Adversarial Robustness of Deep State Space Models for Forecasting

Sribalaji C. Anand, George J. Pappas · 2026

Read Adversarial Robustness of Deep State Space Models for Forecasting by Sribalaji C. Anand, George J. Pappas (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-03427.txt.

Description

Anand and Pappas study adversarial robustness of deep state-space forecasters using control theory to focus on the Spacetime model. They prove the decoder-only architecture can represent the optimal Kalman predictor for autoregressive data, a property no other state-space model holds. They also show model-free attacks needing no gradients cause at least 33 percent higher error than projected gradient descent with tiny steps.

Catalog record

CREATORSribalaji C. Anand, George J. Pappas
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43551
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.03427

Browse the full library · About the Tower of Babel