Search
Patexia Contest

CONTEST

Competed
Prize
$5,000
DEADLINE
This contest is closed.
Winner

Paton Paton

technology
Winner

Yogesh Bansal

technology
Runner Up

Priyal

technology
Runner Up

Biplab Pal

technology
Runner Up

NEHA JAI

technology
Runner Up

Arp Jai

technology
Runner Up

Abha maithil

technology
Runner Up

Ankit Goyal

technology
Runner Up

Roopa V

technology
Runner Up

Bhaskar_K

technology
Runner Up

Kasthuri R

technology
Runner Up
Runner Up

kanak

technology
Runner Up

Ragupathy S P

technology

Problem

This contest is for a prior art search. We are ideally looking for a reference that describes a process for assessing whether an executable file is harmful by analyzing metadata and other characteristics of the file that can be determined without unpacking or executing it. Metadata could be file size, copyright information, creation date, etc. This process involves generating a feature vector from these characteristics, and potentially turning certain features on or off based on whether the metadata or characteristics fall within predefined ranges. The generated feature vector is then processed using a support vector machine to assess the file's risk of harm. The reference should also mention the possibility of determining if the file is known to be malicious, benign, or unwanted. The file size may be used to determine whether certain features in the feature vector are turned on or off, and the reference may define either a minimum size for harmful files or a maximum size for benign files.

Have a question about this contest? Ask a Question

Questions

#QuestionValue
1Is the reference either a U.S. Patent filed, a foreign patent published, or a non-patent document published or otherwise publicly available before June 2015? T/F
2Does the reference describe assessing whether an executable file is harmful (e.g., detecting malware) based on metadata or other characteristics of the file that are ascertainable without unpacking or executing the file (e.g., file size, copyright information, creation date, entropy, etc.)? 10
3Does the reference describe generating a feature vector from the metadata or other file characteristics? 10
4Are any features in the feature vector turned on or off based on whether the metadata or other file characteristics fall with a predefined range (e.g., (1) below a specific value, (2) between two specific values or (3) above a specific value)? 15
5Is a support vector machine used to process the generated feature vector to assess the file’s risk of harm? 10
6Does the reference describe determining if the file is known malicious, known benign, and/or known unwanted? 10
7Is file size used to determine whether any features in the feature vector are turned on or off? 20
8Does the reference define (1) a minimum size for harmful files so any file smaller than that is defined as safe/benign or (2) a file size above which files are assumed to be safe/benign? 25

Additional Notes

Prior Art Search

Prior Art Search

This is a Prior Art Search contest aimed at determining if a patent idea was known and publicly available before a patent was filed.

  • +5 bonus points will be awarded for non-patent literature such as user's manual or peer-reviewed papers
  • +5 bonus points will be awarded for foreign language references.

This contest will close on Sunday, February 5th, 2023, 11:59 pm

Please review the Contest Rules. For more information on how to submit to this Contest type, please read the Intro to Prior Art Search page.

Menu