Description

---

A group of researcher from the Swiss Center of Affective Science investigated the neuronal networks underlying fundamental processes of taste perception.

I processed all the MRI data of 97 individuals to analyze and summarize the results in support of the paper intitled 3D-Printed Mouthpiece for fMRI-Compatible Gustometers.

My work allowed to visualize the parameter distribution of the BOLD signal in 3D.

Read more

---

You can read the scientific paper on eNeuro. The source code of the project are of course on Github.

Description

---

Building and comparing brain maps between several models is a common task for neuroscientist.

Here we performed a Bayesian model selection (BMS) to select the best model given the data for group-level analysis. This procedure computes the probability of the data given the model (model evidence) for a set of candidate models. It also allows to compare characterize maps"heat zones" and to visualize them.

Read more

---

You can read the scientific paper on BiorXiv. The source code of the project are on Github.

Description

---

I collaborated on the Hierarchical Bayesian modeling of Decision-Making tasks library hBayesDM. hBayesDM is a user-friendly package that offers hierarchical Bayesian analysis of various computational models on an array of decision-making tasks. It uses STAN for Bayesian inference and supports both R and Python. hBayesDM's goal is to help interpreting experimental data through computational models using a full Bayesian statistical inference approach with MCMC sampling.

My work involved the implementation of the Q-learning algorithm for reinforcement learning tasks.

Read more

---

You can check my project contributions on Github.

Gaming Reinforcement Learning with PyTorch (Super Mario Bros)

Here I used OpenAI gym tools to build a Super Mario Bros environment, and trained a Proximal Policy Optimization (PPO) algorithm with actor/critic networks being convolutional neural networks (CNNs) with PyTorch to play the game with reduced actions. I also used TensorBoard to visualize and evaluate the quality metrics of the model.

You can check the project Github.

---

Multiplayer AI programming online game

Leek Wars is a multiplayer AI programming game in which you train leeks to fight each others. The language of Leek Wars is LeekScript, built in Java. Follow my leek @leakit22 ! I remain in the top 100 for the time being.

You can check the AI on Github.

---

Conway's Game Of Life implemented in Python

The “game” is a zero-player game, meaning that its evolution is determined by its initial state, requiring no further input. One interacts with the Game of Life by creating an initial configuration and observing how it evolves, or, for advanced “players”, by creating patterns with particular properties. How the game works Because the Game of Life is built on a grid of nine squares, every cell has eight neighboring cells,as shown in the given figure.

A given cell (i, j) in the simulation is accessed on a grid [i][j], where i and j are the row and column indices, respectively. The value of a given cell at a given instant of time depends on the state of its neighbors at the previous time step.

Conway's Game of Life has four rules:

1) If a cell is ON and has fewer than two neighbors that are ON, it turns OFF.
2) If a cell is ON and has either two or three neighbors that are ON, it remains ON.
3) If a cell is ON and has more than three neighbors that are ON, it turns OFF.
4) If a cell is OFF and has exactly three neighbors that are ON, it turns ON.


The bulk of the graphical implementation was made using PyGame.

You can check the project on Github.

---

Description

---

I have been involved in We Data since 2020. We Data is a Swiss organization that shares knowledge on code in social sciences, gives statistical courses, and do coding demonstrations.

Our goal is to promote code in the social science and raise awareness of new digital technologies.

You can check the project on our Blog, Github, or on our YouTube channel.

Description

---

The goal for this project is to provide context for the image quality metrics (IQMs) shown in the MRI quality metric group reports, by showing the distribution of IQMs for your data plotted relative to a larger set of anonymized IQMs pulled from the crowdsourcing web API.

The code provided creates an online interactive data visualization app using Dash and Plotly



You can check the project on Github.