Feb. 9, 2024, 5:43 a.m. | Jamie Hayes Ilia Shumailov Itay Yona

cs.LG updates on arXiv.org arxiv.org

Mixture of Experts (MoE) has become a key ingredient for scaling large foundation models while keeping inference costs steady. We show that expert routing strategies that have cross-batch dependencies are vulnerable to attacks. Malicious queries can be sent to a model and can affect a model's output on other benign queries if they are grouped in the same batch. We demonstrate this via a proof-of-concept attack in a toy experimental setting.

attacks become costs cs.cr cs.lg dependencies expert experts foundation inference inference costs key mixture of experts moe overflow routing scaling show strategies vulnerable

Research Scholar (Technical Research)

@ Centre for the Governance of AI | Hybrid; Oxford, UK

HPC Engineer (x/f/m) - DACH

@ Meshcapade GmbH | Remote, Germany

Senior Analytics Engineer (Retail)

@ Lightspeed Commerce | Toronto, Ontario, Canada

Data Scientist II, BIA GPS India Operations

@ Bristol Myers Squibb | Hyderabad

Analytics Engineer

@ Bestpass | Remote

Senior Analyst - Data Management

@ Marsh McLennan | Mumbai - Hiranandani