Learing Outcomes

• Learn the foundational methods used in NLP from first principles in statistics, algorithms, and lingusitics.
• Understand key facts about human language that motivate NLP methods, and critically disern what problems are possible to solve.
• Implement, experiment with, evaluate, and improve NLP models, gaining practical skills for building natural language systems.
• Learn about and navigate the process of an open-ended NLP research project.
• Reason about the ethical and social implications that arise from NLP systems.

Communication and Piazza

• All communication about CS 375 will take place on Piazza.
• The one exception is for personal matters. In these cases, please email the instructor, kak5@williams.edu
• On Piazza, you are able to post annonymously (your classmates cannot see your identity). There are no such things as "dumb questions" so we encourage you to use Piazza as much as possible.
• Any questions about homeworks will be answered via (1) office hours or (2) Piazza. This means any questions asked over email will be redirected to Piazza.
• Rationale: We are in a collaborative learning environment. Any question you have, your classmates will mostly also have and we would like everyone to benefit. This also helps us answer your questions more quickly.

Assessments

• CS 375 will have several forms of assessements.
• Homeworks are a mix of (a) analytical problems about mathematical/statistical/linguistic foundations of NLP, (b) longer programming assigments, and (c) conceptutal questions.
• The Midterm Exam will take place on the evening of Wednesday, March 15. This will be a closed book written exam (on paper).
• The final project will be in groups. More details will be announced later in the semester. You are encouraged to start browsing other NLP topics early on in the course.

Grade breakdown

• The grade breakdown is as follows:
• Homeworks - 55%
• Final Project - 25%
• Midterm - 20%
• Homework 0 will be graded on completion.
• Homework 6 (in groups) will be worth twice the points as the other homeworks since this homework is roughly double the work of the others.

Late days on homeworks

• You (the student) have 3 late days which you can allocate to homework assignments throughout the semester.
• To use a late day, fill out this Google form. This will be checked after every homework deadline so there is no need to email the instructor.
• If you have used all three late days and require a special exception extension, we require a note from your class Dean.
• If a special exception extension is not granted and all late days are used, the total number of homework points possible will be reduced 20% for each day late. If the work is more than 5 days late, the work will not be accepted.
• Example: The homework is due Thursday, Feb 16 at 9:59pm. The student turns in the homework on Feb 17 at 11am. The student has no late days left and no note from the Dean so that maximum number of points the student can earn is 80/100.
• Late days cannot be used on Homework 0, Homework 6 (group project), the final project, or the Midterm.

Extra Credit

• Extra credit will be offered on some (but not necessarily all) of the assignments.
• Extra credit can only help you, not hurt you. At the end of the semester, if you have a borderline grade, extra credit will be considered to move up your grade.
• Example: At the end of the semester, a student is on the border between a B+ and A-. The student has done all the extra credit so the instructor awards them an A-.
• Rationale: We want you to be invested in the learning process and remain curious about NLP. Extra credit can possibly help this as well as mitigate concerns over small point losses.

Expectations and Norms

You can expect me (the instructor) to:

• Contribute to and support a respectful and welcoming environment.
• Start and end class on time.
• Craft lectures and assignments designed to help you learn the material.
• Release assignments and provide feedback in a timely matter.
• Reply to emails and Piazza posts within 24 hours on weekdays and 48 hours on weekends.

We can expect you (the students) to:

• Contribute to and support a respectful and welcoming environment.
• Attend all lectures in person except for health emergencies or extenuating circumstances.
• Arrive to class and lab on time, and plan to stay until the end.
• Stay engaged in the class and material.
• Reach out for help from the TAs or instructors.
• Adhere to the Honor Code.

Class Norms:

• If you become sick with COVID or another illness please stay home and let us know via email. We will record lectures and then post them to Piazza.
• If you must miss class for other reasons please give us as much advance notice as possible.
• The Computer Science department strives to be a friendly and welcoming community. You may find it slightly less formal (but no less respectful) than what you encountered in previous academic settings. For example, most students and faculty address other faculty by their first names. You are welcome to call me "Katie" as well.
• You are also welcome to address me informally in email (i.e. starting an email with “Hi Katie.”) Here are a few other tips for emailing professors if that is something new to you or out of your comfort zone.
• Katie has set office hours. Feel free to use these times to discuss questions adjacent to the course.
• Katie uses she/her pronouns. We will try to use your prefered pronouns, as indicated in PeopleSoft. Please don’t hesitate to correct us.

Honor Code

For computer assignments in computer science courses, the honor code is interpreted in very specific ways. Homeworks are expected to be the work of the individual student unless otherwise designated, designed and coded by them alone. Help locating errors and interpreting error messages is allowed, but a student may only receive help in correcting errors of syntax; help in correcting errors of logic is strictly forbidden. In general, if you are taking photos of someone else’s screen, looking at someone else’s screen, or telling someone else what to type, it is likely the work is no longer the work of an individual student.

The College and Department also have computer usage policies that apply to courses that make use of computers. Read more about these policies here.

Sharing Solutions.

Please do not post your solutions to our assignments in any public forum, including public GitHub repositories. Students taking the course should not be looking for solutions, but tempting them by making solutions available is inappropriate. This applies not just to the semester you are taking the course, but to the future as well.

If in doubt as to what is appropriate, do not hesitate to ask. I'm happy to discuss this anytime.

Large LMs policy

• Because this is an NLP class, we encourage you to use and interact with large language models (LMs), e.g. ChatGPT.
• However, if you use a large LM you must acknowledge (via writing) what ideas/language is the LMs and what is yours. Be sure to note the specific LM you used, the prompt you inputted, and whether you modified or edited the response.
• Example: "I used ChatGPT with the prompt 'What is the ethics of using a large language model?' Then I edited the output as the first paragraph if my response. The second paragraph of the response is my own thoughts in response to the first paragraph."
• Lack of citation or improperly citing your use of large LMs is a violation of the honor code and will be brought before the honor court.

Accommodations

If formal accommodations need to be made to meet your specific learning or physical abilities, you should contact your instructors as possible to discuss appropriate accommodations. You should also contact the Director of Accessible Education, Dr. G. L. Wallace (x4672) or the Dean’s office (x4171). We will work together to ensure this class is accessible and inclusive.

Mental Health

If you are experiencing mental or physical health challenges that are significantly affecting your academic work, you are encouraged to contact your instructor and/or speak with Dean’s Office staff (x4171).

Public Health

If you feel ill, please do not come to class or lab and let us know if you are unable to attend class due to COVID restrictions. We will work with you to make sure you can make up any missed work, and to develop a plan that allows you to continue making progress in the course during your time in isolation/quarantine. In particular, lectures will be recorded and then posted on Piazza.

Acknowledgement

Parts of this course are adapated from Dan Jurafsky's CS 124 at Stanford and Brendan O'Connor's CS490A at UMass Amherst.