0-1 Validation
A good research project begins with a good scientific question. A scientific question may be a what, how, or why question. Such questions come from three sources: research papers, real-world needs, and flashes of insight during the research process.
Scientific questions are often not too few, but too many. A good scientific question should be stated simply, offer clear benefits, present a reasonable level of difficulty, and have a feasible solution. It should be unprecedented and influential once solved, although the degree of novelty and the magnitude of impact are often inversely related.
Once the question is clear, the student should discuss a series of algorithms with the advisor and examine where each algorithm does and does not apply. With coding agents, the student should be able to develop a minimum validation system quickly and obtain valid results. After discussing the results with the advisor, the student can invest more time.
It is important to recognize that understanding a scientific question often takes time. Practice matters, and revising the definition of a scientific question through practice is common.
Cognitive Alignment
The student should discuss with the advisor what is correct, what is incorrect, and why common errors occur. The student should continuously report what has been validated as correct, the problems encountered, and possible solutions. When facing a problem that cannot be solved independently, the student should seek help from the advisor proactively and promptly to avoid continued ineffective effort.
Investigation and Validation
The student should clearly understand what to do, what not to do, and what to optimize, and should complete system design, implementation, and validation over an extended period. When using coding agents, the student must validate both their processes and their results.
Paper Submission
Once the system design and implementation are largely settled and initial experimental results are available, the student should agree with the advisor on a target conference and submission milestones. The problem, method, experiments, and presentation should then be completed against the review criteria. The student should write the paper, organize reviewer comments, and draft responses, working with the advisor in an alternating and continuous process to advance the submission.
Strengthening Research Outcomes
After the paper is submitted or accepted, the student should organize the outcomes and try to develop the paper into courses, tools, or services for communication, project applications, collaboration, and technology transfer.
Reflection and Improvement
The student and advisor should reflect continuously and improve through practice. This process management approach should be updated continuously.