vix.ing · top · new · best · stats · spec

A One-Layer Decoder-Only Transformer is a Two-Layer RNN: With an Application to Certified Robustness

2024/05/27 by Zhang, Yuhao, Albarghouthi, Aws, D'Antoni, Loris
#Computation and Language (cs.CL) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2405.17361

Abstract

This paper reveals a key insight that a one-layer decoder-only Transformer is equivalent to a two-layer Recurrent Neural Network (RNN). Building on this insight, we propose ARC-Tran, a novel approach for verifying the robustness of decoder-only Transformers against arbitrary perturbation spaces. Compared to ARC-Tran, current robustness verification techniques are limited either to specific and length-preserving perturbations like word substitutions or to recursive models like LSTMs. ARC-Tran addresses these limitations by meticulously managing position encoding to prevent mismatches and by utilizing our key insight to achieve precise and scalable verification. Our evaluation shows that ARC-Tran (1) trains models more robust to arbitrary perturbation spaces than those produced by existing techniques and (2) shows high certification accuracy of the resulting models.

Related