DSpace Collection:
http://dspace.cityu.edu.hk:80/handle/2031/3770
2016-06-25T08:51:44ZActiveCrowd: integrating active learning with Amazon Mechanical Turk
http://dspace.cityu.edu.hk:80/handle/2031/8303
Title: ActiveCrowd: integrating active learning with Amazon Mechanical Turk
Authors: Chi, Fung Cheung (池鳳翔)2015-01-01T00:00:00ZLarge-scale chromosomal 3D structure reconstruction
http://dspace.cityu.edu.hk:80/handle/2031/8304
Title: Large-scale chromosomal 3D structure reconstruction
Authors: Zhang, Yanlin (張延林)2015-01-01T00:00:00ZRegressionMaple: regression coverage of concurrent testing on validating bug-fixing
http://dspace.cityu.edu.hk:80/handle/2031/7467
Title: RegressionMaple: regression coverage of concurrent testing on validating bug-fixing
Authors: Tsui, To (徐韜)
Abstract: Multicore hardware makes performance faster. With the pervasiveness of software and
hardware support, concurrent computing is widely applied. While enjoying its benefits, there
is also a new challenge - concurrency bug. Concurrency bug is an error caused by incorrect
thread interleavings. In concurrent computing, threads are interleaved with each other to
simulate as executing in parallel. But, in fact, threads are executed one by one in a small time
slice, and communicate with each other (for example, via shared memory). Maple is one of
several software of automatic concurrency bugs detection, successfully applying dynamic
analysis to reveal concurrency bugs such as data race and deadlock. In addition, it generates
histories of tested and failed-to-test interleaving schedules. It gives a progressive method for
developers to test their concurrent software. While Maple is good at detecting concurrency
bugs with respect to the same input, it is not without its flaws. This project has observed two
situations, in which Maple is possible to be improved. The two situations are lack of accurate
coverage across versions and inability of validation on concurrency bug-fixing. First, Maple
treats versions of a program as totally different programs. It requires a full set of retest
processes on every version. It is clearly a time consuming process, as developers and testers
are often under stress to release a new version. Second, once a concurrency bug is exposed by
Maple, developers will try to resolve it. However, after suspicious codes were modified, the
developers have no information to determine if the concurrency bug is completely fixed or
not. In the current approach, they can only retest the possible interleaving schedules but
without any target in mind. To this end, this project proposes a new regression coverage
driven testing tool - RegressionMaple. It applies the concept of regression testing (with
assumption of similar execution context) to link testing information across two versions of a
program, thus improves Maple with respect to the above two problems.
Notes: Conference paper developed from this OAPS paper: Tsui, T., Wu, S., & Chan, W. K. (2014). Toward a methodology to expose partially fixed concurrency bugs in modified multithreaded programs. In Proceedings of the International Workshop on Innovative Software Development Methodologies and Practices (pp. 49-56). ACM. doi: 10.1145/2666581.2666592.2014-01-01T00:00:00ZWireless sensor network coverage problem
http://dspace.cityu.edu.hk:80/handle/2031/7466
Title: Wireless sensor network coverage problem
Authors: Fan, Haosheng (樊昊晟)
Abstract: The title of this project is called "Wireless Sensor Network Coverage Problem". The objective of the coverage problem is to use some wireless sensors
to fully cover a target area. In this project, we focus on the problem when
the target area is a line. Therefore, the problem can also be referred as "Barrier Coverage Problem". This problem has some significant applications in
the real world such as highway monitoring, seaway monitoring and border
surveillance, etc.
The aims of the project include finding some efficient algorithms, analysing
these algorithms and developing a system for naive users to find the optimal
solutions of the problem by themselves. We analyse each algorithm from
the following aspects: the approximation ratio, the time complexity of the
algorithm and the performance in experiments.
In this project, we study the objective of minimizing the total cost of a
range assignment that can fully cover the line interval. We can further divide
the problem into two sub-problems. One sub-problem is called "On-the-line
Barrier Coverage Problem" ("On-the-line"), which has been well researched
by researchers. The other sub-problem is called "Off-the-line Barrier Coverage Problem"("Off-the-line"), which very few researchers are working on.
Moreover, the "Off-the-line" problem can be considered as a generally case
of the "On-the-line" problem.
Because the "On-the-line" problem has been well researched and we think
not many extensions can be made further, we spend most of our time on
studying the "Off-the-line" problem and try to make some extensions on the "On-the-line" problem based on some previous works. In the first several
months of the project, we find an approximation algorithm to solve the "Off-the-line" problem and it can achieve an approximation ratio of 1:5. Later,
we improve the algorithm so that it can achieve a better approximation ratio
of 4/3 . Then we try to apply this algorithm to the "On-the-line" problem
and find that it can achieve an approximation ratio of 5/4 , which is the best
approximation ratio for the "On-the-line" problem as far as we know. Besides
looking for constant-approximation algorithms, we also try to use some other
techniques to solve the problem. Later in this report, we will show that this
problem can be solved by dynamic programming and finding the shortest
path under some situations. Some other related extensions have been made in
addition for both the "On-the-line" problem and the "Off-the-line" problem.
We have turned our work into two papers, "Barrier Coverage by Sensors
with Adjustable Ranges" and "Barrier Coverage Using Sensors with Off-sets". The first paper has been accepted by an international journal
called "ACM Transactions on Sensor Networks" and the second one is accepted by an international conference called "The International Conference
on Wireless Algorithms, Systems and Applications"("WASA").
In addition to these achievements in the theoretical side, we have developed a system on the problem as well. The system is written in Python.
Users can find the optimal range assignment for both the "On-the-line" problem and the "Off-the-line" problem by using our system. Our system can get users' inputs either through keyboard or an external file. The system is
robust, user-friendly and easy to use.
Overall speaking, we have achieved most of the aims we have made before the project. Furthermore, we have done quite well on the time management.
Considerable progress has been made every month during the whole year.
Notes: Journal article developed from this OAPS paper: Fan, H., Li, M., Sun, X., Wan, P. J., & Zhao, Y. (2014). Barrier coverage by sensors with adjustable ranges. ACM Transactions on Sensor Networks (TOSN), 11(1), 14. doi: 10.1145/2629518; Conference proceedings paper developed from this OAPS paper: Fan, H., Lee, V. C., Li, M., Zhang, X., & Zhao, Y. (2014). Barrier coverage using sensors with offsets. In B. Liu, A. Bestavros, D. Z. Du, & J. Wang (Eds.), Lecture Notes in Computer Science: Vol. 8491. Wireless Algorithms, Systems, and Applications (pp. 389-400). Berlin, Germany: Springer International Publishing. doi: 10.1007/978-3-319-07782-6_362014-01-01T00:00:00Z