Labs
From 6.034 Wiki
Currently released labs
- Lab 0 ~ due Thursday, September 10 by 10:00pm
Anticipated lab release schedule
Labs are due by 10:00 pm on the due date.
Labs are anticipated to be released by 11:59 pm on the release date. Each lab release will be announced by email.
More lab info coming soon....
| Lab | Subject | Released | Due |
|---|---|---|---|
| lab 0 | Getting Started | 8/26 | 9/10 |
| lab 1 | Search | 9/9 | 9/17 |
| lab 2 | Games | 9/14 | 9/24 |
| lab 3 | Constraint Satisfaction Problems | 9/16 | 10/1 |
| lab 4 | Rule-Based Systems | 9/23 | 10/8 |
| lab 5 | Bayesian Inference | 9/28 | 10/22 |
| lab 6 | k-Nearest Neighbors & ID Trees | 10/5 | 10/29 |
| lab 7 | Neural Nets | 10/7 | 11/5 |
| lab 8 | Support Vector Machines | 10/19 | 11/12 |
| lab 9 | Adaboost | 11/2 | 12/3 |
The anticipated release and due dates for all unreleased labs are subject to change.
The online grader
You will be submitting all of your labs to an online grader. Every lab comes with a file, tester.py, that contains the machinery to test your code and to submit it when you're done.
In order for this to work, you need to securely download a "key" that identifies who you are to the grader.
Make sure you have an MIT certificate, and go to https://ai6034.mit.edu/labs . This page will provide you with a file called key.py. Keep this file secure; for example, don't put it in a publicly-readable Athena directory.
If you do not have a valid certificate, https://ai6034.mit.edu/labs will load, but will display an error message telling you something is wrong with your certificate. If you encounter this issue, please reinstall your certificate and restart your browser.
The only thing the grader cares about is whether you pass the tests. It does not care if your code is pretty or well-commented. However, commenting your code can still be important: if you want a TA to help you with your lab, they will be able to give you more help if your code is understandable.
The grader also submits the code to your lab, so that it can be reviewed later by a human. It should go without saying that you should not try to fool or work around the grader, and that the code you submit must be the code you tested. See our grading and collaboration policy, which also explains how your problem set grade is calculated.
Note that each lab grade is the maximum score out of all of your submissions for that lab. This means that if you re-submit a lab, and the command line output shows a lower grade, your grade for the lab will not actually decrease.
