# US 2026/0099352 A1 · Cognitive kermel: a dockerized "autopilot" system that everyone can deploy privately > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2026/0099352 A1 - Document ID: US20260099352A1 - Status: Published application - Kind: A1 - Application number: 18/905,638 - Filed: 2024-10-03 - Published: 2026-04-09 - Earliest claimed: 2024-10-03 - Assignees: Tencent America LLC - Applicants: Tencent America LLC - Inventors: Hongming Zhang; Xiaoman Pan; Hongwei Wang; Kaixin Ma; Wenhao Yu; Dong Yu - Web page: https://patentgrep.com/patent/US20260099352A1 ## CPC - G06F 9/4806: {Task transfer initiation or dispatching} - G06F 9/4843: {by program, e.g. task dispatcher, supervisor, operating system} - G06F 9/547: {Remote procedure calls [RPC]; Web services} - G06F 16/2455: Query execution - G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/045: Combinations of networks - G06N 3/0455: Auto-encoder networks; Encoder-decoder networks - G06N 3/047: Probabilistic or stochastic networks - G06N 3/08: Learning methods - G06N 3/088: Non-supervised learning, e.g. competitive learning - G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound - G06N 5/022: Knowledge engineering; Knowledge acquisition - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 20/00: Machine learning - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract A method includes receiving, by an autopilot system, a user query; determining, by a perception kernel of the autopilot system, one or more states associated with the user query; executing, by a reasoning kernel of the autopilot system, one or more actions associated with the one or more states; and completing, by the autopilot system, execution of a task associated with the user query. ## Claims ### Claim 1 A method performed by at least one processor, the method comprising: receiving, by an autopilot system, a user query; determining, by a perception kernel of the autopilot system, one or more states associated with the user query; executing, by a reasoning kernel of the autopilot system, one or more actions associated with the one or more states; and completing, by the autopilot system, execution of a task associated with the user query. ### Claim 2 (depends on 1) The method according to claim 1, further comprising determining the one or more actions by inputting the one or more states into a large language model (LLM), wherein an output of the LLM corresponds to the one or more actions. ### Claim 3 (depends on 1) The method according to claim 1, wherein the determining the one or more states includes: generating a web instructions, and executing the web instruction to open a web browser and execute one or more web searches based on the user query, wherein at least one state from the one or more states is determined based on the one or more web searches. ### Claim 4 (depends on 3) The method according to claim 3, wherein the executing the web instruction is associated with one or more actions including: click an element on a webpage, type content in an element on the webpage, scroll up or down, go back to a previously browsed page, return to a homepage and restart execution of the one or more web searches, and send information to the reasoning kernel. ### Claim 5 (depends on 1) The method according to claim 1, wherein the determining the one or more states includes: analyzing one or more local files, wherein at least one state form the one or more states is determined based on the analyzing the one or more local files. ### Claim 6 (depends on 5) The method according to claim 5, the analyzing the one or more local files is associated with one or more actions including: counting a number of occurrences of a term, finding the term, navigating to different sections in file, performing a semantic-based search, extract information from a file. ### Claim 7 (depends on 1) The method according to claim 1, wherein the autopilot system further includes a memory kernel that stores one or more previous states. ### Claim 8 (depends on 7) The method according to claim 7, wherein at least one state from the one or more states is determined based on the one or more previous states. ### Claim 9 (depends on 1) The method according to claim 1, wherein the autopilot system is an application implemented as a Docker container. ### Claim 10 (depends on 9) The method according to claim 9, wherein the autopilot system includes a front end Docker, a backend Docker, a web accessing Docker, a database Docker, a service Docker, and a local file system. ### Claim 11 An apparatus comprising: at least one memory configured to store program code; and at least one processor configured to read the program code and operate as instructed by the program code, the program code including: receiving code configured to cause the at least one processor to receive, by an autopilot system, a user query; determining code configured to cause the at least one processor to determine, by a perception kernel of the autopilot system, one or more states associated with the user query; executing code configured to cause the at least one processor to execute, by a reasoning kernel of the autopilot system, one or more actions associated with the one or more states; and completing code configured to cause the at least one processor to complete, by the autopilot system, execution of a task associated with the user query. ### Claim 12 (depends on 11) The apparatus according to claim 11, wherein the determining code further causes the at least one processor to determine the one or more actions by inputting the one or more states into a large language model (LLM), wherein an output of the LLM corresponds to the one or more actions. ### Claim 13 (depends on 11) The apparatus according to claim 11, wherein the determining code further causes the at least one processor to: generate a web instructions, and execute the web instruction to open a web browser and execute one or more web searches based on the user query, wherein at least one state from the one or more states is determined based on the one or more web searches. ### Claim 14 (depends on 13) The apparatus according to claim 13, wherein execution of the web instruction is associated with one or more actions including: click an element on a webpage, type content in an element on the webpage, scroll up or down, go back to a previously browsed page, return to a homepage and restart execution of the one or more web searches, and send information to the reasoning kernel. ### Claim 15 (depends on 11) The apparatus according to claim 11, wherein the determining code further causes the at least one processor to: analyze one or more local files, wherein at least one state form the one or more states is determined based on the analyzing the one or more local files. ### Claim 16 (depends on 15) The apparatus according to claim 15, the analyzing the one or more local files is associated with one or more actions including: counting a number of occurrences of a term, finding the term, navigating to different sections in file, performing a semantic-based search, extract information from a file. ### Claim 17 (depends on 11) The apparatus according to claim 11, wherein the autopilot system further includes a memory kernel that stores one or more previous states. ### Claim 18 (depends on 17) The apparatus according to claim 17, wherein at least one state from the one or more states is determined based on the one or more previous states. ### Claim 19 (depends on 11) The apparatus according to claim 11, wherein the autopilot system is an application implemented as a Docker container. ### Claim 20 A non-transitory computer readable medium having instructions stored therein, which when executed by a processor cause the processor to execute a method comprising: receiving, by an autopilot system, a user query; determining, by a perception kernel of the autopilot system, one or more states associated with the user query; executing, by a reasoning kernel of the autopilot system, one or more actions associated with the one or more states; and completing, by the autopilot system, execution of a task associated with the user query.