Galen J. Wilkerson: AI & Complex Systems

Flash-Attention-Tutorial

Flash Attention Tutorial Welcome to the Flash Attention Tutorial repository! This repository provides an in-depth tutorial on the concept of Flash Attention, including high-level intuition, detailed explanation, and practical implementation. Repository Overview Notebook: Flash Attention.ipynb This Jupyter Notebook contains all the tutorial content, including explanations and code implementations. Contents Introduction...

CatHerder-reflective-LLM-code-creator-and-debugger

Simple Code Creator & Debugger using LLM Reflection, an agentic framework to improve transformer results.

Socratic-Learn-Anything

This is a program that uses LLM reflection to interact with chatGPT, creating a Socratic dialog about any topic, gradually building up from basic intuitions. This should pique your imagination about what can be possible with LLMs.

LoRA-Low-Rank-Adaptation

A quick tutorial on LoRA - Low Rank Adaptation for efficient fine tuning of transformers.

Transformers-on-Genome-Sequences

Transformers on Genome Sequences This repository demonstrates the use of transformers, including pre-trained BERT and DNABERT, to analyze and visualize genome sequences. The project downloads mitochondrial DNA sequences from various organisms, preprocesses them, and uses these transformer models to generate embeddings. These embeddings are then visualized using UMAP and clustered...

VIsualizing-Attention-Matrices

Visualizing Attention Matrices The attention matrix is like an adjacency matrix that tells you how much each token should attend to the other tokens for contextual meaning.. So, think of it as a graph where each word has connections of various strengths to other words. This histogram shows the distribution...

Radial-Visualization-of-Semantic-Embeddings

Radial Visualization of Semantic Embeddings This repository contains a Jupyter notebook that demonstrates the radial visualization of semantic word embeddings. The visualization is animated to show how word embeddings of a given sentence change over time. Overview This project uses pre-trained GloVe embeddings to create a radial animation of word...

Logistic-Regression

simple tutorial on logistic regression

Solving Min-Cut Using Oscillatory Networks

Oscillatory Networks

Linear-Threshold-Model

Linear Threshold Model simulation. First and Second: Creating a network then activating seed nodes as the inputs, then running cascades yields patterns of activation corresponding to Boolean logic functions at network nodes. Third: The functions computed in this LTM. Third: Cascade sizes correspond to percolation - formation of the Giant...

linear-threshold-model-on-network-topologies

Linear Threshold Model on Network Topologies This repository contains an implementation of the Linear Threshold Model (LTM) on various network topologies using D3.js for visualization. The available network topologies include random, small-world, and preferential attachment. You can adjust parameters such as the number of nodes, minimum and maximum thresholds, mean...

robot-swarm---guided-self-organized-criticality

Robot Swarm Simulation for Oil Spill Cleanup This repository contains the code for a robot swarm simulation that uses percolation and blocking mechanisms of network cascades to control robot swarming. The primary goal is to efficiently clean up oil spills through a guided self-organized criticality approach. This simulation showcases how...

global_cascades

Global Cascades Simulation This repository hosts the code for a "Global Cascades Simulation" based on Watts' influential paper, "A simple model of global cascades". The simulation visualizes the frequency of cascades across different network densities and provides insights into how small initial shocks can lead to widespread effects in a...

robustness-in-large-scale-systems

Robustness in Large-Scale Systems This repository explores the robustness of large-scale systems through the lens of network theory, focusing on the properties of scale-free networks. The key metric investigated is the scaling exponent, a measure of the degree distribution's power-law behavior in such networks. Overview Large-scale systems often exhibit scale-free...

social-influence-game

The game of social influence - add or remove links to maximize your income.

LTM-logic-interactive

LTM Simulation Run the Simulation This repository contains an interactive simulation of the Linear Threshold Model (LTM), which demonstrates the logic functions computed by the model. The LTM represents one of the simplest models of the emergence of information processing in systems and is considered universal due to its connection...

Benfords-Law

A tutorial on how Benford's Law emerges out of multiplicative processes.

Birthday-Paradox-n-other-problems

Birthday Paradox and other Combinatorial and Probability Problems This repository contains a Jupyter Notebook that explores and visualizes the Birthday Paradox and other Combinatorial Problems. The Birthday Paradox is a famous problem in probability theory that demonstrates how our intuition about probabilities can be misleading. It states that in a...

Monty-Hall-Problem

Monty Hall Problem This repository contains a Jupyter Notebook that demonstrates and simulates the Monty Hall Problem, a famous probability puzzle based on a game show scenario. Introduction The Monty Hall problem is a probability puzzle named after the host of the television game show "Let's Make a Deal." The...

micro-canonical-model

A simple web implementation of a single instance of the micro-canonical model.

wavelet-transform-modulus-maxima-WTMM-tutorial

The Wavelet Transform Modulus Maxima (WTMM) method is a powerful tool in signal processing and fractal analysis. It allows for the detection of singularities in signals—points where the signal behaves unusually, such as sharp transitions or changes in frequency. These singularities often correspond to critical features in the data, making WTMM particularly useful in various fields, including physics, biology, and finance.