US 12,405,980 B2
Method and system for multi-level artificial intelligence supercomputer design
Vijay Madisetti, Alpharetta, GA (US); and Arshdeep Bahga, Chandigarh (IN)
Assigned to Vijay Madisetti, Alpharetta, GA (US)
Filed by Vijay Madisetti, Alpharetta, GA (US)
Filed on Mar. 3, 2025, as Appl. No. 19/068,482.
Application 19/068,482 is a continuation of application No. 18/934,384, filed on Nov. 1, 2024, granted, now 12,292,907.
Application 18/934,384 is a continuation of application No. 18/921,923, filed on Oct. 21, 2024.
Application 18/921,923 is a continuation of application No. 18/391,165, filed on Dec. 20, 2023, granted, now 12,153,612, issued on Nov. 26, 2024.
Application 18/391,165 is a continuation of application No. 18/348,692, filed on Jul. 7, 2023, granted, now 12,001,462, issued on Jun. 4, 2024.
Claims priority of provisional application 63/463,913, filed on May 4, 2023.
Claims priority of provisional application 63/469,571, filed on May 30, 2023.
Prior Publication US 2025/0200083 A1, Jun. 19, 2025
Int. Cl. G06F 16/3329 (2025.01); G06F 40/284 (2020.01)
CPC G06F 16/3329 (2019.01) [G06F 40/284 (2020.01)] 14 Claims
OG exemplary drawing
 
1. A method of processing language model input data in a distributed computing environment, comprising:
receiving an input data stream;
tokenizing data received from the input data stream into a plurality of tokens using a map-reduce operation in the distributed computing environment;
processing the plurality of tokens in parallel in the distributed computing environment to produce a plurality of processed tokens;
aggregating the plurality of processed tokens using a reduce operation to produce a plurality of aggregated tokens;
generating one or more updated incrementally-updated family of large language models (h-LLMs) by updating one or more incrementally-updated h-LLMs with the plurality of aggregated tokens in real-time; and
responding to a user query using the one or more updated incrementally-updated h-LLMs.