K-Schubert
I am a M.Sc. Statistics graduate from the University of Geneva (CH) with a passion for NLP and deep learning.
University of GenevaGeneva, Switzerland
Pinned Repositories
BERT_Financial_Sentiment_Analysis
Using Tweepy to pull data from Twitter and scraping world news to predict daily sentiment for stocks using BERT.
hackathon-unigpt-assist
A RAG chatbot for UNIGE developed during the University of Geneva 2023 Hackathon.
iarts
A repo for my IARTS projects (AI generated art, style transfer, stable diffusion, etc.).
LLM-training
Experimenting with LLMs and training.
NLP-progress
Repository to track the progress in Natural Language Processing (NLP), including the datasets and the current state-of-the-art for the most common NLP tasks.
R_Programming_For_Data_Science
Summer School Course 2019
RAG-ML-chatbot
A repository to experiment with a specialized AI Assistant chatbot for ML Research
speech-emotion-recognition
A project to detect emotions from audio samples (Emo-DB: http://emodb.bilderbar.info/start.html).
Swiss_Mountain_app
Development of an app for the identification of swiss mountains. Wondering what peak you are facing while hiking in the swiss Alps? Take a picture with your cellphone and find out!
TransferLearning_ResNet_PETS
Transfer Learning with some pre-trained FastAI resnet models on the PETS image dataset. This is a fine grained classification task with 37 categories (cat and dog breeds). ResNet 34 achieves 94.3% accuracy and ResNet 50 achieves 95% accuracy.
K-Schubert's Repositories
K-Schubert/RAG-ML-chatbot
A repository to experiment with a specialized AI Assistant chatbot for ML Research
K-Schubert/BERT_Financial_Sentiment_Analysis
Using Tweepy to pull data from Twitter and scraping world news to predict daily sentiment for stocks using BERT.
K-Schubert/speech-emotion-recognition
A project to detect emotions from audio samples (Emo-DB: http://emodb.bilderbar.info/start.html).
K-Schubert/TransferLearning_ResNet_PETS
Transfer Learning with some pre-trained FastAI resnet models on the PETS image dataset. This is a fine grained classification task with 37 categories (cat and dog breeds). ResNet 34 achieves 94.3% accuracy and ResNet 50 achieves 95% accuracy.
K-Schubert/LLM-training
Experimenting with LLMs and training.
K-Schubert/NLP-progress
Repository to track the progress in Natural Language Processing (NLP), including the datasets and the current state-of-the-art for the most common NLP tasks.
K-Schubert/YOLOv5_object_detection
Testing out YOLO (v5) object detection model on 1. Cars in my street 2. Professional cycling races to detect race leaders and specific jerseys.
K-Schubert/BERT_Google_Play
Using BERT to do sentiment analysis on Google Play reviews.
K-Schubert/crypto-sentiment
A project to perform sentiment analysis on crypto news articles.
K-Schubert/hackathon-unigpt-assist
A RAG chatbot for UNIGE developed during the University of Geneva 2023 Hackathon.
K-Schubert/iarts
A repo for my IARTS projects (AI generated art, style transfer, stable diffusion, etc.).
K-Schubert/Swiss_Mountain_app
Development of an app for the identification of swiss mountains. Wondering what peak you are facing while hiking in the swiss Alps? Take a picture with your cellphone and find out!
K-Schubert/Brazilian_House_Rent_Linear_Regression
Linear Regression model on the Brazilian House Rent dataset trained with mini-batch SGD.
K-Schubert/CIFAR10_CNN
CNN model on the CIFAR10 image dataset. Used the Adam optimizer. Model trained for 20 epochs achieves 76% test accuracy. No data transformations.
K-Schubert/CIFAR10_Logistic_Regression
Multinomial Logistic Regression model trained with mini-batch SGD on the CIFAR-10 image dataset.
K-Schubert/CIFAR10_NN
Neural Network model on the CIFAR-10 image dataset trained with mini-batch SGD.
K-Schubert/CIFAR10_WideResNet22
Wide ResNet 22 model trained on the CIFAR10 image dataset.
K-Schubert/CIFAR10_WideResNet9
ResNet model on the CIFAR10 image dataset.
K-Schubert/CUDA_test
K-Schubert/Deep-Learning-Papers-Reading-Roadmap
Deep Learning papers reading roadmap for anyone who are eager to learn this amazing tech!
K-Schubert/K-Schubert
K-Schubert/MNIST_GAN
GAN model on the MNIST image dataset.
K-Schubert/MNIST_Logistic_Regression
Multinomial Logistic Regression on the MNIST image dataset trained with mini-batch SGD.
K-Schubert/MNIST_NN
Neural Network model on the MNIST image dataset trained with mini-batch SGD.
K-Schubert/pcs-performance-analytics
This project aims to: 1. Create a dashboard to visualize professional cyclist race data 2. Identify upcoming riders and talent through data analysis 3. Analyze team performance 4. Predict winners in professional cycling (World Tour) 5. Create a racing calendar optimization tool. The dataset was constructed by scraping the 'www.procyclingstats.com' website.
K-Schubert/Python
Python projects from the MSc. Statistics
K-Schubert/strava-analytics
A project for cycling training data analysis and optimal workout prediction.
K-Schubert/strava2TP
Scraper tool to upload all Strava activities to Training Peaks.
K-Schubert/swiss-ultracycling-challenge
This project aims to solve the problem of optimal route creation for the Swiss Ultracycling Challenge (SUCH). Given the mandatory checkpoints, the algorithm will find the shortest path to visit all Swiss cantons and checkpoints using Google Maps data, minimizing elevation and distance. This problem is an instance of the AGTSP (Asymmetric Generalized Traveling Salesman Problem) where a set of points are partitioned into clusters (cantons), and the cyclist has to visit every cluster exactly once.
K-Schubert/VAM_Analysis
Quick linear model for cycling based on my powermeter data and VAM values to predict the effort needed on a certain climb to achieve given time.