Spam-Email-Classification--垃圾邮件分类

group members:

• CHEN Kaiwei

• Guoping Liang

• Cheng Zheping

• Zheng Fuchen

• Liu Renhao

• Wang Zhonghui

• Tang Jingtian

• Lin Jiaxin

• Yao Yu

Learning Objectives:

(a) Practice communication and organization skill; (b) Practice writing skill and presentation skill; (c) Survey current capabilities of AI software; (d) Learn how to evaluate AI software;

Task:

Carry out an evaluation of at least 2 pieces of AI software/package that can solve the same problem (i.e., do not survey one software for knowledge representation the other on planning) so that you can compare their performance. You should select one software from one package and compare more than one package. Based on the commonalities, carry out a comparison between the (two) chosen pieces of software. Finally, you should make appropriate recommendations to the choice of using which AI software for specific tasks.

Topics:

1. RNN/LSTM

2. CNN

3. BERT

Issues for Consideration:

(a) Justify why you have selected the pieces of AI software for comparison; (b) Can you draw a matrix of features, as well as pros and cons, of the chosen pieces of software for comparison? (c) Can you objectively/subjectively evaluate the performance (e.g., effectiveness, efficiency, user friendliness, ease for development, etc.) of the chosen pieces of software? For example, download the demo version of the software and apply it to some test data to measure the effectiveness for comparison. (d) What measurements should you use to compare the software? (e) Do you think that your comparison is fair? (f) What practical problems are there to deal with when using such software, especially when the problem size is big for real-world applications? (g) Which software would you recommend to solve what problems?

Deadlines:

April 13, 2022 (Hand in the softcopy of the report and the powerpoint slides using Blackboard)