CS3 Monthly Research Exchange

The Center for Smart Streetscapes (CS3) Student Leadership Council (SLC) presents the CS3 Monthly Research Exchange! At each CS3 Monthly Research Exchange, we hear from researchers about their latest findings on the future of smart city technology. Attendees will have the opportunity to engage with other CS3 students and faculty at partner institutions, provide feedback to the student presenters, and collaborate on future research.

Lunch will be provided.

Student Presenters:
Ege Ozguroglu
PhD Student, Columbia University
Research Thrust 2: Situational Awareness

• Abstract: We introduce pix2gestalt, a framework for zero-shot amodal segmentation, which learns to estimate the shape and appearance of whole objects that are only partially visible behind occlusions. By capitalizing on large-scale diffusion models and transferring their representations to this task, we learn a conditional diffusion model for reconstructing whole objects in challenging zero-shot cases, including examples that break natural and physical priors, such as art. As training data, we use a synthetically curated dataset containing occluded objects paired with their whole counterparts. Experiments show that our approach outperforms supervised baselines on established benchmarks. Our model can furthermore be used to significantly improve the performance of existing object recognition and 3D reconstruction methods in the presence of occlusions.

Pierre Tholoniat
PhD Student, Columbia University
Research Thrust 3: Privacy, Security, and Fairness

• Abstract: As major browsers phase out third-party cookies, emerging advertising APIs offer an opportunity to improve web privacy. We first present Cookie Monster (published at ACM SOSP ’24), a system that enhances existing advertising measurement APIs from major tech companies with more efficient differential privacy (DP) budgeting. By using an individual form of DP, our approach enables more accurate private measurement queries compared to traditional DP implementations. Cookie Monster lays the foundations for on-device privacy-preserving systems, with applications beyond advertising: in this talk, we propose an analogy between web and smart city privacy, and sketch how insights from our paper can help shape a robust privacy architecture for smart cities.

The CS3 Monthly Research Exchanges are internal and open only to CS3 affiliated students, faculty, and staff. If you are interested in learning more about the research happening at CS3, please contact our team at streetscapes@columbia.edu.