-
Deep transfer learning for system identification using long short-term memory neural networks
Authors:
Kaicheng Niu,
Mi Zhou,
Chaouki T. Abdallah,
Mohammad Hayajneh
Abstract:
Recurrent neural networks (RNNs) have many advantages over more traditional system identification techniques. They may be applied to linear and nonlinear systems, and they require fewer modeling assumptions. However, these neural network models may also need larger amounts of data to learn and generalize. Furthermore, neural networks training is a time-consuming process. Hence, building upon long-…
▽ More
Recurrent neural networks (RNNs) have many advantages over more traditional system identification techniques. They may be applied to linear and nonlinear systems, and they require fewer modeling assumptions. However, these neural network models may also need larger amounts of data to learn and generalize. Furthermore, neural networks training is a time-consuming process. Hence, building upon long-short term memory neural networks (LSTM), this paper proposes using two types of deep transfer learning, namely parameter fine-tuning and freezing, to reduce the data and computation requirements for system identification. We apply these techniques to identify two dynamical systems, namely a second-order linear system and a Wiener-Hammerstein nonlinear system. Results show that compared with direct learning, our method accelerates learning by 10% to 50%, which also saves data and computing resources.
△ Less
Submitted 6 April, 2022;
originally announced April 2022.
-
The Transfer Student Experience: It's A Lot Like Buying a Used Car
Authors:
Gregory L. Heileman,
Chaouki T. Abdallah,
Andrew K. Koch
Abstract:
The experience transfer students encounter as they navigate their journeys from community college to university is similar to that of buying a used car. We demonstrate this by showing how the information asymmetry in the market for used cars also occurs in the market for transfer students, producing inefficient markets in both cases, thereby increasing the chances of adverse selection. We diagnose…
▽ More
The experience transfer students encounter as they navigate their journeys from community college to university is similar to that of buying a used car. We demonstrate this by showing how the information asymmetry in the market for used cars also occurs in the market for transfer students, producing inefficient markets in both cases, thereby increasing the chances of adverse selection. We diagnose the underlying conditions that produce transfer inefficiencies, identifying them as a structural inequity within the system of higher education. Finally, recommendations for alleviating information asymmetry in transfer processes, that would lead to better outcomes for transfer students, are provided.
△ Less
Submitted 1 March, 2022;
originally announced March 2022.
-
Curricular Complexity Versus Quality of Computer Science Programs
Authors:
Gregory L. Heileman,
Hayden W. Free,
Johnny Flynn,
Camden Mackowiak,
Jerzy W. Jaromczyk,
Chaouki T. Abdallah
Abstract:
In this research paper we describe a study that involves measuring the complexities of undergraduate curricula offered by computer science departments, and then comparing them to the quality of these departments, where quality is determined by a metric-based ranking system. The study objective was to determine whether or not a relationship exists between the quality of computer science departments…
▽ More
In this research paper we describe a study that involves measuring the complexities of undergraduate curricula offered by computer science departments, and then comparing them to the quality of these departments, where quality is determined by a metric-based ranking system. The study objective was to determine whether or not a relationship exists between the quality of computer science departments and the complexity of the curricula they offer. The relationship between curricular complexity and program quality was previously investigated for the case of undergraduate electrical engineering programs, with surprising results. It was found that if the US News & World Report Best Undergraduate Programs ranking is used as a proxy for quality, then a statistically significant difference in curricular complexities exists between higher and lower quality electrical engineering programs. Furthermore, it was found that higher quality electrical engineering programs tend to have lower complexity curricula, and vice versa. In the study reported in this paper, a sufficient amount of data was collected in order to determine that an inverse relationship between program quality and curricular complexity also exists in undergraduate computer science departments. This brings up an interesting question regarding the extent to which this phenomenon exists across the spectrum of STEM disciplines.
△ Less
Submitted 11 June, 2020;
originally announced June 2020.
-
Curricular Analytics: A Framework for Quantifying the Impact of Curricular Reforms and Pedagogical Innovations
Authors:
Gregory L. Heileman,
Chaouki T. Abdallah,
Ahmad Slim,
Michael Hickman
Abstract:
In this paper we articulate a framework for quantifying the complexity of curricula based on their fundamental structural and instructional properties. We then introduce the notion of curricular analytics as a means of relating curricular complexity to student success outcomes, and we demonstrate the usefulness of curricular analytics by applying them to a number of important problems.
In this paper we articulate a framework for quantifying the complexity of curricula based on their fundamental structural and instructional properties. We then introduce the notion of curricular analytics as a means of relating curricular complexity to student success outcomes, and we demonstrate the usefulness of curricular analytics by applying them to a number of important problems.
△ Less
Submitted 14 November, 2018;
originally announced November 2018.
-
Route Distribution Incentives
Authors:
Joud Khoury,
Chaouki T. Abdallah,
Kate Krause,
Jorge Crichigno
Abstract:
We present an incentive model for route distribution in the context of path vector routing protocols and we focus on the Border Gateway Protocol (BGP). BGP is the de-facto protocol for interdomain routing on the Internet. We model BGP route distribution and computation using a game in which a BGP speaker advertises its prefix to its direct neighbors promising them a reward for further distributi…
▽ More
We present an incentive model for route distribution in the context of path vector routing protocols and we focus on the Border Gateway Protocol (BGP). BGP is the de-facto protocol for interdomain routing on the Internet. We model BGP route distribution and computation using a game in which a BGP speaker advertises its prefix to its direct neighbors promising them a reward for further distributing the route deeper into the network, the neighbors do the same thing with their neighbors, and so on. The result of this cascaded route distribution is an advertised prefix and hence reachability of the BGP speaker. We first study the convergence of BGP protocol dynamics to a unique outcome tree in the defined game. We then proceed to study the existence of equilibria in the full information game considering competition dynamics. We focus our work on the simplest two classes of graphs: 1) the line (and the tree) graphs which involve no competition, and 2) the ring graph which involves competition.
△ Less
Submitted 19 September, 2009;
originally announced September 2009.
