Adversarial interpolation backdoor detection
One or more computer processors determine a tolerance value, and a norm value
associated with an untrusted model and an adversarial training method. The one or more …
associated with an untrusted model and an adversarial training method. The one or more …
Multi-layer perceptron-based computer vision neural networks
I Tolstikhin, NMT Houlsby, A Kolesnikov… - US Patent App. 17 …, 2022 - Google Patents
BACKGROUND [0002] This specification relates to neural networks that process images to
perform computer vision tasks.[0003] Neural networks are machine learning models that …
perform computer vision tasks.[0003] Neural networks are machine learning models that …
System and method for utilizing perturbation in a multimodal environment
FJC CONDESSA, WY Lin, K Yang… - US Patent App. 17 …, 2022 - Google Patents
A computer-implemented method for training a machine learning network. The method may
include receiving an input data, selecting one or more batch samples from the input data …
include receiving an input data, selecting one or more batch samples from the input data …
Apparatus, method, and computer-readable medium for robust response to adversarial perturbations using hyperdimensional vectors
N Srinivasa - US Patent 11,854,253, 2023 - Google Patents
US11854253B2 - Apparatus, method, and computer-readable medium for robust response to
adversarial perturbations using hyperdimensional vectors - Google Patents US11854253B2 …
adversarial perturbations using hyperdimensional vectors - Google Patents US11854253B2 …