Enhancing database query generation in large language models using evolutionary computation algorithms

A non-transitory computer readable medium comprising computer readable program code, which when executed by a computer processor enables the computer processor to perform a method for servicing user queries, the method comprising:

• obtaining, by a data system and from a client device, a user query for accessing a database;
• in response to the user query:
• configuring an enhanced context embedding model;
• applying a vectorized embedding on the user query using the enhanced context embedding model to obtain a vectorized user query;
• performing a hyperparameter optimization on the vectorized user query using a pre-defined optimization algorithm to obtain an optimal database query,
• wherein the pre-defined optimization algorithm comprises:
• initializing a population of potential solutions of database queries. wherein each potential solution in the population represents a velocity and position;
• evaluating a fitness value of each potential solution in the population;
• for each potential solution in the population:
• evaluating a corresponding database query to obtain a particle score of each database solution, and updating a particle personal best position and a global best position based on the particle score;
• and
• updating a velocity and position of each potential solution based on the global best position and each corresponding particle personal best position to obtain a new velocity and position of each potential solution in the population;
• issuing the optimal database query to the database to obtain a response;
• and
• providing the response to the client device.

A system, comprising:

• a data system operating on a processor;
• and
• memory comprising instructions, which when executed by the processor, perform a method comprising:
• obtaining, from a client device, a user query for accessing a database, wherein the user query is written in a natural language;
• in response to the user query:
• configuring an enhanced context embedding model;
• applying a vectorized embedding on the user query using the enhanced context embedding model to obtain a vectorized user query;
• performing a hyperparameter optimization on the vectorized user query using a pre-defined optimization algorithm to obtain an optimal database query,
• wherein the pre-defined optimization algorithm comprises:
• initializing a population of potential solutions of database queries, wherein each potential solution in the population represents a velocity and position;
• evaluating a fitness value of each potential solution in the population;
• for each potential solution in the population:
• evaluating a corresponding database query to obtain a particle score of each database solution, and updating a particle personal best position and a global best position based on the particle score;
• and
• updating a velocity and position of each potential solution based on the global best position and each corresponding particle personal best position to obtain a new velocity and position of each potential solution in the population;
• issuing the optimal database query to the database to obtain a response;
• and
• providing the response to the client device.