The technical paper should not be longer than 10 pages, and no late submission will be accepted.
Topic 1:
It is known that fuzzy logic technology allows computerized devices to respond to situations that fall between absolute values. These devices operate in a way akin to human reasoning, by gathering complex knowledge as they work, then applying that knowledge to their task. The results are products that are more "intelligent" and user-friendly than ever before.
Fuzzy logic has been implemented in shavers (which sense the length, thickness and density of hair and adjust the speed of the motor respectively), washing machines, dryers, cameras, trains, breaks, etc. The list is rapidly growing. Besides, fuzzy logic has also been applied to real time control applications. For example it has been implemented on a software level in conjunction with artificial neural networks and genetic algorithms. Researchers are today looking at possibilities of using fuzzy logic for pattern recognition, and other applications which generally require some knowledge based expert system.
In this assignment, students are required to carry out a literature review in order to write a technical paper regarding any application of fuzzy logic technology. This helps to demonstrate the students' understanding in the subject of Fuzzy Logic beyond the normal lectures and tutorials. The technical paper must include specific description of the application, the input(s) and output(s) of the fuzzy system, the fuzzy rules and membership functions,
Topic 2:
Multilayer perceptron (MLP) is one key topic for the course of Neural Networks. The assignment for you is to write up on an application of MLP. The purpose is to learn, by looking at the various aspects of an application, the strengths and limitations of MLP. You need to find details associated with this application where MLP has been successfully used. You can find many applications of MLPs, described in a number of web pages. (Use a search engine to get interesting web pages.) Other sources are: IEEE Transactions on Neural Networks, Neurocomputing, and IEEE International Joint Conference on Neural Networks. (These are some of the journal and conference papers which are available in one of the NUS libraries.) You are also free to look for interesting application papers from other resources.
From one of the above, students are required to write a technical paper by getting hold of one paper (which should be referenced in your technical paper) that describes an application of MLP to any problems. This helps to demonstrate the students' understanding in the subject of Neural Networks beyond the normal lectures and tutorials. The technical paper should include the following details:
- A short description of the problem.
- How were the inputs and outputs of the MLP selected?
- How was training data gathered?
- How was the MLP trained (or learned)?
- How did the researchers ensure that the MLP generalized well? (Note: Generalization refers to the ability of the MLP to perform accurately on unseen data.)
- How did the MLP perform in comparison with other methods?
- Any other interesting information that you think is pertinent.
In assignment, either topic 1 or topic 2 (as attached). Preferably in topic 1. Will accept in point form notes, as i would review it myself and prepare the rest. Would be good if have diagrams (please cite sources). Kindly complete in MS Word document. Feel free to revert back to me for any queries.
Please ensure the submitted material is done by tasker and not copied elsewhere (with the exception of some cited material). And kindly ensure submitted material is not made accessible to anyone else.