SILENT-MODE SIGNATURE TESTING IN ANTI-MALWARE PROCESSING

Patent №

US 8,819,835

Granted

2014-08-26

Filed 2013

Owner

KASPERSKY LAB, ZAO

Lab

AI components

3

ml · kr · hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

13740775

Method and computer program product for signature testing used in anti-malware processing. Silent signatures, after being tested, are not updated into a white list and are sent directly to users instead. If the silent signature coincides with malware signature, a user is not informed. A checksum (e.g., hash value) of a suspected file is sent to a server, where statistics are kept and analyzed. Based on collected false positive statistics of the silent-signature, the silent-signature is either valid or invalid. Use of the silent signatures provides for effective signature testing and reduces response time to new malware-related threats. The silent signature method is used for turning off a signature upon first false positive occurrence. Use of silent signatures allows improving heuristic algorithms for detection of unknown malware.

AI classification

AI hardware0.99
Knowledge representation0.91
Machine learning0.69
Planning0.29
Vision0.02
Natural language0.01
Evolutionary computation0.00
Speech0.00

Ownership

KASPERSKY LAB, ZAO

assignment · 296260030

Assignors

NAZAROV, DENIS A.

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC