Théo Gachet's Projects
This repository contains the Julia code to optimize the placement of surveillance points to cover specific targets while avoiding obstacles in a 2D grid. The goal is to minimize the number of surveillance points needed to effectively monitor all targets in the grid.
This repository contains multiple Python implementations of algorithms aimed at solving the surveillance optimization problem. The goal is to strategically position surveillance agents to monitor targets while minimizing the number of agents used.
Here are some ressources for the conferences I give at my engineering school, mostly about Machine Learning (course level: Master)
In May 2023, I took part in the 404 CTF Hackathon where I endeavored to decrypt a code using the prowess of artificial intelligence. Given a folder of handwritten digit images, my task was to create an AI model adept at detecting and classifying these digits. The project showcases the might of AI in both code decryption and pattern recognition.
By building a prediction model based on DNA sequences, I wanted to predict the influence of genetic variation on gene regulation mechanisms, such as DNA receptivity and splicing in eukaryotic cells.
This project simulates fly vision and behavior through a bio-robot using biomimicry and advanced computation. We replicate fly patterns and movements, focusing on their eye photoreceptors' interaction with the environment. Precise calibration and Gaussian sensitivity functions accurately mimic photoreceptor responses to different shades of gray.
A machine learning project predicting real estate prices in France using property attributes like number of rooms, area crime rate, and education level. Featuring detailed data visualization and uncertainty measures with 95% prediction intervals.
This project focuses on automating the translation of French to English texts using deep learning models, specifically Seq2Seq with attention mechanisms. It encompasses the entire workflow, from data preparation to tokenization, and through to text vectorization and model training.
This repository contains the implementation of a Portuguese-to-English translation model. Built on TensorFlow, it utilizes the Transformer architecture to translate sentences from Portuguese to English with high accuracy.
Présentation du fonctionnement de la modularité en C
Welcome to the repository for my image classification project using Convolutional Neural Networks (CNN). The purpose of this project is to classify medical scans into different categories, like "normal" and "pneumonia".
This project is an autonomous object-following robot project. The robot is powered by a PIC18 microcontroller and uses infrared (IR) sensors to detect objects or people within a certain distance range (40cm to 150cm) and follows them at a similar speed.
This project aims to develop an autonomous rover capable of navigating its environment, avoiding obstacles, and adapting to changing wind and environmental conditions. The rover is designed to perform various tasks, including data collection and environmental monitoring.
TESS is a collaborative project by Kawtar El Mamoun and myself, aiming to create a user-friendly human-machine interface with intelligent word suggestions for incomplete input. Utilizing hash tables, TESS captures frequently used words, ensures spelling accuracy, and supports custom names and acronyms.
Collaborative project with Kawtar El Mamoun to develop a C application for preprocessing time series. Aims to predict and clarify the evolution of the average surface material removal rate using historical data.
This project was developed as part of a class project by Timothé Dupuch and myself. The aim of the project was to create a multi-threaded application inspired by the popular game "Questions pour un Champion" to simulate a quiz game.