ORIE 4741: Participation

Participation is worth 15% of your course grade. To get full credit, you will need to complete the participation requirement for all but 4 lectures over the course of the semester. You can fulfill the participation requirement for each lecture by doing either of the following.

• Attending class synchronously and answering questions in class using the iClicker app. You can do this whether you're in the classroom or on Zoom. These questions will be graded on completion. Missing one question is fine. Most students who are attending on Zoom prefer to install the iClicker app on their phones or another device so it doesn't get in the way of viewing the lecture slides.

• Watching the lecture asynchronously and completing the participation form any time before the beginning of the following lecture. The async participation form will be graded for correctness.

Details on async participation

If you're submitting an asynchronous participation form, follow these guidelines to ensure you get full credit.

1. The participation form for each class should be submitted before the beginning of the next class. The responses are timestamped and answers will not be graded if the submission is late.

2. Participation responses for polling questions are graded for correctness. (Often multiple answers are correct.) Use only comma-separated letters A,B, etc depending on the question. Do not include any other symbol or text. Make sure that all the questions are answered. (Note that there's no need to use the iClicker app if you're submitting the async form.)

3. The summary of the lecture must capture the key takeaways from the lecture; notably, it should be more than just an outline of the lecture. The intent of asking you to summarize is to help you learn and review the material. Poor summaries will be graded accordingly.

Here are two examples of good summaries:

• The lecture covered definitions of what it means for the data to be big and messy and the definition of supervised learning. We went over an example about credit decisions that clarified how we define the input space X, output space Y, and function f for the learning problem. Then we learned a little about the hardware limitations. We went through an example dataset from ACS. The jupyter notebook demo looked at a subset of this dataset which we checked for size and plotted the graph which showed that although the summary statistics were the same, scatter plot showed the variation. Plot the data as even a simple scatter plot can illustrate data variations which may be hidden by the summary statistics.

• We first talked about linear classification in the context of the credit card approval example. It was helpful here to represent the data points on a graph and to perform a projective transformation on the data such that the offset b was encoded into w. We then defined the inner product as a measure of how sure you are of the classification of x and formalized the perceptron algorithm. The margin classifier is a signed notion of distance to the boundary, so we were able to show with this fact that by iterating through the algorithm, we increase this margin (which is better!). Lastly, we went through a proof of convergence to show that if the data is linearly separable, the perceptron algorithm will terminate in finite time.

4. Provide a valid comment or question. One word comments (like “nice”) are not valid responses. What did you find exciting? What do you want to explore further? Did you find anything confusing? What more do you want to learn about the topics covered?

Why are we doing this?

It's easy to fall behind in class, especially when you're not attending class in person. These tiny milestones are intended to keep you on top of the course material, so you're not scrambling at the end of the semester. They also help us assess what you're learning, so that we can adapt the course material to help you learn better.

Finally, the small added difficulty of the async option provides mild encouragement to attend class synchronously, which means you can get your questions answered real-time and shape your own class experience. If you want your professor to teach what you want to learn, come to class and ask your questions!