About Stacy Fenstermacher

Poster Keerthana Deepti Karunakaran BioMedical Engineering And Imaging Institute


Keerthina De Ponte


PhD Candidate, Computer Science – University of Cambridge




Keerthina is pursuing her doctorate in artificial‑intelligence ethics at the Department of Computer Science and Technology. Her work bridges formal verification methods with practical policy design, aiming to make machine‑learning systems both trustworthy and transparent for real‑world deployment.



---




Keerthina De Ponte


PhD Student, Electrical Engineering – University of Michigan




At U‑M’s School of Electrical & Computer Engineering, Keerthina studies the integration of neuromorphic hardware with spiking‑neural‑network algorithms. Her research seeks to reduce energy consumption for edge‑AI applications while preserving performance on vision and speech tasks.



---




Keerthina De Ponte


PhD Candidate, Computer Science – Stanford University




In Stanford’s AI Lab, Keerthina focuses on causal inference in deep learning models. She develops techniques that allow neural networks to generate counterfactual explanations, improving transparency for medical diagnosis and autonomous driving systems.



---




Keir De Pante


Research Interests




Causal Inference – Developing methods for discovering cause‑effect relationships from observational data.


Generative Models – Building scalable architectures for image and text synthesis.


Statistical Learning Theory – Theoretical analysis of learning algorithms under distributional shift.







Keir De Pante


Current Projects



Project Brief Description


CausalGAN A generative adversarial network that incorporates causal constraints to produce realistic counterfactuals.


ShiftRobustness Framework for quantifying and mitigating performance degradation when test data diverges from training distribution.


Explainable AI Toolkit Open‑source library providing transparent explanations for deep learning models across modalities.


---




Keir De Pante


Publications




CausalGAN: Learning Causally Structured Generative Models, JMLR, 2023.


ShiftRobustness in Deep Neural Networks, ICML, 2024.


Explainable AI Toolkit for Cross‑Modal Models, arXiv preprint, 2025.







Keir De Pante


Patents




US Patent 10,123,456 – "Method and Apparatus for Causally Guided Generative Modeling."


WO Patent 2021/012345 – "Framework for Cross‑Modal Explainability."







Keir De Pante


Professional Affiliations




IEEE Fellow, AI & Machine Learning Society.


Member, Association for Computing Machinery (ACM).







Keir De Pante


Contact Information




Email: keir.depante@ai-research.org


Phone: +1 555‑123‑4567


LinkedIn: linkedin.com/in/keirdepante


ResearchGate: researchgate.net/profile/KeirDepante






End of CV.



---
Note: This document is a stylized representation for illustrative purposes.

Female

---