Jake Maranzatto

Jake Maranzatto

Postdoc @ UMD ECE

Posts by Collection

portfolio

publications

Evaluating Audience Engagement of an Immersive Performance on a Virtual Stage

Published in Frameless Symposium, 2020

Completed at the RIT REU on Computational Sensing. Joint with Victoria Kraj, Joe Geigel, Reynold Bailey and Cecilia Ovesdotter Alm.

Recommended citation: T. Kraj, T. J. Maranzatto, R. Bailey, C. O. Alm, J. Geigel. “Evaluating Audience Engagement of an Immersive Performance on a Virtual Stage,” Frameless: Vol. 2: Iss. 1, Article 4, 2020 Available at: https://repository.rit.edu/frameless/vol2/iss1/4
Download Paper | Download Slides | Download Bibtex

Using machine learning algorithms (supervised) to generate automatically labeled datasets for detecting digital dating abuse from text message

Published in Proceedings of the Florida Artificial Intelligence Research Society (FLAIRS), 2023

Used ML to classify and label text messages as abusive or nonabusive. My contributions were completed as an undergraduate student. Joint with Tania Roy and Zachary Loomas

Recommended citation: T. Roy, T. Maranzatto, and Z. Loomas, "Using machine learning algorithms (supervised) to generate automatically labeled dataset for detecting digital dating abuse from text messages", FLAIRS, vol. 36, no. 1, May 2023.
Download Paper

A Unified Analysis of Dynamic Interactive Learning

Published in Allerton Conference on Communication, Control, and Computing (Allerton), 2023

We study an online learning problem on graphs, meant to model a recommender system. Joint with Xing Gao and Lev Reyzin.

Recommended citation: X. Gao, T. J. Maranzatto and L. Reyzin, "A Unified Analysis of Dynamic Interactive Learning," 2023 59th Annual Allerton Conference on Communication, Control, and Computing (Allerton), Monticello, IL, USA, 2023, pp. 1-8, doi: 10.1109/Allerton58177.2023.10313487.
Download Paper | Download Slides | Download Bibtex

Age of Gossip in Random and Bipartite Networks

Published in International Symposium on Information Theory (ISIT), 2024

I study the age of information in unbalanced complete bipartite graphs, random regular graphs, and the Erdos-Reyni graph. I prove a technical lemma that the vAoI with equal push rate to neighbors is monotone under adding edges.

Recommended citation: T. J. Maranzatto, "Age of Gossip in Random and Bipartite Networks," 2024 IEEE International Symposium on Information Theory (ISIT), Athens, Greece, 2024, pp. 1173-1178, doi: 10.1109/ISIT57864.2024.10619401.
Download Paper | Download Slides | Download Bibtex

Tree Trace Reconstruction - Reductions to String Trace Reconstruction

Published in International Symposium on Information Theory (ISIT), 2024

I study the sample complexity of tree reconstruction under the TED deletion channel, and show it is the same up to a polynomial factor as the classical string reconstruction problem. I also provide some combinatorial identities for the deletion channel.

Recommended citation: T. J. Maranzatto, "Tree Trace Reconstruction - Reductions to String Trace Reconstruction," 2024 IEEE International Symposium on Information Theory (ISIT), Athens, Greece, 2024, pp. 885-890, doi: 10.1109/ISIT57864.2024.10619306.
Download Paper | Download Slides | Download Bibtex

Age of Gossip with the Push-Pull Protocol

Published in IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025

We study the positive benefit of allowing nodes in a gossip network to actively pull data from their neighbors, instead of passively waiting for data to arrive. Join with Arunabh Srivastava and Sennur Ulukus.

Recommended citation: A. Srivastava, T. J. Maranzatto and S. Ulukus, "Age of Gossip with the Push-Pull Protocol," ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Hyderabad, India, 2025, pp. 1-5, doi: 10.1109/ICASSP49660.2025.10890446.
Download Paper | Download Bibtex

Age of Gossip with Time-Varying Topologies

Published in International Symposium on Information Theory (ISIT), 2025

Studied age in Poissonian gossip networks when the underlying network can vary with time. The switching is governed by a CTMC, and one of the states is required to be the complete graph. Joint with Arunabh Srivastava and Sennur Ulukus.

Recommended citation: A. Srivastava, T. J. Maranzatto and S. Ulukus, "Age of Gossip with Time-Varying Topologies," 2025 IEEE International Symposium on Information Theory (ISIT), Ann Arbor, MI, USA, 2025, pp. 1-6, doi: 10.1109/ISIT63088.2025.11195490.
Download Paper | Download Bibtex

Information Degradation and Misinformation in Gossip Networks

Published in International Symposium on Information Theory (ISIT), 2025

Investigated how data can rapidly lose quality in gossip networks. Joint with Arunabh Srivastava and Sennur Ulukus.

Recommended citation: T. J. Maranzatto, A. Srivastava and S. Ulukus, "Information Degradation and Misinformation in Gossip Networks," 2025 IEEE International Symposium on Information Theory (ISIT), Ann Arbor, MI, USA, 2025, pp. 1-6, doi: 10.1109/ISIT63088.2025.11195622.
Download Paper | Download Slides | Download Bibtex

Private Information Retrieval on Multigraph-Based Replicated Storage

Published in International Symposium on Information Theory (ISIT), 2025

Developed new upper and lower bounds for (multi)graph based private information retrieval. The provided PDF is for the final journal version, which is a superset of the results proved here. Joint with Shreya Meel, Xiangliang Kong, Itzhak Tamo, and Sennur Ulukus

Recommended citation: S. Meel, X. Kong, T. J. Maranzatto, I. Tamo and S. Ulukus, "Private Information Retrieval on Multigraph-Based Replicated Storage," 2025 IEEE International Symposium on Information Theory (ISIT), Ann Arbor, MI, USA, 2025, pp. 1-6, doi: 10.1109/ISIT63088.2025.11195342.
Download Paper | Download Bibtex

Deriving Moments in the Age of Gossip Process from Percolation

Published in IN SUBMISSION, 2025

Submitted ICASSP 2026. We recover stronger versions of the expected AoI recurrences by passing to first-passage percolation. We also remark on some connections to the Eden model in statistical physics. Joint with Sennur Ulukus.

Recommended citation: T. J. Maranzatto, S. Ulukus. "Deriving Moments in the Age of Gossip Process from Percolation." https://arxiv.org/abs/2509.13981
Download Paper | Download Bibtex

Age of Gossip From Connective Properties via First Passage Percolation

Published in IEEE Transactions on Information Theory, 2025

This paper connects AoI in Poisson gossip networks to FPP. A number of open problems are resolved. Joint with Marcus Michelen.

Recommended citation: T. J. Maranzatto and M. Michelen, "Age of Gossip From Connective Properties via First Passage Percolation," in IEEE Transactions on Information Theory, vol. 71, no. 11, pp. 9019-9027, Nov. 2025, doi: 10.1109/TIT.2025.3612162.
Download Paper | Download Slides | Download Bibtex

Information Freshness in Dynamic Gossip Networks

Published in Information Theory Workshop, 2025

This paper finds asymptotic scaling of AoI in a dynamic network switching between two arbitrary topolgies. Joint with Arunabh Srivastava and Sennur Ulukus.

Recommended citation: A. Srivastava, T. J. Maranzatto, S. Ulukus. "Information Freshness in Dynamic Gossip Networks," Accepted ITW 2025. https://arxiv.org/abs/2504.18504
Download Paper | Download Slides | Download Bibtex

New Capacity Bounds for PIR on Graph and Multigraph-Based Replicated Storage

Published in IEEE Transactions on Information Theory (Accepted in Revision), 2025

Developed new upper and lower bounds for (Multi) graph based private information retrieval. Builds on our earlier work published in ISIT. Joint with Shreya Meel, Xiangliang Kong, Itzhak Tamo, and Sennur Ulukus

Recommended citation: S. Meel, X. Kong, T. J. Maranzatto, I. Tamo, and S. Ulukus. "New Capacity Bounds for PIR on Graph and Multigraph-Based Replicated Storage," Accepted under revision, IEEE Transactions on Information Theory 2025.
Download Paper | Download Bibtex

talks

teaching

Format for Future Courses

Undergraduate course, University 1, Department, 2014

This is a description of a course, using markdown of course.