Presumption of Innocence (IUPG), Adversarial-Resistant, and False-Positive-Resistant Deep Learning Models

The Presumption of Innocence (IUPG) framework addresses the vulnerabilities of existing malware detection systems by training deep neural networks to resist adversarial attacks and reduce false positives, enhancing the accuracy of malware classification.

JP7763301B2Active Publication Date: 2025-10-31PALO ALTO NETWORKS INC
4 Cites 0 Cited by

Patent Information

Application Number
JP2024120874
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2021-05-26
Filing Date
2024-07-26
Publication Date
2025-10-31
Estimated Expiration
2041-06-03

AI Technical Summary

Technical Problem

Existing malware detection systems are vulnerable to adversarial attacks and produce high false-positive rates, failing to effectively distinguish between benign and malicious applications.

Method used

The development of adversary-resistant and false-positive-resistant deep learning models, known as the Presumption of Innocence (IUPG) framework, which utilizes a hybrid discriminative and generative loss function to train deep neural networks, learning unique input space prototypes and maximizing the distance between target and off-target classes.

Benefits of technology

The IUPG framework enhances malware detection by reducing false positives and improving the ability to identify adversarial attacks, ensuring accurate classification of applications.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

To provide a system and method including innocent until proven guilty (IUPG) model for building and using adversary resistant and false positive resistant deep learning models.SOLUTION: A method of performing static analysis of a sample includes: storing a set comprising one or more IUPG models for static analysis of a sample; performing static analysis of a content associated with the sample, performing the static analysis of the model including using at least one stored IUPG model. The method further includes: determining that the sample is malicious on the basis of at least partially, the static analysis of the content associated with the sample; and in response to determining that the sample is malicious, performing an action on the basis of a security policy.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art