingNotes
https://medium.com/@azimkhan8018/a-beginners-guide-to-deep-learning-with-mnist-dataset-0894f7183344https://ranasinghiitkgp.medium.com/principal-component-analysis-pca-with-code-on-mnist-dataset-da7de0d07c22https://github.com/mkosaka1/MNIST_PCA_CNN/blob/main/PCA%20%2B%20CNN.ipynbhttps://gist.github.com/tommct/1490cdf856d745ba41c1ac99ada2b579

Reinforcement Learning

by ComputingNotes
https://au.mathworks.com/content/dam/mathworks/ebook/gated/reinforcement-learning-ebook-all-chapters.pdf Reinforcement Learning: Type of machine learning, where a computer agent learns to perform a task through a process of repeated trial and error within a dynamic environment. The goal is to learn what to do, how to map situations to take certain actions that minimize a numerical reward signal. Eg. RL trained computer gains ... Read More

Optimization

by ComputingNotes
Rosenbrock Function $latex f(x,y) = (1-x)^2 + 100(y-x^2)^2$ https://docs.scipy.org/doc/scipy/tutorial/optimize.html

What is Gradient? Loss Function? Model Parameters? Gradient Descent?

by ComputingNotes
Gradient Descent: Optimization algorithm; minimizes errors between predicted and actual results; updates parameters by moving against the gradient Loss Function: aim is to minimize this function, closer to zero; measures how bad the prediction is in comparison to the actual true value; various methods are used; one is Mean Squared Error (MSE) Gradient: Slope; direction ... Read More

Conda Environment

by ComputingNotes
conda create -n env python=version conda activate env conda deactivate Useful tools: Starting jupyter notebook (env) file_path > jupyter notebook
error: