Dr. S. SangeethaAI & Privacy · PSG iTech
Research library

Thirty papers, five questions

Thirty papers from 2012 onwards. Each has its own page so search engines can index it, with authors, venue, citation counts, a copy-ready citation, and links to the publisher, the dataset or the code where they exist.

Differential privacy and private recommenders · 7Healthcare and federated privacy · 2Location privacy and blockchain · 5Applied AI with student teams · 10Foundations: data mining and big data privacy · 6
30papers
110citations
5h-index
2i10-index
Papers published per year
View as a table
YearPapers
20122
20181
20191
20201
20217
20223
20237
20243
20254

Record as of 2026-09-10, from Google Scholar. Refreshed weekly from OpenAlex and Semantic Scholar.

Differential privacy and private recommenders

Can a recommender learn what you like without learning who you are?

7 papers · 20 citations

The core of my doctoral work. Differential privacy adds calibrated noise so that no single person can be reverse-engineered out of a model. The catch is that noise costs accuracy, and these papers measure that trade-off and find where a recommender stays useful.

New to this? Differential privacy, in five minutes

Healthcare and federated privacy

Can a hospital share what it learned without sharing its patients?

2 papers · 20 citations

Health records are the most sensitive data there is, and anonymised releases have been re-identified before. This line of work releases the trained model instead of the data, and looks at federated learning, where the data never leaves the hospital.

New to this? Sharing a model instead of the patients

Location privacy and blockchain

Can a system use where you are without recording who you are?

5 papers · 8 citations

Crowd-sensing and location-based services leak position data by design. These papers combine cryptography and distributed ledgers to keep a contribution useful while keeping the contributor anonymous.

New to this? Using where you are without recording who you are

Applied AI with student teams

What can a model see that a person cannot?

10 papers · 23 citations

Projects I have run with undergraduate teams: infection detection in mushroom farms, solar output under cloud cover, fake review and fake news detection, gesture recognition, text summarisation, and carbon-efficient inference in large language models.

New to this? What a student team can actually build in a semester

Foundations: data mining and big data privacy

Which privacy methods survive at scale?

6 papers · 40 citations

My earliest thread, running from association-rule hiding in 2012 through the Springer handbook review that is still my most cited work. This is where the privacy question starts for me, five years before the doctorate.

New to this? Which privacy methods survive at scale

Getting the full text

Links point at the publisher where I hold the record. Accepted manuscripts will be hosted here for the papers whose publishers permit it, which covers most of this list: Springer, Taylor & Francis and IEEE all allow the author’s accepted version on a personal site. Open copies are read and cited substantially more often than links behind a paywall.

5 recurring co-authors appear across this work, alongside the student teams credited on the applied projects.