Research
CogAI lab focuses on research & development in the areas of Computer Science and Cognitive Science involving the intersection and interplay between humans and software. Topics of interest include:
Artificial Intelligence
Cognition-Inspired Machine Learning
Artificial General Intelligence (AGI)
Human-AI teaming, collaboration, and symbiosis
AI usability, explainability (XAI), and trust
Human-Computer Interaction (HCI)
Improving UI/UX in everyday technology
AR / VR usability and effectiveness
Adaptive UI
Augmented Cognition
Simulation-based predictions of human behavior
Apps for improving cognitive/behavioral outcomes
Current Research Projects:
Artificial Intelligence
Addressing computational efficiency and catastrophic interference issues in neural networks
Brain-inspired approaches to Artificial Intelligence
Using Machine Learning to detect plagiarism
Integrating GPT with Reinforcement Learning
Real-time video to audio translator for the visually impaired
Human-Computer Interaction
Touch-based software development and keyboard-free programming
Machine-readable cross-framework GUI interaction language as an alternative to HTML/REST
Web components library development (as an alternative to monolithic code bases)
Low-resource fast data collection framework
Recent Publications and Presentations:
Gedenidze, N. & Veksler, V. D. (2024). A Layer-Freezing Approach for Reduced Backpropagation Demand. 28th annual conference of the Consortium for Computing Sciences in Colleges, Northeastern Region (CCSCNE). Albany, NY.
Salazar, J. & Veksler, V. D. (2024). Comparing Touch-Based Coding to Traditional Computer Programming Methods. 28th annual conference of the Consortium for Computing Sciences in Colleges, Northeastern Region (CCSCNE). Albany, NY.
Khatri, K., Khadka, S., & Veksler, V. D. (2024). ML for Plagiarism Detection in Coding Assignments. 28th annual conference of the Consortium for Computing Sciences in Colleges, Northeastern Region (CCSCNE). Albany, NY.
Gogitidze, G. & Veksler, V. D. (2024). Using a GPT with Reinforcement Learning for More Personalized or Targeted Text Generation. 28th annual conference of the Consortium for Computing Sciences in Colleges, Northeastern Region (CCSCNE). Albany, NY.
Veksler, V.D., Gedenidze, N., & Yadav, R. (2023). Visual Cortex Doesn't Change, Why should Convolutional Layers? The 16th International Conference on Brain Informatics. Hoboken, NJ, USA.
Gedenidze, N. & Veksler, V. D. (2023). Backpropagation is expensive. Is it necessary? 27th annual conference of the Consortium for Computing Sciences in Colleges, Northeastern Region (CCSCNE). Ithaca, NY.
Veksler, V. D., Hoffman, B. E., & Buchler, N. (2021). Symbolic Deep Networks: A psychologically-inspired light-weight and efficient approach to Deep Learning. Topics in Cognitive Science. https://doi.org/10.1111/tops.12571