Brian L. Mark

Omar Alotaibi, Brian L. Mark, Mohammad Reza Fasihi

We present a novel filtering algorithm that employs Bayesian transfer learning to address the challenges posed by mismatched intensity of the noise in a pair of sensors, each of which tracks an object using a nonlinear dynamic system model. In this setting, the primary sensor experiences a higher noise intensity in tracking the object than the source sensor. To improve the estimation accuracy of the primary sensor, we propose a framework that integrates Bayesian transfer learning into an Unscented Kalman Filter (UKF) and a Cubature Kalman Filter (CKF). In this approach, the parameters of the predicted observations in the source sensor are transferred to the primary sensor and used as an additional prior in the filtering process. Our simulation results show that the transfer learning approach significantly outperforms the conventional isolated UKF and CKF. Comparisons to a form of measurement vector fusion are also presented.

Massieh Kordi Boroujeny, Ya Jiang, Kai Zeng et al.

Methods for watermarking large language models have been proposed that distinguish AI-generated text from human-generated text by slightly altering the model output distribution, but they also distort the quality of the text, exposing the watermark to adversarial detection. More recently, distortion-free watermarking methods were proposed that require a secret key to detect the watermark. The prior methods generally embed zero-bit watermarks that do not provide additional information beyond tagging a text as being AI-generated. We extend an existing zero-bit distortion-free watermarking method by embedding multiple bits of meta-information as part of the watermark. We also develop a computationally efficient decoder that extracts the embedded information from the watermark with low bit error rate.

Ya Jiang, Chuxiong Wu, Massieh Kordi Boroujeny et al.

Watermarking for large language models (LLMs) offers a promising approach to identifying AI-generated text. Existing approaches, however, either compromise the distribution of original generated text by LLMs or are limited to embedding zero-bit information that only allows for watermark detection but ignores identification. We present StealthInk, a stealthy multi-bit watermarking scheme that preserves the original text distribution while enabling the embedding of provenance data, such as userID, TimeStamp, and modelID, within LLM-generated text. This enhances fast traceability without requiring access to the language model's API or prompts. We derive a lower bound on the number of tokens necessary for watermark detection at a fixed equal error rate, which provides insights on how to enhance the capacity. Comprehensive empirical evaluations across diverse tasks highlight the stealthiness, detectability, and resilience of StealthInk, establishing it as an effective solution for LLM watermarking applications.