# Alexander Barg

###### Contact
Office: AVW 2361
Phone: +1 301 405 7135
[abarg@umd.edu]

Department of ECE
University of Maryland
8223 Paint Branch Dr.
College Park, MD 20742, USA

###### Research
• Information storage and management in networks: Suppose that information is encoded with a code of length $n$ over a finite alphabet $F$. Coordinates of the codeword are stored at the vertices of a graph $G(V,E)$ with $|V|=n.$ This setting models distributed storage systems in which servers are represented by the vertices, and connections of the graph represent communication constraints in the system. The problems considered here include protocols for the recovery of erased data at the vertices, bounding the communication complexity of recovery, data recovery in the presence of corrupted nodes, dynamic data maintenance in random environment, and other similar questions.
[more]

• Discrepancy of codes and uniform distributions: Here one is interested in characterizing binary codes and codes in other finite metric spaces that approximate the uniform distribution on the space. Applications of such codes could include derandomizing algorithms, approximation theory, probability of decoding error, image processing, and concept learning (uniform laws of large numbers and VC dimension). These problems are also connected with constructing energy-minimizing configurations in metric spaces.
[more]

• Other problems: Algebraic constructions of codes for storage [1], [2]; Private distribution estimation

Continual support of the US National Science Foundation is gratefully acknowledged.