Guidelines for Paper Review

Guidelines for Paper Review:

1. A survey/tutorial paper should have good "classification" and "comparison" of the research "issues" and their existing "solutions".
2. A survey/tutorial paper should have "new" observations or insights on the effectiveness of existing solutions. It does not need to provide a new solution. But new observations or insights are required.
3. A survey/tutorial paper should point out "potential" further research directions for the subject under study.
4. The first and the most important criterion for a research paper is the degree of innovation, be it a new algorithm, a new model or architecture, a new experiment, or a new result. This innovation should be valuable in terms of the percentage of improvement or the degree of impact. An innovation with little value is not appreciated.
5. The second criterion for a research paper is the logic flow in the writing. Why, how, and what should be logically reasoned. A good paper should have more arguments instead of full of plain descriptions (論述而非敘述).
6. The third criterion for a research paper is the quality of English writing in terms of (1) the number of grammar or non-grammar errors, (2) organization of the entire article, (3) accuracy of the words used, (4) structure and conciseness of the sentences, (5) correctness of the references, etc.
7. A manuscript should be rejected if it does not pass some of the above criteria, revised major if it passes partially and could be fixed within 1-2 months, revised minor if could be fixed within 2-4 weeks, and accepted if it satifies the above criteria well.