University of Maryland

Pomona College

Research Assistant at University of Maryland

Washington, District Of Columbia, United States

- Designed and wrote (~90%) implementation of the Metacognitive Integrated Dual-­Cycle           Architecture (MIDCA)• MIDCA is a flexible, modular architecture for autonomous or semi-­‐autonomous agents.• MIDCA is currently under development and incorporates several tools describedbelow. [<a href="https://github.com/mclumd/MIDCA]%E2%80%A2">https://github.com/mclumd/MIDCA]•</a> Developed tools to generate problem sets, run tests, and statistically analyze results for experiments in anomaly detection• Created a method for classifying detected anomalies using unsupervised machine learning techniques• Combined above results with research on autonomous goal generation to confront an unexpected challenge (fire) in a construction domain• Separately, adapted The Metacognitive Loop (MCL), which provides anomaly-­‐linked metacognitive oversight in robotic domains, to a story understanding domain.• Provided technical support for adaptation of MCL to a dialogue agent

Research Assistant

TA for CMSC421, Intro to Artificial Intelligence

Teaching Assistant

- Worked on award winning Contrast application security tool-Developed a tool to search through large code bases for security controls and integrated it into Contrast; Presented at AppSec USA 2012-Contributed to Contrast test suite-Began development on rule set to automatically customize firewalls based ondetected security risks; intended for later Contrast integration

Matthew Paisner