List of projects from the Machine Learning Diploma on OpenClassroom with Centrale Supelec.
Learning Strategy
Data exploration and Data Cleaning techniques.
Linear predictions models framework with Supervised Learning.
- Feature Engineering
- Train, Validation, Test datasets
- Fine-Tuning (Hyperparametrisation and Cross-Validation)
- Model Evaluation and Error Analysis
Data segmentation with Unsupervised Learning.
- Database engineering
- Dimensionality Reduction (LDA, PCA)
- Unsupervised Models for Clustering (KMeans, PCA, KNN, (H)DBSCAN)
NLP categorisation for tag prediction.
- Text feature engineering (word tokenization, word2vec)
- Transformers and BERT models
- Text2Vec
Image classification.
- Pytorch
- Image feature engineering (Image normalisation, data augmentation)
- CNN architectures and theory (Convolution)
- GPU Acceleration
- Transfer Learning for Large CNN models
Enhancing a project through a POC and Cloud production.
- AWS manipulation to deploy back and front of a live app
- Tensorflow and Deep Learning Supervised Learning
- ELMo and BERT models
Participating in a Kaggle Competition. NLP, classfication of twitter messages.
- Tf-IDF, ELMo and BERT and state of the art methods in text classification (fine-tuning)
- Error analysis and model improvements (F1 Score analysis)