# MVA-2021
My solutions of the homeworks of several courses of the Masters [MVA](<https://www.master-mva.com/>) (Mathématiques, Vision, Apprentissage) 


## First Semester

### [3D Computer Vision](<https://github.com/pierrefdz/MVA-2021/tree/main/3D_computer_vision>)

Instructors: Pascal Monasse and Mathieu Aubry.

Includes:
- [Panorama with homography computation and perspective projection](<https://github.com/pierrefdz/MVA-2021/tree/main/3D_computer_vision/Panorama>)
- [Fundamental matrix computation with SIFT keypoints](<https://github.com/pierrefdz/MVA-2021/tree/main/3D_computer_vision/Fundamental>)
- [Disparity map computation by propagation of seeds](<https://github.com/pierrefdz/MVA-2021/tree/main/3D_computer_vision/Seeds>)
- [Disparity map computation by Graph-Cuts](<https://github.com/pierrefdz/MVA-2021/tree/main/3D_computer_vision/Gcdispar>)


### [Object Recognition](<https://github.com/pierrefdz/MVA-2021/tree/main/object_recognition>)

Instructors: Jean Ponce, Ivan Laptev, Cordelia Schmid and Josef Sivic

Includes:
- [Instance Level recognition](<https://github.com/pierrefdz/MVA-2021/blob/main/object_recognition/hw1_instance_level_recognition.ipynb>)
- [Neural Networks](<https://github.com/pierrefdz/MVA-2021/blob/main/object_recognition/hw2_neural_nets.ipynb>)
- Bird Classification Data-Challenge: [Code](<https://github.com/pierrefdz/MVA-2021/blob/main/object_recognition/hw3_bird_classification.ipynb>) & [Report](<https://github.com/pierrefdz/MVA-2021/blob/main/object_recognition/hw3_bird_classification.pdf>) 


### [Computational Statistics](<https://github.com/pierrefdz/MVA-2021/tree/main/computational_stats>)
Instructor: Stéphanie Allassonnière

Includes:
- [Simple implementation of SGD](<https://github.com/pierrefdz/MVA-2021/blob/main/computational_stats/hw1_sol.ipynb>)
- [Implementation of Gaussian-Mixtures Model with EM algorithm and Adaptative Importance Sampling with Monte-Carlo method](<https://github.com/pierrefdz/MVA-2021/blob/main/computational_stats/hw2_sol.ipynb>)
- [Implementation of Hastings-Metropolis Sampler and some extensions (Adaptative HM sampler within Gibbs and Parallel Tempering)](<https://github.com/pierrefdz/MVA-2021/blob/main/computational_stats/hw4_sol.ipynb>)
- [Implementation of Tempering SAEM from *A New Class of EM Algorithms. Escaping Local Minima and Handling Intractable Sampling*](<https://github.com/pierrefdz/MVA-2021/blob/main/computational_stats/temp_saem.ipynb>)


### [Convex Optimization](<https://github.com/pierrefdz/MVA-2021/tree/main/convex_optim>)
Instructor: Alexandre d'Aspremont

Includes:
- [Convexity, Conjugate Function](<https://github.com/pierrefdz/MVA-2021/blob/main/convex_optim/hw1_sol.pdf>)
- [Duality](<https://github.com/pierrefdz/MVA-2021/blob/main/convex_optim/hw2_sol.pdf>)
- [LASSO Regularization](<https://github.com/pierrefdz/MVA-2021/blob/main/convex_optim/hw3_sol.ipynb>)
- [Final Exam](<https://github.com/pierrefdz/MVA-2021/blob/main/convex_optim/exam_sol.pdf>)


### [Reinforcement Learning](<https://github.com/pierrefdz/MVA-2021/tree/main/reinforcement_learning>)
Instructors: Alessandro Lazaric and Matteo Pirotta

Includes:
- [Dynamic Programming](<https://github.com/pierrefdz/MVA-2021/tree/main/reinforcement_learning/hw1_dynamic_programming>) (MDP, Policy Iteration and Value Iteration)
- [Approximate RL](<https://github.com/pierrefdz/MVA-2021/tree/main/reinforcement_learning/hw2_approximate_rl>) (TRPO, TD, REINFORCE, DQN)
- [Exploration](<https://github.com/pierrefdz/MVA-2021/tree/main/reinforcement_learning/hw3_exploration>) (UCB, Best Arm Identification, Bernoulli Bandits)
- [Model Selection in Contextual Bandits](<https://github.com/pierrefdz/MVA-2021/tree/main/reinforcement_learning/hw4_model_selection_bandits>)



## Second Semester

### [Point Clouds & 3D Modelling (NPM3D)](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d>)
Instructors: François Goulette, Jean-Emmanuel Deschaud and Tamy Boubekeur

Includes:
- [Basic Transformations of 3D Point Clouds](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d/hw1_basics>)
- [Iterative Closest Point](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d/hw2_icp>)
- [3D Descriptors](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d/hw3_descriptors>)
- [Surface Reconstruction](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d/hw4_reconstruction>)
- [Modeling and Shape Detection](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d/hw5_modelisation>)
- [Deep Learning for Point CLouds](<https://github.com/pierrefdz/MVA-2021/tree/main/npm3d/hw6_deep>)


### [Deep Learning in Practice](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice>)
Instructor: Guillaume Charpiat

Includes:
- [Hyperparameter and training basics](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice/hw1_basics>)
- [Visualization with grad-CAM](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice/hw2_gradcam>)
- [Graph Neural Network](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice/hw3_gnn>)
- [Transfer Learning](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice/hw4_transfer>)
- [Learning Dynamical Systems](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice/hw5_physics>)
- [Generative Models](<https://github.com/pierrefdz/MVA-2021/tree/main/dl_in_practice/hw6_generative>)


### [Graphs in Machine Learning](<https://github.com/pierrefdz/MVA-2021/tree/main/graphs_in_ml>)
Instructors: Daniele Calandriello and Michal Valko

Includes:
- [Spectral Clustering](<https://github.com/pierrefdz/MVA-2021/tree/main/graphs_in_ml/hw1_spectral_clustering>)
- [SSL](<https://github.com/pierrefdz/MVA-2021/tree/main/graphs_in_ml/hw2_ssl>)
- [Spectral GNN](<https://github.com/pierrefdz/MVA-2021/tree/main/graphs_in_ml/hw3_spectral_gnn.ipynb>)

### [Kernel Methods](<https://github.com/pierrefdz/MVA-2021/tree/main/kernel_methods>)
Instructors: Julien Mairal and Jean-Philippe Vert

Includes:
- [Positive definite kernels and boundedness](<https://github.com/pierrefdz/MVA-2021/blob/main/kernel_methods/hw1_sol.pdf>)
- [RKHS](<https://github.com/pierrefdz/MVA-2021/blob/main/kernel_methods/hw2_sol.pdf>)
- [COCO: Covariance Kernel](<https://github.com/pierrefdz/MVA-2021/blob/main/kernel_methods/hw3_sol.pdf>)
- [Bn-Splines and diffusion kernel](<https://github.com/pierrefdz/MVA-2021/blob/main/kernel_methods/hw4_sol.pdf>)
- [Final exam](<https://github.com/pierrefdz/MVA-2021/blob/main/kernel_methods/hw4_sol.pdf>) (Positive Definiteness, Equivalence classes, Kernel mean embedding, Dot-product kernel)

### [Sequential Learning](<https://github.com/pierrefdz/MVA-2021/tree/main/sequential_learning>)

Instructors: Pierre Gaillard and Rémi Degenne

Includes:
- [Rock Paper Scissors with bandits and Bernoulli bandits](<https://github.com/pierrefdz/MVA-2021/blob/main/sequential_learning/hw_sol.pdf>)