VictorLetzelter's Projects
This repository contains an algorithmic trading project whose aim is to detect automatically divergences between a given asset, an a given momentum indicator. The algorithm is still to improve, and its efficiency and robustness to be tested (See the Notebook file in the Trading_project folder).
This project has been realised in cooperation with two students : Hugo BESSON and Adrien CORTES. We tried to understand the analogy between the Brownian Motion used in physics, and the geometric Brownian used in financial mathematics. For those who are interested in the field, we provide you with our technical report.
Code for associated with the NeurIPS23 publication 'Resilient Multiple Choice Learning: A learned scoring scheme with application to audio scene analysis'.
Geometric loss functions between point clouds, images and volumes
[ECCV 2020] In-Domain GAN Inversion for Real Image Editing (PyTorch code)
Code, Dataset, and Pretrained Models for Audio and Speech Large Language Model "Listen, Think, and Understand".
Some of the projects performed during my master.
The official repository for the CVPR 2019 paper "Overcoming Limitations of Mixture Density Networks: A Sampling and Fitting Framework for Multimodal Future Prediction"
This project consists on a technical study on the Stellarator TJ-II, an experimental device allowing to carry out experiments of turbulent plasmas at high temperature confined by magnetic fields. More precisely, an algorithm for the automatic detection of patterns in the time series of the electrostatic potential measured in the plasma was constructed. For this purpose, a probabilistic model was designed to simulate additional data to those provided and to train a neural network. These patterns correspond to specific events within the reactor: the "Zonal Flows"; this detection is relevant insofar as the understanding and detection of these hazards allows the improvement of the magnetic confinement of the plasma.
Audio preprocessing framework for Deep Learning audio applications
Fast EMD for Python: a wrapper for Pele and Werman's C++ implementation of the Earth Mover's Distance metric
Tensors and Dynamic neural networks in Python with strong GPU acceleration
A scalable template for PyTorch projects, with examples in Image Segmentation, Object classification, GANs and Reinforcement Learning.
PyTorch deep learning projects made easy.
scikit-learn: machine learning in Python
Benchmark for sound event localization task of DCASE 2019 challenge
Baseline method for sound event localization task of DCASE 2022 challenge
Sound event localization, detection, and tracking of multiple overlapping and moving sources in 2D spherical space using convolutional recurrent neural network
StyleGAN2-ADA - Official PyTorch implementation
A best practice for tensorflow project template architecture.
Config files for my GitHub profile.
Academic website.
Official code of the ICML24 paper: "Winner-takes-all learners are geometry-aware conditional density estimators"