# US 2023/0409293 A1 · System and method for developing an artificial specific intelligence (asi) interface for a specific software > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2023/0409293 A1 - Document ID: US20230409293A1 - Status: Published application - Kind: A1 - Application number: 18/337,770 - Filed: 2023-06-20 - Published: 2023-12-21 - Earliest claimed: 2022-06-20 - Assignees: Lucinetic Inc - Applicants: Lucinetic Inc - Inventors: Konstantin Y. Batygin; Gregory P. Laughlin - Web page: https://patentgrep.com/patent/US20230409293A1 ## CPC - G06F 8/34: Graphical or visual programming - G06F 9/453: {Help systems} - G06F 8/38: for implementing user interfaces - G06F 11/3438: {monitoring of user actions (tracking the activity of the user H04L67/535)} - G06F 40/166: Editing, e.g. inserting or deleting - G06F 40/226: Validation - G06N 3/045: Combinations of networks - G06N 3/0475: Generative networks - G06N 3/08: Learning methods - 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 system and method for developing an artificial specific intelligence (ASI) interface for specific software. The system identifies a specific software package and the user base. The system creates an interaction model for the software based on how users interact with it. Using this model, an initial version of the ASI interface is generated. The system deploys this initial version to a subset of the user base and collects interaction data between the ASI interface and the users. By analyzing the interaction data, the system identifies interaction patterns and pre-defined issues faced by the users. A machine learning model is trained using these patterns and issues. Finally, the system optimizes the initial ASI interface based on the trained machine-learning model. ## Claims ### Claim 1 A computing system for developing an artificial specific intelligence (ASI) interface for a specific software, the computing system comprising: one or more hardware processors; and a memory coupled to the one or more hardware processors, wherein the memory comprises a plurality of modules in the form of programmable instructions executable by the one or more hardware processors, wherein the plurality of modules comprises: an interaction model generation module configured to: identify a specific software package and a user base for generating an artificial specific intelligence (ASI) interface; and generate an interaction model for the identified specific software package based on a medium of interaction with a user; an artificial specific intelligence (ASI) interface generation module configured to: generate an initial version of the ASI interface comprising the generated interaction model; and deploy the generated initial version of the ASI interface with a subset of the user base; a pattern and issue identification module configured to: periodically obtain interaction data between the ASI interface and the user; and identify interaction patterns and pre-defined issues associated with the user by analyzing the obtained interaction data; a machine learning module configured to: train a machine learning model based on the identified interaction patterns and the pre-defined issues associated with the user; and an ASI interface optimizer module configured to: optimize the initial version of the ASI interface based on the trained machine learning model. ### Claim 2 (depends on 1) The computing system of claim 1, wherein the ASI interface optimizer module is further configured to: generate one or more machine learning-based insights for the identified interaction patterns and pre-defined issues associated with the user. ### Claim 3 (depends on 1) The computing system of claim 1, wherein to identify the interaction patterns and the pre-defined issues associated with the user by analyzing the obtained interaction data, the pattern and issue identification module is configured to: identify user behavior patterns and user preference patterns on the ASI interface; and store the identified user behavior pattern and the user preference pattern in the ASI interface. ### Claim 4 (depends on 1) The computing system of claim 1, wherein to optimize the initial version of the ASI interface based on the trained machine learning model, the ASI interface optimizer module is configured to: learn one or more types of tasks performed by the user on the ASI interface; generate accessible instances of the one or more types of tasks performed by the user; and provide the generated accessible instances of the one or more types of tasks to the user at run-time. ### Claim 5 (depends on 1) The computing system of claim 1, wherein to optimize the initial version of the ASI interface based on the trained machine learning model, the ASI interface optimizer module is configured to: analyze the identified specific software package received as input from the user; execute one or more test assessments on the specific software package via an application programming interface; and optimize the initial version of the ASI interface based on the results of the one or more test assessments. ### Claim 6 (depends on 1) The computing system of claim 1, wherein to optimize the initial version of the ASI interface based on the trained machine learning model, the ASI interface optimizer module is configured to: automate complex sequences of actions required to be performed on the specific software package based on the identified interaction patterns and the pre-defined issues associated with the user. ### Claim 7 (depends on 1) The computing system of claim 1, wherein to optimize the initial version of the ASI interface based on the trained machine learning model, the ASI interface optimizer module is configured to: interpret user input commands received from the user with a disability; and generate one or more user-preferred output commands corresponding to the ASI interface based on the interpreted user input commands received from the user with the disability, wherein the one or more user-preferred output commands correspond to a type of disability of the user. ### Claim 8 (depends on 1) The computing system of claim 1, wherein to train the machine learning model based on the identified interaction patterns and the pre-defined issues associated with the user, the machine learning module is configured to: receive unstructured text data as the user base from the user; classify and select language models for the received unstructured text data based on the user base to generate model variables; assign the generated model variables to the identified interaction patterns to generate desired text data; analyze the interaction patterns of the user with the ASI interface using a pre-trained language generation model to generate the desired text data; display the desired text data on the ASI interface for reviewing and formatting; tune and modify the desired text data based on the user preference and the identified pre-defined issues associated with the user; and retrain the machine learning model based on the tuned and modified desired text data. ### Claim 9 (depends on 8) The computing system of claim 8, wherein the machine learning module is further configured to: pre-process the unstructured text data obtained from letters of reference extracted from the user base; generate machine-learning-based summaries of the pre-processed unstructured text data using an AI-based summarization algorithm; validate the generated machine-learning-based summaries based on one or more predefined rules; and map text with the model variables to generate a long-form structured text data desired by the user. ### Claim 10 (depends on 9) The computing system of claim 9, wherein to pre-process the unstructured text data obtained from letters of reference extracted from the user base, the machine learning module is further configured to: parse of letters comprised in the unstructured text data using a decision tree-based model to filter segments lacking actionable intent; filter unstructured text data to remove gendered language; and tune of stereotyped phrases to generate a neutral tone in the unstructured text data. ### Claim 11 (depends on 9) The computing system of claim 9, wherein to generate the machine-learning-based summaries of the pre-processed unstructured text data using the AI-based summarization algorithm, the machine learning module is further configured to: generate the machine-learning-based summaries of the pre-processed unstructured text data using a transformer encoder-encoder machine learning model with a bidirectional encoder and an autoregressive decoder. ### Claim 12 A computer-implemented method for developing an artificial specific intelligence (ASI) interface for specific software, the computer-implemented method comprising: identifying, by one or more hardware processors, a specific software package and a user base for generating an artificial specific intelligence (ASI) interface; generating, by the one or more hardware processors, an interaction model for the identified specific software package based on a medium of interaction with a user; generating, by the one or more hardware processors, an initial version of the ASI interface comprising the generated interaction model; deploying, by the one or more hardware processors, the generated initial version of the ASI interface with a subset of the user base; periodically obtaining, by the one or more hardware processors, interaction data between the ASI interface and the user; identifying, by the one or more hardware processors, interaction patterns and pre-defined issues associated with the user by analyzing the obtained interaction data; training, by the one or more hardware processors, a machine learning model based on the identified interaction patterns and the pre-defined issues associated with the user; and optimizing, by the one or more hardware processors, the initial version of the ASI interface based on the trained machine learning model. ### Claim 13 (depends on 12) The computer-implemented method of claim 12 further comprising: generating, by the one or more hardware processors, one or more machine learning-based insights for the identified interaction patterns and pre-defined issues associated with the user. ### Claim 14 (depends on 12) The computer-implemented method of claim 12, wherein identifying the interaction patterns and the pre-defined issues associated with the user by analyzing the obtained interaction data, further comprises: identifying, by the one or more hardware processors, user behavior patterns and user preference patterns on the ASI interface; and storing, by the one or more hardware processors, the identified user behavior pattern and the user preference pattern in the ASI interface. ### Claim 15 (depends on 12) The computer-implemented method of claim 12, wherein optimizing the initial version of the ASI interface based on the trained machine learning model, further comprises: learning, by the one or more hardware processors, one or more types of tasks performed by the user on the ASI interface; generating, by the one or more hardware processors, accessible instances of the one or more types of tasks performed by the user; and providing, by the one or more hardware processors, the generated accessible instances of the one or more types of tasks to the user at run-time. ### Claim 16 (depends on 12) The computer-implemented method of claim 12, wherein optimizing the initial version of the ASI interface based on the trained machine learning model, further comprises: analyzing, by the one or more hardware processors, the identified specific software package received as input from the user; executing, by the one or more hardware processors, one or more test assessments on the specific software package via an application programming interface; optimizing, by the one or more hardware processors, the initial version of the ASI interface based on the results of the one or more test assessments; automating, by the one or more hardware processors, complex sequences of actions required to be performed on the specific software package based on the identified interaction patterns and the pre-defined issues associated with the user; interpreting, by the one or more hardware processors, user input commands received from the user with a disability; and generating, by the one or more hardware processors, one or more user-preferred output commands corresponding to the ASI interface based on the interpreted user input commands received from the user with the disability, wherein the one or more user-preferred output commands correspond to a type of disability of the user. ### Claim 17 (depends on 12) The computer-implemented method of claim 12, wherein training the machine learning model based on the identified interaction patterns and the pre-defined issues associated with the user, further comprises: receiving, by the one or more hardware processors, unstructured text data as the user base from the user; classifying and selecting, by the one or more hardware processors, language models for the received unstructured text data based on the user base to generate model variables; assigning, by the one or more hardware processors, the generated model variables to the identified interaction patterns to generate desired text data; analyzing, by the one or more hardware processors, the interaction patterns of the user with the ASI interface using a pre-trained language generation model to generate the desired text data; displaying, by the one or more hardware processors, the desired text data on the ASI interface for reviewing and formatting; tuning and modifying, by the one or more hardware processors, the desired text data based on the user preference and the identified pre-defined issues associated with the user; and retraining, by the one or more hardware processors, the machine learning model based on the tuned and modified desired text data. ### Claim 18 (depends on 19) The computer-implemented method of claim 19 further comprising: pre-processing, by the one or more hardware processors, the unstructured text data obtained from letters of reference extracted from the user base; generating, by the one or more hardware processors, machine-learning-based summaries of the pre-processed unstructured text data using at least one of an AI-based summarization algorithm, a transformer encoder-encoder machine learning model with a bidirectional encoder and an autoregressive decoder; validating, by the one or more hardware processors, the generated machine-learning-based summaries based on one or more predefined rules; and mapping, by the one or more hardware processors, text with the model variables to generate a long form structured text data desired by the user. ### Claim 19 (depends on 9) The computing system of claim 9, wherein pre-processing the unstructured text data obtained from letters of reference extracted from the user base, the machine learning module is further configured to: parse of letters comprised in the unstructured text data using a decision tree-based model to filter segments lacking actionable intent; filter unstructured text data to remove gendered language; and tune of stereotyped phrases to generate a neutral tone in the unstructured text data. ### Claim 20 A non-transitory computer-readable storage medium having programmable instructions stored therein, that when executed by one or more hardware processors, cause the one or more hardware processors to: identify a specific software package and a user base for generating an artificial specific intelligence (ASI) interface; generate an interaction model for the identified specific software package based on a medium of interaction with a user; generate an initial version of the ASI interface comprising the generated interaction model; deploy the generated initial version of the ASI interface with a subset of the user base; periodically obtain interaction data between the ASI interface and the user; identify interaction patterns and pre-defined issues associated with the user by analyzing the obtained interaction data; train a machine learning model based on the identified interaction patterns and the pre-defined issues associated with the user, and optimize the initial version of the ASI interface based on the trained machine learning model.