Reinforcement Learning for a maze-like GridWorld

This project uses reinforcement learning techniques to teach a computer agent how to navigate a maze-like environment in order to maximize a reward signal by evading bombs that are fixed in location. The agent ultimately learns the optimal path to cell within the grid that is most rewarding; a chest. I use an on-policy first-visit Monte Carlo and a Q-Learning approach. 

Supervised Learning for Binary Classification

In this mock-up paper, I use three supervised-learning algorithms to conduct a binary classification on three datasets acquired from the UCI machine learning repository. I compare the accuracy of a logistic regression, k-nearest neighbors, and random forest classifier. 

Cognitive Ethnography | Hands-on

In this short post, I take some time to elaborate on my experience working as a research assistant at the San Diego Zoo. Routine visits for data collection on primates would allow me to deepen my own understanding of observational research. Abstractions pertaining to cognitive phenomena involving apes are supplementary to the understanding of human sociality. 

Fish Market Analysis

As a team member, I helped contribute to an assignment where multivariate models were used to predict the weight of a fish given relevant parameters.

This was done using an existing dataset, and models were compared using a cross-validation technique to assess the degree of error regarding the model's predictions. 

Bilingualism & Executive Function

A mock-up research proposal that seeks to evaluate our current understanding of second language acquisition and its contribution to working memory, cognitive flexibility, and inhibition. 

Programming the Lego NXT

The construction and development of a mobile robot was part of a team project completed throughout the fall quarter of 2019.  The continuous reconstruction of its structure allowed for new functions to be programmed into the device. 

At varying points in time. this robot could draw, speak, avoid obstacles, and more.