Sotabase

Career

· Invited Lecturer, Ilia State University2019–
· Associate Professor, European University (Tbilisi, Georgia) (Current)2018–
· Invited Professor, European University (Current)2015–
· Researcher, Ilia State University2013–
· Scientific Researcher, Ilia State University2013–
· Scientific Researcher, E. Andronikashvili Institute of Physics (Department of Physics of Biological Systems) (Current)2009–
· Doctor of Philosophy - PhD, Biophysics, Ilia State University2007–2014
· Microbiologist - Analyser, LTD "Multitest"2006–2008
· Master's degree, Biology/Biological Sciences, General, Tbilisi State University2001–2006

Publications (98)

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context
arXiv.org · 2024
3,137
cited
Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
arXiv.org · 2025
1,636
cited
AI models collapse when trained on recursively generated data
Nature · 2024
558
cited
The Curse of Recursion: Training on Generated Data Makes Models Forget
arXiv.org · 2023
413
cited
When the Curious Abandon Honesty: Federated Learning Is Not Private
European Symposium on Security and Privacy · 2021
226
cited
On the Necessity of Auditable Algorithmic Definitions for Machine Unlearning
USENIX Security Symposium · 2021
204
cited
Scalable watermarking for identifying large language model outputs
Nature · 2024
178
cited
Sponge Examples: Energy-Latency Attacks on Neural Networks
European Symposium on Security and Privacy · 2020
161
cited
Bad Characters: Imperceptible NLP Attacks
IEEE Symposium on Security and Privacy · 2021
128
cited
Manipulating SGD with Data Ordering Attacks
Neural Information Processing Systems · 2021
109
cited
Defeating Prompt Injections by Design
arXiv.org · 2025
79
cited
LLM Censorship: A Machine Learning Challenge or a Computer Security Problem?
arXiv.org · 2023
72
cited
Inexact Unlearning Needs More Careful Evaluations to Avoid a False Sense of Privacy
2025 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) · 2024
65
cited
Fairness Feedback Loops: Training on Synthetic Data Amplifies Bias
Conference on Fairness, Accountability and Transparency · 2024
54
cited
Turning Up the Dial: the Evolution of a Cybercrime Market Through Set-up, Stable, and Covid-19 Eras
ACM/SIGCOMM Internet Measurement Conference · 2020
53
cited
UnUnlearning: Unlearning is not sufficient for content regulation in advanced generative AI
arXiv.org · 2024
50
cited
To compress or not to compress: Understanding the Interactions between Adversarial Attacks and Neural Network Compression
USENIX workshop on Tackling computer systems problems with machine learning techniques · 2018
44
cited
Hearing your touch: A new acoustic side channel on smartphones
arXiv.org · 2019
37
cited
Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy, Research, and Practice
arXiv.org · 2024
36
cited
Blackbox Attacks on Reinforcement Learning Agents Using Approximated Temporal Information
2020 50th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W) · 2019
35
cited
Sotabase
Ilia Shumailov | Researcher Profile | Sotabase | Sotabase