Midterm
The CS 159 midterm will take place in class on Thursday, November 5. It will be a traditional pen(cil)-and-paper exam, worth 100 points.
Please read the following details carefully to understand your responsibilities and expectations before, during, and after the day of the exam.
BEFORE the day of the exam
There will be two tasks distributed to you the week before the exam (i.e., the week of October 26), during which Prof Chang will be out of town for a conference.
The first task is a practice exam. Technically, this isn't really a task since finishing it is not mandatory, and it is not worth any points. However, you are encouraged to work on it (since there's not much else happening that week) so you can get a feel for what kinds of questions (and what kinds of math) you might expect to see on the actual exam.
The second task is required and is worth 10 points on the exam: as a class, you will work together to create a shared notesheet.
For context, the exam itself is closed-book, with one exception: you are allowed access to a notesheet, which you will construct together as a class in a shared Google Doc. Everyone in the class must contribute; there will be a Gradescope assignment where you must submit proof of your contribution. Again, you are required to do this (but as long as you remember to submit something, this is basically 10 free points on the exam!)
This notesheet is the only reference material you are allowed to have on the exam. You may not bring any supplementary notes of your own. Therefore, please work together to help each other! I have full confidence that, when you put your heads together, you'll be able to create a notesheet that is comprehensive but also concise enough to work as reasonable reference material during the exam.
In order to help guide you in creating your notesheet, here is a list of the titles of the questions that will appear on the midterm:
- Tokenizer Basics: Terminology, Concepts, and Basic Tokenization
- Training and Running Simple Language Models
- Training and Running a Naive Bayes Classifier
- Evaluating a Text Classifier
- Hidden Markov Model Decoding Using the Viterbi Algorithm
- Information Retrieval Using Cosine Similarity
ON the day of the exam
As noted above, the exam will take place in class. It will have a time limit of one hour (leaving enough buffer during the class period to distribute and collect the exams).
If you think you won't be able to make it to that session, please notify Prof Chang immediately. Additionally, I currently do not have any documentation of students in the class who need an extra-time accomodation; if you believe this is in error, you should also notify Prof Chang immediately.
The exam is closed-book and closed-notes with the exception of the shared notesheet described above. You are allowed to have a device through which to access the notesheet (which will be set to read-only during the exam), or you may print a physical copy. You are also allowed to have a calculator to help with the math parts of the exam.
AFTER the day of the exam
When I grade your exam, I will not provide full explanations of what you got wrong; I will only tell you how many points you got on each question (marking where you specifically lost points).
You can recover up to 50% of points lost by following these instructions:
- Book an appointment with Prof Chang any time before the end of the semester, to discuss the questions you got wrong.
- Within 24 hours after the appointment, submit a written explanation, in your own words, of why your original answer was wrong and what the right answer was. For each question you submit an explanation for, you can recover up to 50% of lost points for that question (the exact number of points you get back is up to Prof Chang's discretion based on what exactly you say in your explanation, but the number will not be zero).
(When logged in, completion status appears here.)