IP Library Granted Patent US 12,238,213
Granted Patent B1
US 12,238,213 · App. 18/882,398 · Granted Feb 25, 2025

Methods and systems for verifying a worker agent

Inventors: Mohammad Naanaa (Hillsborough, CA); Volodymyr Panchenko (Hillsborough, CA); Manav Mehra (Toronto, CA); Ricardo Fornari (San Jose, CA)
Assignee: Portal AI Inc.
H04L9/32
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,238,213
App. No.
18/882,398
Granted
Feb 25, 2025
Kind
B1
Abstract

A method for verifying a worker agent includes receiving, by a core node, from a worker agent, a capability description describing a plurality of tasks and, for each of the plurality of tasks, (i) at least one parameter of the task and (ii) an outcome expected to be produced by performing the task. The method includes generating, based on the capability description, a plurality of request-output pairs, each representing a particular request and a corresponding baseline output expected to be produced upon processing the request. The core node receives, from the worker agent, a plurality of outputs, each of the plurality of outputs generated by the worker agent and corresponding to one of the plurality of request-output pairs. The core node compares the plurality of baseline outputs to the plurality of actual outputs to produce comparison output and determines whether to approve the worker agent based on the comparison output.

Claims (30)

1. A method, performed by at least one computer processor executing computer program instructions stored on at least one non-transitory computer-readable medium, for verifying a worker agent, the method comprising:

(A) receiving, by a core node, from a worker agent, a capability description describing a plurality of tasks and, for each of the plurality of tasks, (i) at least one parameter of the task and (ii) an outcome expected to be produced by performing the task;

(B) generating, based on the capability description, a plurality of request-output pairs, each representing a particular request and a corresponding baseline output expected to be produced when the request is processed;

(C) receiving, by the core node, from the worker agent, a plurality of outputs, each of the plurality of outputs generated by the worker agent and corresponding to one of the plurality of request-output pairs;

(D) comparing, by the core node, the plurality of baseline outputs to the plurality of actual outputs to produce comparison output; and

(E) determining whether to approve the worker agent based on the comparison output.

2. The method of claim 1 , wherein (A) further comprises transmitting the predefined set of user requests subsequent to verifying, by the core node, from the worker agent, a response to a cryptographic challenge.

3. The method of claim 1 , wherein (D) further comprises executing a check of accuracy to determine whether the received plurality of outputs satisfies a threshold level of accuracy identified in the received capability description.

4. The method of claim 1 , wherein (D) further comprises executing a check of consistency to determine that the worker agent generates a second plurality of outputs that is substantially consistent with the plurality of outputs received in (C).

5. The method of claim 1 , wherein (D) further comprises executing a response time assessment to determine whether an amount of time lapsed between generating the plurality of request-output pairs and receiving the plurality of outputs satisfies a threshold amount of time identified in the received capability description.

6. The method of claim 1 , wherein (D) further comprises executing a response time assessment to determine whether an amount of utilization of a resource by the worker agent between generating the plurality of request-output pairs and receiving the plurality of outputs satisfies a threshold amount of resource utilization identified in the received capability description.

7. The method of claim 1 further comprising modifying a directory of approved agents to include a worker agent identifier associated with the worker agent.

8. The method of claim 1 further comprising modifying a directory of approved agents to include a cryptographic public key associated with the worker agent.

9. The method of claim 1 further comprises determining, by the core node, to provisionally approve the worker agent.

10. The method of claim 9 further comprises transmitting, by the core node, to the worker agent, an identification of a period of time for which the worker agent is provisionally approved.

11. A system comprising at least one non-transitory computer-readable medium having computer program instructions stored thereon, the computer program instructions being executable by at least one computer processor to perform a method for verifying a worker agent, the method comprising:

(A) receiving, by a core node, from a worker agent, a capability description describing a plurality of tasks and, for each of the plurality of tasks, (i) at least one parameter of the task and (ii) an outcome expected to be produced by performing the task;

(B) generating, based on the capability description, a plurality of request-output pairs, each representing a particular request and a corresponding baseline output expected to be produced when the request is processed;

(C) receiving, by the core node, from the worker agent, a plurality of outputs, each of the plurality of outputs generated by the worker agent and corresponding to one of the plurality of request-output pairs;

(D) comparing, by the core node, the plurality of baseline outputs to the plurality of actual outputs to produce comparison output; and

(E) determining whether to approve the worker agent based on the comparison output.

12. The system of claim 11 , wherein (A) further comprises transmitting the predefined set of user requests subsequent to verifying, by the core node, from the worker agent, a response to a cryptographic challenge.

13. The system of claim 1 , wherein (D) further comprises executing a check of accuracy to determine whether the received plurality of outputs satisfies a threshold level of accuracy identified in the received capability description.

14. The system of claim 11 , wherein (D) further comprises executing a check of consistency to determine that the worker agent generates a second plurality of outputs that is substantially consistent with the plurality of outputs received in (C).

15. The system of claim 11 , wherein (D) further comprises executing a response time assessment to determine whether an amount of time lapsed between generating the plurality of request-output pairs and receiving the plurality of outputs satisfies a threshold amount of time identified in the received capability description.

16. The system of claim 11 , wherein (D) further comprises executing a response time assessment to determine whether an amount of utilization of a resource by the worker agent between generating the plurality of request-output pairs and receiving the plurality of outputs satisfies a threshold amount of resource utilization identified in the received capability description.

17. The system of claim 11 further comprising modifying a directory of approved agents to include a worker agent identifier associated with the worker agent.

18. The system of claim 11 further comprising modifying a directory of approved agents to include a cryptographic public key associated with the worker agent.

19. The system of claim 11 further comprises determining, by the core node, to provisionally approve the worker agent.

20. The system of claim 19 further comprises transmitting, by the core node, to the worker agent, an identification of a period of time for which the worker agent is provisionally approved.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 1, 2024
From: NAANAA, MOHAMMAD; PANCHENKO, VOLODYMYR; MEHRA, MANAV; FORNARI, RICARDO
To: PORTAL AI INC.
Reel/Frame 069106/0527 →
Continuity (1)
Provisional Application 63537979 · Sep 12, 2023
References Cited (87)
US 8856196B2 · Yamada · 2014 [cited by applicant]
US 8856335B1 · Yadwadkar · 2014 [cited by examiner]
US 10042636B1 · Srivastava · 2018 [cited by applicant]
US 10042886B2 · Saadat-Panah · 2018 [cited by applicant]
US 11037554B1 · Le Chevalier · 2021 [cited by applicant]
US 11095579B1 · De Mazancourt · 2021 [cited by applicant]
US 11463585B1 · Ahani · 2022 [cited by applicant]
US 11494851B1 · Novak · 2022 [cited by applicant]
US 11632341B2 · Wang · 2023 [cited by applicant]
US 11715042B1 · Liu · 2023 [cited by applicant]
US 11748128B2 · Chakraborti · 2023 [cited by applicant]
US 11848833B1 · Govindaraju · 2023 [cited by applicant]
US 11922515B1 · Lombard · 2024 [cited by applicant]
US 12126769B1 · Koul · 2024 [cited by applicant]
US 20020032723A1 · Johnson · 2002 [cited by examiner]
US 20040249951A1 · Grabelsky · 2004 [cited by applicant]
US 20070011281A1 · Jhoney · 2007 [cited by applicant]
US 20070038610A1 · Omoigui · 2007 [cited by applicant]
US 20090254336A1 · Dumais · 2009 [cited by applicant]
US 20110138053A1 · Khan · 2011 [cited by applicant]
US 20130080761A1 · Garrett · 2013 [cited by applicant]
US 20130197954A1 · Yankelevich · 2013 [cited by applicant]
US 20140341566A1 · Smith · 2014 [cited by applicant]
US 20140359439A1 · Lyren · 2014 [cited by applicant]
US 20140372160A1 · Nath · 2014 [cited by applicant]
US 20150186154A1 · Brown · 2015 [cited by applicant]
US 20150186189A1 · Venkataraman · 2015 [cited by applicant]
US 20160292011A1 · Colson · 2016 [cited by applicant]
US 20170039239A1 · Saadat-Panah · 2017 [cited by applicant]
US 20170061356A1 · Haas · 2017 [cited by examiner]
US 20170091590A1 · Sawhney · 2017 [cited by applicant]
US 20170139789A1 · Fries · 2017 [cited by applicant]
US 20170262304A1 · Williams · 2017 [cited by applicant]
US 20170287038A1 · Krasadakis · 2017 [cited by applicant]
US 20180096284A1 · Stets · 2018 [cited by applicant]
US 20180189266A1 · Venkataraman · 2018 [cited by applicant]
US 20180232255A1 · Nordin · 2018 [cited by applicant]
US 20180268244A1 · Moazzami · 2018 [cited by applicant]
US 20180293103A1 · Kalmus · 2018 [cited by applicant]
US 20190004815A1 · Povalyayev · 2019 [cited by applicant]
US 20190066016A1 · Ghosh · 2019 [cited by applicant]
US 20190171822A1 · Sjouwerman · 2019 [cited by applicant]
US 20190171984A1 · Irimie · 2019 [cited by applicant]
US 20190243916A1 · Ashoori · 2019 [cited by applicant]
US 20190266254A1 · Blumenfeld · 2019 [cited by applicant]
US 20200027553A1 · Vaughn · 2020 [cited by applicant]
US 20200092232A1 · Tojima · 2020 [cited by applicant]
US 20200142930A1 · Wang · 2020 [cited by applicant]
US 20200195731A1 · Guo · 2020 [cited by applicant]
US 20210043099A1 · Du · 2021 [cited by applicant]
US 20210081250A1 · Chauhan · 2021 [cited by applicant]
US 20210173682A1 · Chakraborti · 2021 [cited by applicant]
US 20210173718A1 · Patel · 2021 [cited by applicant]
US 20210288927A1 · Wang · 2021 [cited by applicant]
US 20220036153A1 · O'Malia · 2022 [cited by applicant]
US 20220179635A1 · Fang · 2022 [cited by applicant]
US 20220197306A1 · Cella · 2022 [cited by applicant]
US 20230040094A1 · Huang · 2023 [cited by applicant]
US 20230060753A1 · Matsuoka · 2023 [cited by applicant]
US 20230074406A1 · Baeuml · 2023 [cited by applicant]
US 20230076327A1 · Matsuoka · 2023 [cited by applicant]
US 20230089596A1 · Huffman · 2023 [cited by applicant]
US 20230136226A1 · Mo · 2023 [cited by applicant]
US 20230156690A1 · Shuwei · 2023 [cited by applicant]
US 20230274094A1 · Tunstall-Pedoe · 2023 [cited by applicant]
US 20230316172A1 · Ayat · 2023 [cited by applicant]
US 20230368284A1 · Sheikh · 2023 [cited by applicant]
US 20230376700A1 · Bista · 2023 [cited by applicant]
US 20240267464A1 · Koneru · 2024 [cited by applicant]
US 20240289863A1 · Smith Lewis · 2024 [cited by applicant]
US 20240346254A1 · Liu · 2024 [cited by applicant]
US 20240346256A1 · Qin · 2024 [cited by applicant]
CA 2773326A1 · 2012 [cited by applicant]
EP 4184322A1 · 2023 [cited by applicant]
WO 2015061976A1 · 2015 [cited by applicant]
WO 2024182285A2 · 2024 [cited by applicant]
'What Are AI Agents? Benefits, Examples, Types' by Salesforce, 2024. (Year: 2024). [cited by applicant]
Foy “Understanding Tokens & Context Windows” (2023) (https://blog.mlq.ai/tokens-context-window-llms/ (Year: 2023). [cited by applicant]
Guan, Lin, et al. “Leveraging pre-trained large language models to construct and utilize world models for model-based task planning.” Advances in Neural Information Processing Systems 36 (2023): 79081-79094. [cited by applicant]
International Search Report and Written Opinion issued in International App. No. PCT/US2024/046328 dated Dec. 12, 2024, 8 pages. [cited by applicant]
Martineau, K, What is retrieval-augmented generation (RAG) ?. Aug. 22, 2023, IBM Research., https://research.ibm.com/blog/ retrieval-augmented-generation-RAG pp. 1-8 (Year: 2023). [cited by applicant]
N Sonnino (Large Language Models as smart space aware social conversational agents)—2022—politesi.polimi.it (Year: 2022). [cited by applicant]
Notice of Allowance dated Dec. 18, 2024 for U.S. Appl. No. 18/882,373 (pp. 1-12). [cited by applicant]
Notice of Allowance dated Dec. 27, 2024 for U.S. Appl. No. 18/882,284 (pp. 1-15). [cited by applicant]
Office Action dated Nov. 15, 2024 for U.S. Appl. No. 18/882,344 (pp. 1-24). [cited by applicant]
Office Action dated Nov. 27, 2024 for U.S. Appl. No. 18/882,326 (pp. 1-20). [cited by applicant]
Office Action dated Dec. 10, 2024 for U.S. Appl. No. 18/882,388 (pp. 1-28). [cited by applicant]
Cited By (3)
US 12,395,337 US 12,407,510 US 12,640,924