The Structure of Information Networks (Jon Kleinberg)

The Structure of Information Networks: Coursework
Fall 2026

The assigned work for the course will consist of

two problem sets,

an exploratory paper with accompanying presentation, and

a final project with accompanying presentation (and beginning with a short roadmap document).

The grading will be based 10% on each problem set, 30% on the exploratory paper, and 50% on the project. Here is a more detailed description of each of these components.

Problem Sets

We will begin with two problem sets; the goal here is to get everyone to practice some skills that will be crucial in doing the project. Specifically, in the project, we will be looking for you to work with real network data, and also to work with mathematical models of networks. Thus, the first problem set will be a computational one concerned with a specific network dataset, and the second problem set will be a mathematical one concerned with network models.

The first problem set will be handed out on Sep. 2 and due on Sep. 16. The second problem set will be handed out on Sep. 16 and due on Sep. 30. You may discuss the problem sets with other students in the class, but since the goal is to practice skills, you should each hand in an individual solution.

Exploratory Papers

The course includes a lot of material from recent research; this means that much of it exists mainly in the form of papers on the Web, and the existing literature raises a lot of interesting issues that have yet to be explored.

As a way to get everyone thinking about the research issues underlying the course, and to prepare for the project, there will be a short exploratory paper of roughly 5 pages in length. You can work in groups of up to 4 people on the paper. The exploratory paper is a good way to work on formulating a topic for the course project: your group for the exploratory paper can become your group for the project (which also can be done in groups of 4), and your exploratory paper can form the basis for your project if you are interested in continuing and extending it.

The exploratory papers will be due on November 2, and the work will be evaluated in two ways: first from the paper itself; and second, during the week of November 2-6 each group will give a 20-minute one-on-one presentation of the exploratory paper to one of the members of the course staff (the faculty or one of the TAs).

The exploratory paper should be structured as follows. First, you should read at least two closely related papers relevant to a particular section of the course, at least one of which is not linked from the course home page. You should then write approximately 5 pages in which you address the following points:

What is main technical content of the papers?

Why is it interesting in relation to the corresponding section of the course?

What are the weaknesses of the papers, and how could they be improved?

What are some promising further research questions in the direction of the papers, and how could they be pursued?

What are some initial steps you can make in the direction of follow-up research starting from these papers?

Regarding this last point, exploratory papers should not just be summaries of the papers you read; most of your text should be focused on synthesis of the underlying ideas, and your own perspective on the papers. To make this concrete, you should make sure that you devote much of the content to the last two bullet points above: promising directions for further research, and initial steps in the direction of this further research. In particular, the exploratory paper should contain at least some amount of each of the following types of content:

A proposal for a model or algorithm -- potentially extending, varying, or improving something in the papers you've read -- together with some mathematical analysis of it.

A test of a model or algorithm (either your own or something from one of the papers) on a dataset or on simulated data.

The one-on-one presentation of the exploratory paper will involve the project group and one member of the course staff, and it will be 20 minutes in length. The goal is to have a discussion of the exploratory paper rather than a straight recitation of it, and therefore the presentation will involve no slides, scripts, or devices, but will just be a discussion with questions and answers interspersed. We'll use a format without written notes as a way to make sure the presentation doesn't lapse into something that just follows a script, but instead has back-and-forth discussion. The one-on-one presentations will all take place in person, with the natural exception that members of project groups taking the class from Cornell Tech will need to be on Zoom.

Project

The final piece of the work for the course will be a project. You can work on this in groups of up to 4 people, and it is largely up to you to define the topic and scope of the project.

The first step in the project will be a short roadmap document. This is meant just to be a brief description of what you're intending for the project -- about a page in length, with a discussion of relevant background work and tentative plans for how you'll proceed. If your project is based on your exploratory paper, then you don't need to repeat things you've said in the exploratory paper -- it's enough to describe how you're planning to turn the ideas from the exploratory paper into a larger project.

The roadmap document will be due November 16.

The basic genres of project are the following:

An experimental evaluation of an algorithm, model, or measure on an interesting dataset. The papers on the course home page suggest some possible domains in which to think about such experiments; but you can also assemble your own data.

A theoretical project that considers an algorithm, model, or measure in the area of some course topic, and derives rigorous results about it.

An extended, critical survey of one the course topics, going into significant depth and offering a novel perspective on the area.

As with the exploratory paper, the project should contain at least some amount of mathematical analysis, and some computations on real or synthetic data.

The result of the project will be three pieces of work. First, you should hand in a 10-15 page paper, describing the approach, the results, and the related work. The references on the course home page serve as examples of what such papers tend to look like; of course, the overall form of the paper will depend on the nature of the project.

Second, there will be an open presentation of the projects in class by each group. The open presentations will take place during the final three classes of the semester, on November 30, December 2, and December 7. The open presentations will be very short in format, 4-6 minutes in length (depending on the exact number of groups and the available time), and based on slides (a maximum of two slides per presentation) that each group should submit in advance by November 23.

Third, there will be a one-on-one presentation of the project to a member of the course staff, 20 minutes in length, which will be scheduled to take place between November 30 and December 7 inclusive. This will have the same one-on-one question-and-answer style as the exploratory paper presentation, with no slides, scripts, or devices so as to ensure a discussion-oriented format. (And the same exception that we'll need to use Zoom for members of project groups taking the class from Cornell Tech.)

The exact time slots for both the open and one-on-one presentations will be worked out later in the semester. Since the presentations will happen before the final write-up of the project is due, you can view these as progress reports on the project, previewing the work done so far, rather than presentations on the full completed project.

The final write-up of the project will be due on a date during the final exam period, chosen by the university's central scheduling process. (The university's process gives them the choice of final project due dates, rather than allowing individual classes to choose these dates.) This date will be announced later in the semester when it is available.

Use of AI

AI models are often used by research groups to help with projects like the ones being assigned in this course, and so accordingly we will allow the use of AI models for doing the class projects as well, as well as the steps leading up to it: the two problem sets, the exploratory paper, the roadmap, and the preparation of the presentations. As with all things involving the use of a new technology like AI language models and agents where the norms are still in flux, it helps to articulate some guidelines.

First, you should document the ways in which you use AI in your work, including the workflow (which models, what prompts, and other interactions). This is useful so that we can understand what you did, and it's also useful to see how students are using AI in a class like this in Fall 2026. You will not lose credit from your grade for using AI, but we do want to know what you did.

Second, for the parts of the project that AI models help with, we want to make sure that you are able to stand behind the work and vouch for it, all the way to being able to present the work without the assistance of any AI models who may have helped. In this way, while the one-on-one presentations are helping you practice extremely useful expositional skills in their own right, the design choice to have them involve no slides, scripts, or devices is partly in service of ensuring that you can present the work without the assistance of AI.

Third, since the point of the course is to learn techniques, it is very useful to attempt a productive engagement with the material and the work before turning to AI models. This is particularly true of the initial problem sets: they are a small portion of the course grade, and AI models are able to handle them, but since the entire purpose of the problem sets is to practice skills that will be useful in the eventual course project, we strongly recommend you try the questions yourself before deciding whether you want to invoke the assistance of AI models. Having said that, it is within the scope of the course policies for you to use the assistance of AI models on the problem sets, as it is with other parts of the coursework.

Academic Integrity and Plagiarism

You are expected to maintain the utmost level of academic integrity in the course. Any violation of the code of academic integrity will be penalized severely, and can lead to failing the course.

Plagiarism deserves special mention here. Including text from other sources in an exploratory paper or project write-up without quoting it and providing a proper citation constitutes plagiarism. This is a serious form of academic misconduct, and instances of plagiarism will very likely result in failing the course. Plagiarism is not just word-for-word copying; it also includes cases in which no full sentence has been copied from the original source, but large amounts of text have been closely paraphrased without proper attribution.