CS224U: Natural Language Understanding

Your grade is determined based on:

• Quizzes: 15%
• Homeworks and bake-offs: 35%
• Literature review: 10%
• Experimental protocol: 10%
• Final project paper: 30%

The letter grade policy is detailed below.

There will be a series of short quizzes covering core class content. They will be done on Canvas. They will always be open notes, open book, etc., but no discussion or collaboration with others is permitted, nor can others' work for CS224u (this quarter or previous quarters) be consulted in any way. Quizzes cannot be completed for credit after their due date. The goal of these quizzes is to create a course-related incentive for individual students to study the material beyond what is required for the more free-form and collaborative assignments.

Our goals for the homeworks and bake-offs: (i) to raise important questions, (ii) to exemplify best-practices for NLU experiments, and (iii) to help you master central NLU concepts.

The homeworks and bake-offs come in pairs and follow a common pattern:

1. In the homework, you build some baseline models and develop your own model, and this accounts for 9 of the 10 points.
2. In the associated bake-off, you enter your model into a competition centered around evaluation on a previously unseen test set. All entries get the additional homework point, and top entries receive a bit of extra credit. Precisely what it means to be a top entry might vary a bit by bake-off.

Policies:

• All homeworks must be submitted by the start of class on the day they are due. The associated bakeoff submission is always due at that same date-time as well.
• Collaboration on weekly assignments and bake-offs is permitted, with a maximim group size of 3.
• All the homeworks culminate in an "original system" question that becomes your bake-off entry. Grading these is more subjective than it is for the other assignment questions. Here are the basic guidelines we will adopt for grading:
  1. We want to emphasize that the submissions needs to be an original system. It doesn't suffice to download code from the Web, retrain, and submit, even if this leads to an outstanding bakeoff score. You can build on others' code, but you have to do something new and meaningful with it.
  2. Systems that are very creative and well-motivated will be given full credit even if they do not perform well on the bakeoff data. We want to encourage creative exploration!
  3. Other systems will receive less than full credit, based on the judgment of the teaching team. The specific criteria will vary based on the nature of the assignment. Point deductions will be justified in feedback.

1. Each student will have a total of 4 free late (calendar) days applicable to any assignment (including the lit review and project milestone) except the final project paper. Final project papers cannot be turned in late under any circumstances.
2. Free late days can be used at any time, no questions asked. Each 24 hours or part thereof that a homework is late uses up one full late day. Once these late days are exhausted, any homework turned in late will be penalized 10% per late day.
3. If a group's assignment is late n days, then each group member is charged n late days.
4. Late days are never transferrable between students, even students in the same group.
5. All late work must be turned in by June 10, 11:30 am Pacific

For details, see the final projects page.

On the one hand, we want to encourage you to pursue unified interdisciplinary projects that weave together themes from multiple classes. On the other hand, we need to ensure that final projects for this course are original and involve a substantial new effort.

To try to meet both these demands, we are adopting the following policy on joint submission: if your final project for this course is related to your final project for another course, you are required to submit both projects to us by our final project due date. If we decide that the projects are too similar, your project will receive a failing grade. To avoid this extreme outcome, we strongly encourage you to stay in close communication with us if your project is related to another you are submitting for credit, so that there are no unhappy surprises at the end of the term. Since there is no single objective standard for what counts as "different enough", it is better to play it safe by talking with us.

Fundamentally, we are saying that combining projects is not a shortcut. In a sense, we are in the same position as professional conferences and journals, which also need to watch out for multiple submissions. You might have a look at the ACL/NAACL policy, which strives to ensure that any two papers submitted to those conferences make substantially different contributions – our goal here as well.

Please familiarize yourself with Stanford's honor code. We will adhere to it and follow through on its penalty guidelines.

Important: the policy on plagiarism says,

For purposes of the Stanford University Honor Code, plagiarism is defined as the use, without giving reasonable and appropriate credit to or acknowledging the author or source, of another person's original work, whether such work is made up of code, formulas, ideas, language, research, strategies, writing or other form(s). Moreover, verbatim text from another source must always be put in (or within) quotation marks.

We intepret "another person's original work" to include content that was produced by an AI writing assistant like ChatGPT. This follows either by treating the AI assistant as a person for the purposes of this policy (controversial) or acknowledging that the AI assistant was trained directly on people's original work. Thus, while you are not forbidden from using these tools, you should consider the above policy carefully and quote where appropriate. Prose that is in large part quoted from an AI assistant are very unlikely to be evaluated positively. In addition, if a student's work is substantially identical to another student's work, that will be grounds for an investigation of plagiarism regardless of whether the prose was produced by an AI assistant.

From Stanford's Office of Accessible Education:

Stanford is committed to providing equal educational opportunities for disabled students. Disabled students are a valued and essential part of the Stanford community. We welcome you to our class.

If you experience disability, please register with the Office of Accessible Education (OAE). Professional staff will evaluate your needs, support appropriate and reasonable accommodations, and prepare an Academic Accommodation Letter for faculty. To get started, or to re-initiate services, please visit oae.stanford.edu.

If you already have an Academic Accommodation Letter, we invite you to share your letter with us. Academic Accommodation Letters should be shared at the earliest possible opportunity so we may partner with you and OAE to identify any barriers to access and inclusion that might be encountered in your experience of this course.

This is the system we will use at the end of the quarter to map numerical final grades to letter grades. No curve is applied, and there are no other factors shaping the mapping from weighted averages (details here) to letter grades.