WO2024261525 - SECURE AND PRIVACY PRESERVING QUERYING OF LARGE LANGUAGE MODELS
National phase entry is expected:
Publication Number
WO/2024/261525
Publication Date
26.12.2024
International Application No.
PCT/IB2023/058212
International Filing Date
16.08.2023
Title **
[English]
SECURE AND PRIVACY PRESERVING QUERYING OF LARGE LANGUAGE MODELS
[French]
INTERROGATION DE GRANDS MODÈLES DE LANGAGE SÉCURISÉE ET PRÉSERVANT LA CONFIDENTIALITÉ
Applicants **
NEC LABORATORIES EUROPE GMBH
Inventors
GONZALEZ SANCHEZ, Roberto
BIFULCO, Roberto
SIRACUSANO, Giuseppe
SANVITO, Davide
Priority Data
63/522,180
21.06.2023
US
Application details
| Total Number of Claims/PCT | * |
| Number of Independent Claims | * |
| Number of Priorities | * |
| Number of Multi-Dependent Claims | * |
| Number of Drawings | * |
| Pages for Publication | * |
| Number of Pages with Drawings | * |
| Pages of Specification | * |
| * | |
| Number of Office Actions | * |
| * | |
International Searching Authority |
EPO
* |
| Recordal of a Change of the Applicant's Name/Address |
Change of Applicant's Name and Address
* |
| Type of Assignment |
The Standard Agent's Assignment
* |
| Applicant's Legal Status |
Legal Entity
* |
| * | |
| * | |
| * | |
| * | |
| * | |
| Entry into National Phase under |
Chapter I
* |
| Patent Delivery |
Send the Letters Patent by Courier
* |
| Translation |
|
* The data is based on automatic recognition. Please verify and amend if necessary.
** IP-Coster compiles data from publicly available sources. If this data includes your personal information, you can contact us to request its removal.
Quotation for National Phase entry
| Country | Stages | Total | |
|---|---|---|---|
| China | Filing, Examination, Granting | 2027 | |
| EPO | Filing, Examination, Granting | 8221 | |
| Japan | Filing, Examination, Granting | 2201 | |
| South Korea | Filing, Examination, Granting | 2023 | |
| USA | Filing, Examination, Granting | 4740 |
Total:
19,212
The term for entry into the National Phase has expired. This quotation is for informational purposes only
Abstract[English]
A computer-implemented method operates a large language model (LLM) in a secure and privacy preserving manner. The method includes: identifying an entity in a query, the query being configured to be processed by the LLM; generating a modified query by substituting the identified entity with a semantically similar entity; and submitting the modified query to the LLM. The present invention can be used in a variety of applications including, but not limited to, several anticipated use cases in customer service, cybersecurity, finance, and medical and healthcare.[French]
Un procédé mis en oeuvre par ordinateur exploite un grand modèle de langage (LLM) d'une manière sécurisée et qui préserve la confidentialité. Le procédé consiste à : identifier une entité dans une requête, la requête étant configurée pour être traitée par le LLM ; générer une requête modifiée en substituant l'entité identifiée par une entité sémantiquement similaire ; et soumettre la requête modifiée au LLM. La présente invention peut être utilisée dans diverses applications comprenant, de manière non limitatitive, plusieurs cas anticipés d'utilisation de service à la clientèle, de cybersécurité, de la finance ainsi que des soins médicaux et de santé.