CUNY Graduate Center RSS Feedshttp://www.gc.cuny.eduCopyright 2015 CUNY Graduate Center. All rights reserved.Fall 2015 Computer Science Colloquium<p>Event Date: 9/3/2015 4:15:58 PM</p>Every Thursday, distinguished researchers in Computer Science present their current research.http://www.gc.cuny.edu/News/Calendar/Detail?id=325802_309/03/20154:15 PM12/17/20151:14 PM365 Fifth AvenueThe Graduate Center9204FreeEvery Thursday, distinguished researchers in Computer Science present their research.32333 Felisa Vazquez-Abad: Understanding the influence of uncertainty on algorithms<p>Event Date: 9/3/2015 4:15:00 PM</p>Speaker: Felisa Vázquez-Abad (Hunter College)<br />
<br />
Title: Understanding the influence of uncertainty on algorithms: application of stochastic models for optimization;<br />
<br />
Abstract:<br />
This talk will present examples of applications of stochastic modeling and optimization. These examples are co-authored by students who took the "stochastic modeling and simulation” course and the “stochastic optimization” course at the graduate center over the past four years.<br />
<br />
We first talk about the problem of scheduling robotic bees back to the hive. The deterministic algorithm is shown to be optimal and run in O(N logN) time. However when noisy GPS information is considered, it is no longer possible to avoid collisions with probability one. We will emphasize how the problem should be re-formulated introducing a risk constraint. A dynamic algorithm can correct for the risk, but the CPU time may be too large, so an alternative method is to solve smaller “local” problems to guide the bees for a short time and then re-evaluate the positions.<br />
<br />
The second problem deals with correcting circadian cycles. Biochemical equations lead to deterministic models that describe the dynamics of protein concentration in terms of ordinary differential equations (ODE’s). However when dealing with a single cell model with only one or two (integer-valued) DNA molecules the ODE (fluid) description is not adequate. A stochastic model can be used to simulate the biological system. Our first result shows the exact relationship between the stochastic model and the ODE model. While accurate, simulations can be slow with the stochastic model, so we propose algorithms that are hybrid.<br />
<br />
The third project deals with an intelligent transportation model for one-way vehicles for hire. We present our ideas for incorporating intelligence and, again, show that because of the uncertainties, it is impossible to ensure that reservations will be kept. Thus a measure of risk must be considered when modeling the optimization problem. Our project includes modeling human social behavior and we incorporate humans as sensors in a complex network.<br />
<br />
If time permits, we will finalize with a problem in cyber-security. Cognitive Radio Networks (CRN) may be subject to jamming attacks. In our research we propose a statistical learning model to detect when channels are attacked and what is the type of the attacker. According to the type of attack a number of strategies are proposed implementing honeypot (decoy) strategies.<br />
<br />
We will present the main problems and methodologies broadly, without going too much into the details. The emphasis will be on understanding how uncertainty may dramatically change the focus of the algorithmic solutions.<br />
<br />
For Prof Felisa Vázquez-Abad's bio, you may use the website<br />
<a href="http://www.cs.hunter.cuny.edu/~felisav/CoSSMO/about/our-people.html" target="_blank">http://www.cs.hunter.cuny.edu/~felisav/CoSSMO/about/our-people.html</a>http://www.gc.cuny.edu/News/Calendar/Detail?id=325812_309/03/20154:15 PM09/03/20156:00 PM365 Fifth AvenueThe Graduate Center9204FreeSpeaker: Felisa Vazquez-Abad, Hunter College
Title: Understanding the influence of uncertainty on algorithms: application of stochastic models for optimization;32334Rohit Parikh, Brooklyn Coolege Title: The logic and strategy of political campaigning<p>Event Date: 9/24/2015 4:15:32 PM</p><p>
Rohit Parikh, Brooklyn Coolege<br />
Title: The logic and strategy of political campaigning</p>
http://www.gc.cuny.edu/News/Calendar/Detail?id=326212_309/24/20154:15 PM09/24/20156:00 PM365 Fifth AvenueThe Graduate Center9204FreeRohit Parikh, Brooklyn Coolege
Title: The logic and strategy of political campaigning32374Noson Yanofsky: Kolmogorov Complexity of Categorical Structures<p>Event Date: 9/17/2015 4:15:48 PM</p>Speaker: Noson Yanofsky, Brooklyn College<br />
<br />
Title: Kolmogorov Complexity of Categorical Structures<br />
<br />
Abstract: Kolmogorov complexity (KC) is a measure of the informational content of a string. For a given string, the KC is the size of the shortest program that can produce the string. If the KC is small, then the string can be compressed and does not have much information. If the KC is large then the string cannot be compressed and it contains a lot of information. We are interested in the KC of general mathematical structures. Most mathematical structure can be described by constructions in category theory. We formulate a programing language to describe categorical constructions. We then define the KC of a mathematical structure to be the length of the shortest such program that can produce that structure. We show basic properties of this KC and explore its relationship with classical KC. We also discuss what can and cannot be constructed by category theory. The talk will begin with a short introduction to classical KC and to the basics of category theory.http://www.gc.cuny.edu/News/Calendar/Detail?id=326202_309/17/20154:15 PM09/17/20156:00 PM365 Fifth AvenueThe Graduate Center9204FreeSpeaker: Noson Yanofsky, Brooklyn College
Title: Kolmogorov Complexity of Categorical Structures32373