A simple explanation of the Pearson correlation coefficient with examples
This book provides an introductory, textbook-like treatment of multi-armed bandits. It covers various algorithms and techniques for decision-making under uncertainty, with a focus on theoretical foundations and practical applications.
* **Multi-Armed Bandit Framework:** The document introduces the core concept of multi-armed bandits – a model for decision-making under uncertainty, often used as a simplified starting point for more complex reinforcement learning problems.
* **Applications:** It highlights several applications, including news website optimization, dynamic pricing, and medical trials.
* **Key Concepts:** Defines crucial concepts like arms, rewards, regret, exploration vs. exploitation, and different feedback mechanisms (bandit, full, partial).
* **Algorithms:** Presents and analyzes simple algorithms like Explore-First and Epsilon-Greedy.
* **Regret Bounds:** Focuses heavily on bounding the regret of these algorithms, which measures how much worse the algorithm performs compared to always choosing the best arm.
* **Adaptive Exploration:** Introduces the idea of improving performance through adaptive exploration strategies (adjusting exploration based on observed rewards).
* **Clean Event:** Introduces the concept of the "clean event" to simplify analysis by focusing on high probability events.
* **Table of Contents:** Shows a detailed table of contents, indicating the breadth of topics covered in the full book including Bayesian Bandits, Contextual bandits, Adversarial bandits and connection with economics.
* **Multi-Armed Bandit Framework:** The document introduces the core concept of multi-armed bandits – a model for decision-making under uncertainty, often used as a simplified starting point for more complex reinforcement learning problems.
* **Applications:** It highlights several applications, including news website optimization, dynamic pricing, and medical trials.
* **Key Concepts:** Defines crucial concepts like arms, rewards, regret, exploration vs. exploitation, and different feedback mechanisms (bandit, full, partial).
* **Algorithms:** Presents and analyzes simple algorithms like Explore-First and Epsilon-Greedy.
* **Regret Bounds:** Focuses heavily on bounding the regret of these algorithms, which measures how much worse the algorithm performs compared to always choosing the best arm.
* **Adaptive Exploration:** Introduces the idea of improving performance through adaptive exploration strategies (adjusting exploration based on observed rewards).
* **Clean Event:** Introduces the concept of the "clean event" to simplify analysis by focusing on high probability events.
* **Table of Contents:** Shows a detailed table of contents, indicating the breadth of topics covered in the full book including Bayesian Bandits, Contextual bandits, Adversarial bandits and connection with economics.
Multi-armed bandits a simple but very powerful framework for algorithms that make decisions over time under uncertainty. This book provides a more introductory, textbook-like treatment of the subject, covering IID and adversarial rewards, contextual bandits, and connections to economics.
This article explores TinyML, a branch of machine learning run on microcontrollers like the ESP32. It details how TinyML can be used for local inference, anomaly detection, and efficient data processing with minimal power consumption, using an example of temperature and humidity monitoring.
3D simulations and movement control with PyBullet. This article demonstrates how to build a 3D environment with PyBullet for manually controlling a robotic arm, covering setup, robot loading, movement control (position, velocity, force), and interaction with objects.
An exploration of simple transformer circuit models that illustrate how superposition arises in transformer architectures, introducing toy examples and analyzing their behavior.
This paper provides a theoretical analysis of Transformers' limitations for time series forecasting through the lens of In-Context Learning (ICL) theory, demonstrating that even powerful Transformers often fail to outperform simpler models like linear models. The study focuses on Linear Self-Attention (LSA) models and shows that they cannot achieve lower expected MSE than classical linear models for in-context forecasting, and that predictions collapse to the mean exponentially under Chain-of-Thought inference.
This notebook provides an introduction to Naive Bayes classification, covering concepts, formulas, and implementation.
This article explores how prompt engineering can be used to improve time-series analysis with Large Language Models (LLMs), covering core strategies, preprocessing, anomaly detection, and feature engineering. It provides practical prompts and examples for various tasks.
Hierarchical Reasoning Model (HRM) is a novel approach using two small neural networks recursing at different frequencies. This biologically inspired method beats Large Language models (LLMs) on hard puzzle tasks such as Sudoku, Maze, and ARC-AGI while trained with small models (27M parameters) on small data (around 1000 examples). HRM holds great promise for solving hard problems with small networks, but it is not yet well understood and may be suboptimal. We propose Tiny Recursive Model (TRM), a much simpler recursive reasoning approach that achieves significantly higher generalization than HRM, while using a single tiny network with only 2 layers. With only 7M parameters, TRM obtains 45% test-accuracy on ARC-AGI-1 and 8% on ARC-AGI-2, higher than most LLMs (e.g., Deepseek R1, o3-mini, Gemini 2.5 Pro) with less than 0.01% of the parameters.
