Full-Privacy Secured Search Engine Empowered by Efficient Genome-Mapping Algorithms.

Area of Science:

• Computer Science
• Information Retrieval
• Bioinformatics

Background:

• Traditional search engines rely on keyword queries, exposing user data and search history.
• Existing search methods lack robust privacy protections, often utilizing user data for targeted advertising or database expansion.

Purpose of the Study:

• To introduce a novel search and plagiarism detection system, Sapiens Aperio Veritas Engine (S.A.V.E.), that prioritizes user privacy.
• To enable users to search the internet and perform plagiarism checks using full articles as queries without revealing content.

Main Methods:

• Developed S.A.V.E., employing an encoding scheme and FM-index search adapted from next-generation genome sequencing.
• Transformed query words into a pseudo-biological sequence (PBS) using 12 "amino acids" on the user's device.
• Utilized cloud service for plagiarism checks by submitting locally generated PBSs against a database of Wikipedia and Open Access journals.

Main Results:

• S.A.V.E. accurately detects identical content with less than 0.8% false positives for PBSs longer than 12 "amino acids".
• The system achieves genome-mapping speeds comparable to Bowtie and is over 5 orders of magnitude faster than BLAST.
• S.A.V.E. offers both standard and private search modes, enhancing user confidentiality.

Conclusions:

• S.A.V.E. presents a revolutionary, privacy-first search and plagiarism check system.
• The method sets a precedent for future search engines designed with user confidentiality as a priority.
• The system is accessible at https://dyn.life.nthu.edu.tw/SAVE/.