Publications
Peer-reviewed work on differentiation, privacy, and language systems.
Six publications spanning scalable automatic differentiation, privacy-preserving machine learning, LLM-based language systems, and speech processing, listed newest first.
Speaker-attributed meeting transcription refinement with constrained open-weight language models
Costin-Alexandru Deonise, T.M. Coconu, Muhammad Bajwa, Cătălin Negru, B.C. Mocanu, Aniello Castiglione, Florin Pop
Future Generation Computer Systems (Elsevier), June 2026
Homomorphic Encryption for Privacy-Preserving Machine Learning Inference
Costin-Alexandru Deonise, A.M. Roangheși, L.I. Gheorghe, E. Simion, B.C. Mocanu, D. Țurcanu, Florin Pop
Annals of the Academy of Romanian Scientists, Series on Science and Technology and Information, Vol. 18, No. 2, 2025
Parallel and Distributed Computation of High-Order Derivatives in Neural Networks using Stochastic Taylor Derivative Estimator
Costin-Alexandru Deonise, J. Kołodziej, Florin Pop
2025 24th RoEduNet Conference: Networking in Education and Research, 2025
- Best Scientific PaperMost Impactful Paper
Stochastic Taylor Derivative Estimator for High-Order Differential Operators
Costin-Alexandru Deonise, Florin Pop
UPB Students' Scientific Communications Session (SCSS 2025), 2025
Microphone Speaker Analysis: Audio Segmentation and Frequency Insights
T.M. Coconu, Costin-Alexandru Deonise, C. Anghel, C. Negru, Florin Pop
Annals of the Academy of Romanian Scientists, 2024
Improved Speech Activity Detection Model Using Convolutional Neural Networks
Costin-Alexandru Deonise, T.M. Coconu, T. Rebedea, Florin Pop
2023 22nd RoEduNet Conference: Networking in Education and Research, 2023