hey there! I'm, Anshu.

Professional Me

Privacy Researcher | AI Researcher | Software Engineer

Anshu Singh is a privacy researcher at GovTech (Government Technology Agency), Singapore. Her current interests lie in privacy threat modelling, privacy by design, and privacy-enhancing technologies. She graduated from the National University of Singapore (NUS) with Masters of Computing degree, specialization in Artificial Intelligence (AI). She also spent 2 years in AI research particularly in computer vision and privacy at NUS Centre for Research in Privacy Technologies (N-CRiPT) under the supervision of Prof. Mohan Kankanhalli and Dr. Shaojing Fan . Her research interests also include privacy-preserving machine learning. Reach out at firstname.lastname173 at gmail.com


Human Attributes Prediction Under Privacy-preserving Conditions
Anshu Singh, Shaojing Fan, Mohan Kankanhalli
ACM International Conference on Multimedia 2021
Abstract: Human attributes prediction in visual media is a well-researched topic with a major focus on human faces. However, face images are often of high privacy concern as they can reveal an individual's identity. How to balance this trade-off between privacy and utility is a key problem among researchers and practitioners. In this study, we make one of the first attempts to investigate the human attributes (emotion, age, and gender) prediction under the different de-identification (eyes, lower-face, face, and head obfuscation) privacy scenarios. We first constructed the Diversity in People and Context Dataset (DPaC). We then performed a human study with eye-tracking on how humans recognize facial attributes without the presence of face and context. Results show that in an image, situational context is informative of a target's attributes. Motivated by our human study, we proposed a multi-tasking deep learning model - Context-Guided Human Attributes Prediction (CHAPNet), for human attributes prediction under privacy-preserving conditions. Extensive experiments on DPaC and three commonly used benchmark datasets demonstrate the superiority of CHAPNet in leveraging the situational context for a better interpretation of a target's attributes without the full presence of the target's face. Our research demonstrates the feasibility of visual analytics under de-identification for privacy.
The work had been selected and presented at various venues — Content-Based Multimedia Indexing Lille (France), Indian Institute of Science Bangalore (India), SGInnovate (Singapore), and Singapore Cybersecurity Consortium (Singapore).
[Project Webpage] [Code] [Poster]


Conducted a workshop on data privacy (focus on differential privacy) at STACK Developer Conference 2022.
Gave an invited talk at the Singapore Cybersecurity Consortium; spoke on privacy-preserving human attributes prediction using AI [Video]
Presented the published paper virtually at the main conference held in Chengdu, China [Video]


Solutions to catalysing PETs adoption at Privacy-Enhancing Technology Summit Asia-Pacific
Nov 15, 2022
I had a great time discussing privacy-enhancing technologies' (PETs) topics with global experts in the roundtable discussions at Privacy-Enhancing Technology Summit Asia-Pacific today. 😊 One of the major concerns was the adoption of PETs, and I have succinctly jotted down the gripping solutions brought to the table
2022 USENIX Conference on Privacy Engineering Practice and Respect Key Takeaways.
Sep 12, 2022
The talks at PEPR'22 gave encouragement that a new "privacy protection industry" is emerging, with new privacy technologies being developed and discoveries being made to assist users in gaining more control and transparency over their personal data. Such advancements unquestionably reflect a step toward making privacy a "human-value-focused" need rather than a "compliance-only" mandate.....
Why Should You Care About Facial Privacy?
Apr 21, 2021
In multiple research domains, a human face is a well-studied object from visual content since it is a source of rich information, most notably the identity of an individual. Identity revelation from visual data can make one vulnerable to leakage of personal and sensitive information (e.g., sexual orientation [1], health condition [2], religious beliefs [3]), mental and social harassment [4], and much more.....
The Role of Artificial Intelligence in Future Technology
May 13, 2019
We stand at the brink of a new economy: a digital economy, where the global workforce will be deeply impacted by Artificial Intelligence(AI) and globally more developments are expected tomorrow.....

Work Experience

Privacy Researcher at GovTech
- Working on a whitespace project to benchmark differential privacy (DP) libraries (such as OpenDP library, Tumult Core/Analytics, and Google's DP libraries)
- Building an anonymization tool to be adopted by the whole of government (agencies such as in healthcare and education domain); engaging with policymakers to operationalize rules and regulations into the tool
- Researching, documenting, and guiding the team on the implementation of anonymization techniques (such as pseudonymization, K-anonymity, L-diversity)
- Actively involved in discussions and advising government agencies on privacy-enhancing technologies (such as DP) to solve privacy-preserving data publishing and processing
Research Assistant at NUS's N-CRiPT
- Geocoded Tweets' Insights Dashboard - built an interactive, configurable and generic dashboard to analyse country-specific or global influential users, local and global news and events, public perception, user communities and so on. These insights can help the regulators and decision-makers take appropriate actions
- Did cutting edge research and published a paper in the domain of computer vision for facial analytics in privacy settings; did comprehensive literature survey, exhaustive geo-diverse data collection and wrangling, experimentation and statistical testing, explainable deep learning modeling and hyperparameter optimization, evaluation, and visualizations
Software Engineer at Development Bank of Singapore
- Revamped a high priority project, Current Account Savings Account (CASA), used by millions to create a DBS or POSB account
- Led and won a hackathon and was featured on DBS India’s website
Software Engineer at HashedIn by Deloitte
- Developed a SaaS application to help teams run effective meetings with clearly defined agendas, action items, collaborative notes taking, and so on; clients reported productivity increase by 25%
- Ideated and developed an MVP product to track a user's activities across GitHub, Google drive, and Jira