← Search

Hanna Foerster

2 accepted papers

2025

Hardware and Software Platform Inference

ICML 2025poster

It is now a common business practice to buy access to large language model (LLM) inference rather than self-host, because of significant upfront hardware infrastructure and energy costs. However, as a buyer, there is no mechanism to verify the authenticity of the advertised service including the ser…

Cited by 0SourcePDFScholar
2024

Beyond Slow Signs in High-fidelity Model Extraction

NeurIPS 2024poster

Deep neural networks, costly to train and rich in intellectual property value, are increasingly threatened by model extraction attacks that compromise their confiden- tiality. Previous attacks have succeeded in reverse-engineering model parameters up to a precision of float64 for models trained on r…