†§Z. Carmichael, H. F. Langroudi, C. Khazanov, J. Lillie, J. L. Gustafson, D. Kudithipudi, "Deep Positron: A Deep Neural Network Using Posit Number System," Proceedings of the IEEE Conference and Exhibition on Design, Automation and Test in Europe (DATE), Florence, Italy, 2019 [24% acceptance rate].
†§Z. Carmichael, H. F. Langroudi, C. Khazanov, J. Lillie, J. L. Gustafson, D. Kudithipudi, "Performance-Efficiency Trade-off of Low-Precision Numerical Formats in Deep Neural Networks," Proceedings of the ACM Conference for Next Generation Arithmetic (CoNGA), Singapore, 2019.
†§Z. Carmichael, B. Glasstone, F. Cwitkowitz, K. Alexopoulos, R. Relyea, R. Ptucha, "Autonomous Navigation Using Localization Priors, Sensor Fusion, and Terrain Classification," Proceedings of IS&T International Symposium on Electronic Imaging: Image Processing Algorithms and Systems, San Francisco, CA, 2019.
†§Z. Carmichael, H. Syed, D. Kudithipudi, "Analysis of Wide and Deep Echo State Networks for Multiscale Spatiotemporal Time Series Forecasting," Proceedings of the ACM Neuro Inspired Computational Elements (NICE) Workshop, Albany, NY, 2019.
‡Z. Carmichael, H. Syed, S. Burtner, D. Kudithipudi, "Mod-DeepESN: Modular Deep Echo State Network," Annual Conference on Cognitive Computational Neuroscience, Philadelphia, PA, 2018.
† Proceedings Paper | § Oral Presentation | ‡ Poster Presentation
I have been a member of Prof. Kudithipudi’s Neuromorphic Artificial Intelligence (Nu.AI) Lab since January of 2018, whose strong focus on novel and explainable learning techniques and advocation of a student-centric research environment attracted me. In the Nu.AI Lab, I modernized the Nu.AI website to better broadcast our research and attract interested researchers, set up and contributed content to a Confluence wiki space, communicated an internalized GitLab workflow to members to share codebases, and revived a weekly reading group to discuss novel and pertinent research papers. I provide members with programming lessons and have also mentored several undergraduate research assistants. In my involvement with the lab, I have published several papers on echo state networks (ESNs), the posit number system applied to deep learning, and am currently completing my Master's thesis that aims at exploiting knowledge distillation to support model interpretability.
As a digital engineering intern, I worked on several projects: some for Plexus customers, and others for internal usage. I designed RTL code for the generic instantiation of power-hungry components for Xilinx FPGAs. The module uses a given proportion of a chip to generate as much heat as possible when enabled to test the efficacy of its cooling systems. I worked with the digital engineering team at Plexus in parametrizing the design and ensuring weekly project goals were met. I also worked on an internal battery testing unit using the development board FRDM-K64F. I collaborated with mechanical, electrical, and software engineers to verify and validate the design.
At CUBRC, I joined the recently spawned health care team who were striving to push for a breakthrough in data analysis in the health care industry. The project, called Heartwood Analytics, has a vision to provide a broad genre of agencies with the tools to reduce the perplexity of the possession and potential utilization of huge amounts of data; through automated predictive analysis and accessible centralization of multi-source data, Heartwood Analytics provides its customers with the ability to build on their core competencies, identify market opportunities, and manage outcomes for their highest risk patients.
I contributed to the analytical backend of the proposed product by developing an automated machine learning framework that produces a fine-tuned model for any given health record data. Data is abstracted through a common ontology to allow for the use of various data sources, and then a model search is executed given a budget to produce a predictive model. Analysis is also run to interpret important factors, such as the higher weighted features of a regression. Additionally, I documented relevant research and the software itself, and provided receptive, effective support for our customers.
At the Information Services and Technology (ITS) Service Desk (SD), I was a senior representative and a Right Answers (RA) publisher and editor. As a SD senior, in addition to routine duties, I was responsible for the training of new employees, presentation of processes and support tips to student employees, and assessment of sensitive scenarios when full-time employees were unavailable. As a RA publisher, I was responsible for writing, editing, and publishing articles for the RIT community: faculty, staff, and students across multiple campuses.
CPUtiful Tech was a company comprising four partners with the mutual interest of creating a consumer electronics retail company with a business platform based on personable interactions. The company was dissolved in 2016.