Ilia Shumailov

机构：Google DeepMind (United Kingdom) · ORCID：0000-0003-3100-0727

发表论文 117 篇 · 总被引 1777 次 · h-index 17

代表论文

• AI models collapse when trained on recursively generated data (2024 · Nature · 被引 749)
• Scalable watermarking for identifying large language model outputs (2024 · Nature · 被引 121)
• Fairness Feedback Loops: Training on Synthetic Data Amplifies Bias (2024 · 被引 26)
• Inexact Unlearning Needs More Careful Evaluations to Avoid a False Sense of Privacy (2025 · 被引 11)
• Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy, Research, and Practice (2025 · SSRN Electronic Journal · 被引 11)
• Measuring memorization in language models via probabilistic extraction (2025 · 被引 11)