Scholar
Nathalie Baracaldo
Google Scholar ID: 3ACndBYAAAAJ
IBM Almaden Research Center, Senior Research Staff Member, Ph.D.
Citations & Impact
Citations
6,617
H-index
28
i10-index
46
Publications
20
Co-authors
11
list available
Publications
Resume (English only)
Academic Achievements
- Named IBM Master Inventor (2020)
- Received IBM Corporate Technical Recognition (2021) for contributions to Trusted AI
- Associate Editor of IEEE Transactions on Service Computing
- Guest Editor for IEEE S&P Magazine Special Issue on ML Security and Privacy
- Publication Chair (S&P 2022) and Poster Chair (S&P 2021) at IEEE Symposium on Security and Privacy
- Organized federated learning workshops: MLSys CrossFL 2022, ICML-FL 2021, ICML-FL 2020
- Advisor for 2021 Workshop on Security, Privacy, and Ethics in Health and Biomedical Research
- Published numerous papers (available via Google Scholar)
- Holds multiple patents (available via Google Patents)
Research Experience
- Leads the AI Security and Privacy Solutions team at IBM Almaden Research Center
- Principal Investigator for DARPA’s GARD (Guaranteeing AI Robustness Against Deception) program
- Contributor to the Adversarial Robustness 360 Toolbox (ART)
- Designed secure systems in cloud computing, Platform as a Service (PaaS), secure data sharing, and Internet of Things
- Developed scalable systems for monitoring, performance management, and SLA management in cloud environments
Background
- Manager of AI Security and Privacy Solutions and Research Staff Member at IBM Almaden Research Center
- Research interests lie at the intersection of information security, privacy, and trust
- Passionate about developing highly accurate machine learning solutions that resist adversarial attacks and protect data privacy
- Team focuses on federated learning (training models without direct access to raw data) and adversarial machine learning (designing defenses against attacks on ML pipelines)
- Contributor to IBM's Trusted AI initiative
Miscellany
- Invited tutorial speaker at IEEE TPS 2021
- Keynote speaker at EMISA 2021
- Featured guest on Data Science Podcast discussing federated learning
- Authored blog posts on 'Accountable Federated Learning' (in English and German)
- Presented at NeurIPS 2020 workshop 'Beyond AutoML: Scaling & Automating AI'
- Interviewed on YouTube (DC_THURS) about federated learning and adversarial ML
