My research interests lie in the broad area of multimedia coding and communications. My primary focus is on the analysis and design of constructive approaches for reliable data transmission over modern communications networks. More specific topics of my research are network aware data compression, signal quantization, image coding and transmission, steganalysis. My earlier research interests were in formal languages and automata theory.
Sorina Dumitrescu received the B.Sc. and Ph.D. degrees in mathematics from the University of Bucharest, Romania, in 1990 and 1997, respectively. From 2000 to 2002 she was a Postdoctoral Fellow in the Department of Computer Science at the University of Western Ontario, London, Canada. Since 2002 she has been with the Department of Electrical and Computer Engineering at McMaster University, Hamilton, Canada, where she held Postdoctoral, Research Associate, and Assistant Professor positions, and where she is currently an Associate Professor. Her current research interests include multimedia coding and communications, network-aware data compression, signal quantization, steganalysis. Her earlier research interests were in formal languages and automata theory. Dr. Dumitrescu was a recipient of the NSERC University Faculty Award during 2007-2012.
B.Sc. (University of Bucharest, Romania) ; Ph.D. (University of Bucharest, Romania)
NSERC University Faculty Award (2007 – 2012)
LL (PEO).
Undergraduate Advisor for the Computer Engineering program (ECE Department, McMaster Universit
Dumitrescu, S., Wu, X., and Wang, Z. (2003) Detection of LSB Steganography via Sample Pair Analysis IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 51, ISSUE 7, PP. 1995 – 2007
Shao, M., Dumitrescu, S., and Wu, X. (2011) Layered Multicast with Inter-layer Network Coding for Multimedia Streaming IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 13, NO. 2, PP. 353 – 365
Dumitrescu, S. and Wu, X. (2009) On Properties of Locally Optimal Multiple Description Scalar Quantizers with Convex Cells IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 55, NO. 12, PP. 5591 – 5606
Dumitrescu, S. (2016) On the Design of Optimal Noisy Channel Scalar Quantizer with Random Index Assignment IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 55, NO. 12, PP. 5591 – 5606
Recent 1) D. Elzouki, S. Dumitrescu, and J. Chen, “Lattice-based Robust Distributed Source Coding’’, IEEE Transactions on Information Theory, vol. 65, no. 3, pp. 1764-1781, Mar. 2019, DOI: 10.1109/TIT.2018.2878219.
2) H. Wu and S. Dumitrescu, „Design of Optimal Entropy-constrained Unrestricted Polar Quantizer for Bivariate Circularly Symmetric Sources”, to appear in IEEE Transactions on Communications, vol. 66, no.5, pp. 2169-2180, May 2018, DOI: 10.1109/TCOMM.2017.2789221.
3) Wu, H. , Zheng, T., and Dumitrescu, S., „On the Design of Symmetric Entropy-constrained Multiple Description Scalar Quantizer with Linear Joint Decoders”, IEEE Transactions on Communications, vol. 65, no. 8, pp. 3453—3466, Aug. 2017, DOI: 10.1109/TCOMM.2017.2704585.
4) Lin, K., and Dumitrescu, S., „Cross-layer Resource Allocation for Scalable Video over OFDMA Wireless Networks: Trade-off between Quality Fairness and Efficiency”, IEEE Transactions on Multimedia, vol. 19, no. 7, pp. 1654—1669, July 2017, DOI: 10.1109/TMM.2017.2678198..
5) Dumitrescu, S. and Wan, Y., “Bit-error Resilient Index Assignment for Multiple Description Scalar Quantizers”, IEEE Transactions on Information Theory, vol. 61, no. 5, pp. 2748—2763, May 2015.
5) Gao, Z. and Dumitrescu, S., “Flexible Symmetric Multiple Description Lattice Vector Quantizer with L ≥ 3 Descriptions”, IEEE Transactions Communications, vol. 62, no. 12, pp. 4281 – 4292, Dec. 2014.
3 unit(s) The course objective is to provide a broad introduction to common machine learning approaches and their underlying principles. The intended topics are: linear methods for regression and classification, nearest neighbours, decision trees, bias-variance trade-off, neural networks (including deep neural nets, convolutional nets, generative adversarial nets and variational autoencoders), clustering, principal component analysis, Gaussian mixture models and the EM algorithm, basics of reinforcement learning including temporal difference learning (Sarsa, Q-learning, expected SARSA, SARSA with function approximation) and policy gradient methods.
Instructor
Dr. Sorina Dumitrescu
Fundamental principles and algorithms of machine learning: linear and logistic regression, nearest neighbours, decision trees, neural networks, support vector machines, ensemble methods; clustering and principal component analysis; basics of reinforcement learning; Three lectures, one tutorial, first term Prerequisite(s): COMPENG 2SI4 or 2SI3, STATS 3Y03 or HTHSCI 2G03, and ELECENG 3TQ3 or ELECENG 3TQ4 Antirequisite(s): CHEMENG 4H03, COMPSCI 4ML3, STATS 3DS3
Fundamental principles and algorithms of machine learning: linear and logistic regression, nearest neighbours, decision trees, neural networks, support vector machines, ensemble methods; clustering and principal component analysis; basics of reinforcement learning; Three lectures, one tutorial, first term Prerequisite(s): COMPENG 2SI4 or 2SI3, STATS 3Y03 or HTHSCI 2G03, and ELECENG 3TQ3 or ELECENG 3TQ4 Antirequisite(s): CHEMENG 4H03, COMPSCI 4ML3, STATS 3DS3
Design and analysis of correct and efficient algorithms and related discrete mathematics concepts and data structures. Topics include: sets, function relations; graph theory; graph algorithms (graph traversals, topological sort, minimum spanning trees, shortest paths); balanced trees and advanced data structures; algorithmic design strategies (dynamic programming, greedy algorithms, divide-and-conquer, backtracking); introduction to NP completeness and approximation algorithms; introduction to parallel algorithms. Three lectures, one tutorial, one lab every other week; second term Prerequisite(s): COMPENG 2SH4, and COMPENG 2SI4 or COMPENG 2SI3 Antirequisite(s): COMPSCI 2C03
