rmko4 Goto Github PK

analytical-displacement-mapping

Implementation of analytic displacement mapping using hardware tessellation based on Nießner & Loop (2013). The model is extended with procedural generation of displacement textures.

asteroid-shooter

backprop-mlp

Explicit backpropagation (no AD) of a multilayer perceptron in a single binary classification task in TensorFlow. Includes validation with automatic differentiation (AD).

hydrodynamic-imaging

Hydrodynamic image with the artificial lateral line using physics-informed informed neural networks and other proven methods in 2D dipole localization.

ik-nlp-tutorials

Lab tutorials for the MSc NLP course at the University of Groningen 🐮

mas-trust-in-exchange

Multi-agent system for modeling trust in market exchanges.

ml-covid-forecasting

C implementation of a Multilayer perceptron with gradient descent through explicit backpropagation for use in time-series prediction tasks. The project targets forecasting of Covid-19 data.

ml-unfinished-fugue

Music generation with machine learning: Completing Bach's unfinished Contrapunctus XIV. A variety of models including linear regression, multilayer perceptrons and echo state networks are used as autoregressive models to take on the challenging time-series task.

nlp-explanations-for-nli

Fine-tuning and evaluation of a language model for explanation generation of natural language inference. Fine-tuning scripts for a pre-trained T5 model supporting both full model fine-tuning as well as LoRA are included in this repository.

rl-catch-value-based

Deep Reinforcement Learning: Value-Based methods. An implementation of DQN, DDQN, Dueling Architectures, DQV, DQV-Max on the PyTorch Lightning framework.

rl-karmed-bandit

K-armed bandit problem approached with a variety of action-selection learning algorithms.

rl-on-policy-actor-critic

Deep Reinforcement Learning: On-Policy Actor Critic methods. An implementation of Advantage Actor-Critic (A2C) and Proximal Policy Optimization (PPO) on the PyTorch Lightning framework.

rl-tabular-rubikscube

Reinforcement Learning with tabular methods: TD-learning (Q-learning and SARSA) and MENACE-like approach applied to a Rubik's cube with a move set restricted to 180-degree turns.

rmko4

Config files for my GitHub profile.