1. Grading Reading Assiginments
The summaries will be graded on the following four-point scale:
- 4: All main concepts and skills mastered and all major questions answered, probably with minor errors
- 3: Important points made, but contains some significant omissions or errors
- 2: Substantial missing concepts or errors
- 1: Effort shown, but not a significant amount of relevant or correct content
- 0: Not turned in, or almost no effort or understanding demonstrated
2. Grading Paper Presentations
In a paper presentation, please include the following components
- Problem definition (20%): What is the research problem studied in this paper? Why this problem is important to the NLP community?
- Related work (10%): A brief explanation of previous works and their limitations.
- Proposed method (30%): A detail description of proposed method and its novelty. (If necessary, also provide some background knowledge, so the audience can fully understand the technical content.)
- Evaluation (30%): (1) A brief description of evaluation setup, including datasets, baselines and competitive systems, evaluation measurements etc.; (2) Experimental results and key observations from experiments. (If necessary, also provide some background knowledge, so the audience can fully understand the technical content.)
- Takeaways (10%): important things to be remembered from this paper
3. Mapping Betweeen Points and Letter Grade