# US 2026/0099491 A1 · End-to-end SQL generation

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0099491 A1
- Document ID: US20260099491A1
- Status: Published application
- Kind: A1
- Application number: 19/366,345
- Filed: 2025-10-22
- Published: 2026-04-09
- Earliest claimed: 2024-10-08
- Applicants: Distyl AI, Inc.
- Inventors: Karime Maamari; Yonas Tesfaye; Amine Mhedhbi
- Web page: https://patentgrep.com/patent/US20260099491A1

## CPC

- G06F 16/2452: Query translation
- G06F 16/24522: {Translation of natural language queries to structured queries}
- G06F 16/242: Query formulation
- G06F 16/243: {Natural language query formulation}
- G06F 16/2433: {Query languages}
- G06N 3/08: Learning methods
- G06N 5/02: Knowledge representation; Symbolic representation
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N 20/00: Machine learning
- G06F 16/2453: Query optimisation
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Embodiments include systems and methods for user authentication using behavioral biometrics. In some embodiments, the method includes collecting raw behavioral data regarding user interactions with a device; extracting a plurality of features from the raw behavioral data; mapping the features to a plurality of mathematical spaces; generating, based on the mapped features, a unified behavioral profile; and determining, based on the unified behavioral profile, an authentication decision for a user of the device.

## Claims

### Claim 1

A computer-implemented method for structured language query generation, comprising:
generating a knowledge set based on at least one query log and at least one domain-specific document;
receiving a natural language query for a database;
retrieving data relevant to the natural language query from the knowledge set;
generating, based on the relevant data and the natural language query, a candidate structured language query using a machine learning model; and
presenting the candidate structured language query for execution against the database.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, wherein the candidate structured language query comprises a SQL query.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein the knowledge set comprises generation instructions, sub-expressions representing structured language queries at the query log, and a schema representation indicating the structure of the database.

### Claim 4 (depends on 3)

The computer-implemented method of claim 3, wherein the generating the knowledge set comprises:
retrieving the structured language queries from the query log;
reformatting the structured language queries into a sketch based on common table expressions; and
decomposing the sketch into sub-queries; and
generating the sub-expressions based on the sub-queries.

### Claim 5 (depends on 3)

The computer-implemented method of claim 3, wherein the generation instructions and sub-expressions are partitioned in the knowledge set according to user intent.

### Claim 6 (depends on 3)

The computer-implemented method of claim 3, wherein the schema representation indicates the structure of the database using at least one of a table name, column name, column type, or column data sample.

### Claim 7 (depends on 3)

The computer-implemented method of claim 3, wherein the retrieving the relevant data comprises:
reformatting the natural language query in a canonical form;
identifying a user intent of the natural language query;
identifying relevant generation instructions and sub-expressions from the knowledge set based on the user intent; and
identifying irrelevant elements of the schema representation using a second machine learning model.

### Claim 8 (depends on 1)

The computer-implemented method of claim 1, wherein the generating the candidate structured language query comprises constructing, based on the natural language query and the relevant data, a chain-of-thought reasoning plan comprising at least one pseudo-SQL statement.

### Claim 9 (depends on 1)

The computer-implemented method of claim 1, wherein the machine learning model comprises a large language model configured to receive the natural language query and the relevant data as input.

### Claim 10 (depends on 1)

The computer-implemented method of claim 1, further comprising:
receiving feedback regarding the execution of the candidate structured language query; and
updating the knowledge set based on the feedback; and
updating the candidate structured language query based on the updated knowledge set.

### Claim 11 (depends on 10)

The computer-implemented method of claim 10, wherein the feedback comprises at least one of user feedback or feedback from an evaluation model.

### Claim 12

A system for structured language query generation, comprising:
memory storing instructions; and
a processor executing the instructions to perform the steps of:
generating a knowledge set based on at least one query log and at least one domain-specific document;
receiving a natural language query for a database;
retrieving data relevant to the natural language query from the knowledge set;
generating, based on the relevant data and the natural language query, a candidate structured language query using a machine learning model; and
presenting the candidate structured language query for execution against the database.

### Claim 13 (depends on 12)

The system of claim 12, wherein the candidate structured language query comprises a SQL query.

### Claim 14 (depends on 12)

The system of claim 12, wherein the knowledge set comprises generation instructions, sub-expressions representing structured language queries at the query log, and a schema representation indicating the structure of the database.

### Claim 15 (depends on 14)

The system of claim 14, wherein the generating the knowledge set comprises:
retrieving the structured language queries from the query log;
reformatting the structured language queries into a sketch based on common table expressions; and
decomposing the sketch into sub-queries; and
generating the sub-expressions based on the sub-queries.

### Claim 16 (depends on 14)

The system of claim 14, wherein the retrieving the relevant data comprises:
reformatting the natural language query in a canonical form;
identifying a user intent of the natural language query;
identifying relevant generation instructions and sub-expressions from the knowledge set based on the user intent; and
identifying irrelevant elements of the schema representation using a second machine learning model.

### Claim 17 (depends on 12)

The system of claim 12, wherein the generating the candidate structured language query comprises constructing, based on the natural language query and the relevant data, a chain-of-thought reasoning plan comprising at least one pseudo-SQL statement.

### Claim 18 (depends on 12)

The system of claim 12, wherein the machine learning model comprises a large language model configured to receive the natural language query and the relevant data as input.

### Claim 19 (depends on 12)

The system of claim 12, wherein the processor is further configured to perform the steps of:
receiving feedback regarding the execution of the candidate structured language query; and
updating the knowledge set based on the feedback; and
updating the candidate structured language query based on the updated knowledge set.

### Claim 20

A computer program product embodied in a non-transitory computer readable storage medium and comprising computer instructions for:
generating a knowledge set based on at least one query log and at least one domain-specific document;
receiving a natural language query for a database;
retrieving data relevant to the natural language query from the knowledge set;
generating, based on the relevant data and the natural language query, a candidate structured language query using a machine learning model; and
presenting the candidate structured language query for execution against the database.
