Derek Prijatelj's Projects
code of project that corresponds with the presentation "Towards Cross-Language Automated Authorship Attribution” at the 2018 Keystone DH Conference at Penn State University..
Code for the paper "A Bayesian Evaluation Framework for Ground Truth-Free Visual Recognition Tasks".
Probablistic models for certainty.
Simulations to evaluate probablistic models of certainty.
Tensorflow backend for certainty: Probablistic models of certainty including those that are Bayesian or information theoretic.
Torch backend for certainty: Probablistic models of certainty including those that are Bayesian or information theoretic.
This is a practice compiler for a very simpified version of c written in c++. This project was done as a course project for the class "System Programming & Design" my sophomore year. The code is not optimized to the fullest extent and is not as efficient as possible, for it was not necessarially meant for that. The main objective was to recreate a compiler for a simplified code language to better understand the compiling process and the intricacies involved in basic compiling of code.
Notre Dame scripts and notes for ease of computation on the CRC machines.
A Computational Approach to Crowd Noise Generation, as presented at the 2018 Keystone Digital Humanities Conference at Penn State University.
Computer Security undergraduate coursework from Fall 2017, Duquesne University
A TensorFlow implementation of the Differentiable Neural Computer.
Generate a hierarchical ConfigArgParser from docstrings.
Dot files for Arch Linux custom configuration and ricing.
Commonly used experiment utilities
Various graphics programs from both academics and in personal time.
The code for the ICDAR 2021 paper, "Handwriting Recognition with Novelty" by Derek S. Prijatelj, Samuel Grieggs, Futoshi Yumoto, Eric Robertson, and Walter J. Scheirer.
Code for the paper "Human Activity Recognition in the Open World."
The Java Graphical Authorship Attribution Program
Deep Learning for humans
Memory Networks implementations
Natural Language Processing Duquesne Fall 2017 Course
JGAAP log Parser and pairwise independence calculator for different JGAAP methods on same test set
Duquesne Senior Project: Persona Simulation in Chatbots. PerSim v0.1 (proof of concept)
Problem rendering engine for WeBWorK
python3 bindings for FLANN - Fast Library for Approximate Nearest Neighbors. The result of python3 version of pyflann when run thru 2to3.
Structure, sample, and savor hyperparameter searches.