Professor Araújo studies how individual behaviors give rise to emergent patterns in communities — from political movements to religious shifts. If you're fascinated by the complex forces shaping societies around the world (especially in the Global South), this might be the research home for you.

Working as a computational modeler means sitting at the intersection of disciplines: you'll draw on sociology, history, and psychology just as much as on mathematics and algorithms. Students who thrive here enjoy reading broadly, thinking abstractly, and translating real-world phenomena into equations and simulations.

Current projects include:

• Modeling political polarization in Brazil, with attention to religious identity and political friction
• Modeling church communities in the United States and examining how political polarization strains their social fabric
• Implementing and testing classical sociological theories in computational environments

Methods used include data collection and agent-based modeling (ABM). No prior experience with ABM is required — just curiosity and a willingness to learn.

• Thoughtful AI software development (continuation) Continue work on developing and deploying the Thoughtful AI add-in for MS Word. Build and test interactive tools that help writers using LLM APIs. Specific sub-projects include: merging in research prototype code, improving UX, extending the add-in to work in other applications (Google Docs, Outlook, PowerPoint, etc.), and deploying the add-in to broader audiences.
• Research (continuation): Design and execute studies on how writers (and other people) use AI tools for thought, aiming for academic publications.
• AI for Education: Prof Arnold has prototyped interactions with AI to provide students personalized feedback and in-the-moment instruction with the goal of collecting evidence of learning. A student team would flesh out these prototypes and potentially study their use in education.
• AI for Translation ✓ active: Prof Arnold has also prototyped various interactive tools for helping people understand each other, especially in church contexts. For example, a careful translation workflow.
• AI for Reflection: Flesh out one of the projects listed in my blog post on AI in reflective mode.

Team members should have taken or be taking CS336 or have significant experience with web development.

The app will allow participants to upload pictures and descriptions of items they are willing to share. It will then help participants track the items, with schedules, locations, etc.

This project was begun by Bryn Lamppa, Rose Campbell, and others as a CS262 project.

See, for example, projects like The Fediverse, social media like Mastodon, video sharing like PeerTube, workspace management like NextCloud and even newer AI-focused projects like Ollama.

You will practice some hands-on Linux server administration, DNS configuration, SSL certificates, containerization with Docker, and service orchestration, and beyond deployment, maybe even design a cooperative governance framework — defining how the community makes decisions about data, access, moderation, and the use of AI — grounded in platform cooperativism literature and reflection on how Christian values and virtues would inform these decisions.

1. Use generative AI to replicate a small-scale network data analysis problem. We will investigate whether generative AI can identify additional or previously overlooked patterns and insights.
2. Scale the same network data analysis to increasingly large network measurement datasets using generative AI, and evaluate how analysis time, computational cost, analytical coverage, and quality of findings change with data volume.
3. Develop an agentic AI system that automates the major stages of the workflow—including data acquisition, preprocessing, analysis, and visualization—with minimal human intervention. This step can be done in parallel with (2) or as a new project.

• Why is this problem important? Studying darknet is important for gathering threat intelligence, understandng the attack behavior, measuring the amount of attack activities, and even preventing cyber attacks.
• What specific tasks am I responsible for?According to the proposal, there are three major tasks. Tentatively, you will be involved in the development and testing of anomaly detection method for malicious traffic.
• What skills do I need to have? Basic computer netwoking (esp. interdomain routing), pcap(ng) analysis of the network data, basic machine learning, working on Linux, data analysis, paying attention to details
• What else do I need to know? This is a collaborative research project with CAIDA at UC San Diego. Therefore, this is not only your senior project. We must be able to deliver according to a research plan.

Prof Mark Taylor in the Business Department proposed the following project for us:

AI has introduced challenges to accurate assessment, especially unproctored assessment occurring outside of the classroom. Out of a desire to meet these new challenges without continually squeezing ever more assignments into limited class periods, professors in the school of business are piloting a new assessment type: group oral exams (GOEs). So far, students have responded very positively to GOE; it gives them the opportunity to meet in teams to prepare together for the exam while still requiring each student to answer individually when the exam is in process.

There is an instructor-facing problem, however, in the logistics. Professors currently have to track three different grades for different pieces of the assignment (including a slightly complex shared grade component) and a randomized pool of exam questions with their own logic. This is a lot of work and there are many junctions in the workflow where grades are manually moved in ways that create opportunities for user-introduced errors. Automating this process with the right software app would save professors a great deal of time and increase reliability.