# US 2026/0087012 A1 · Executing machine learning models stored as objects in response to queries in a database system

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

- Number: US 2026/0087012 A1
- Document ID: US20260087012A1
- Status: Published application
- Kind: A1
- Application number: 19/406,048
- Filed: 2025-12-02
- Published: 2026-03-26
- Earliest claimed: 2023-04-28
- Assignees: Ocient Holdings LLC
- Applicants: Ocient Holdings LLC
- Inventors: Andrei Lougovtsov; Jason Arnold; Kevin Garner
- Web page: https://patentgrep.com/patent/US20260087012A1

## CPC

- G06F 16/24542: {Plan optimisation}
- G06F 11/3419: {by assessing time}
- G06F 16/24532: {of parallel queries}
- G06F 16/2456: {Join operations}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A database system includes multiple computing device clusters having computing devices with computing nodes and processing core resources. A set of the processing core resources receives a query regarding a machine learning model stored as an object in memory of the database system, the query includes an object name that identifies the machine learning model as a non-standard data object and specifies a requested machine learning purpose. The system identifies data for use by the model to achieve the requested purpose, obtains a query plan that allocates computing resources for executing the model, and retrieves the model from memory based on the object name. The system executes the machine learning model using the identified data in accordance with the query plan to produce a machine learning result corresponding to the requested purpose.

## Claims

### Claim 1

A database system comprises:
a plurality of computing device clusters, wherein a computing device cluster of the plurality of computing device clusters includes a plurality of computing devices, wherein a computing device of the plurality of computing devices includes a plurality of computing nodes, wherein a computing node of the plurality of computing nodes includes a plurality of processing core resources, wherein a set of processing core resources of the computing device cluster is operable to:
receive a query regarding a machine learning (ML) model, wherein the query includes an object name of the machine learning model and a requested ML purpose, wherein the ML model is stored as an object within memory of the database system based on the object name, wherein the object name identifies the ML model as a non-standard data object;
identify data for use by the ML model to achieve the requested ML purpose;
obtain a query plan for the query, wherein the query plan allocates computing resources of the database system to execute the ML model in light of the requested ML purpose;
retrieve the ML model from the memory of the database system based on the object name; and
execute the ML model using the identified data in accordance with the query plan to produce an ML result that corresponds to the requested ML purpose.

### Claim 2 (depends on 1)

The database system of claim 1 further comprises:
the query, including a request for training the ML model, wherein the requested ML purpose is training, and wherein the identified data is training data.

### Claim 3 (depends on 2)

The database system of claim 2 further comprises:
the training includes tuning parameters of the ML model.

### Claim 4 (depends on 2)

The database system of claim 2 further comprises:
the training data is data of a dataset stored as a set of data segments in the memory of the database system.

### Claim 5 (depends on 1)

The database system of claim 1 further comprises:
the query including a request to execute the ML model, wherein the requested ML purpose is to execute the ML model on the identified data to produce an ML data result.

### Claim 6 (depends on 1)

The database system of claim 1, wherein the set of processing core resources are further operable to store the ML model by:
segmenting a plurality of operational instructions that constitute the ML model into a plurality of segments;
processing the plurality of segments in accordance with a long-term storage (LTS) protocol to produce a plurality of LTS segments; and
storing the plurality of LTS segments as the object in the memory of the database system.

### Claim 7 (depends on 6)

The database system of claim 6, wherein the LTS protocol comprises one or more of:
dictionary compression;
data compression;
data deduplication;
data integrity; and
error correction encoding.

### Claim 8 (depends on 1)

The database system of claim 1, wherein the set of processing core resources is further operable to:
interpret the query and/or the object name to determine that the query is regarding storage of a non-standard data object of the database system;
determine that the non-standard data object is the ML model based on metadata associated with the object; and
modify, when required, the query plan to indicate that the query is regarding the non-standard data object, wherein standard data objects include datasets, query operations, and various types of query results, wherein a general non-standard data object is an executable algorithm.

### Claim 9 (depends on 1)

The database system of claim 1, wherein the set of processing core resources is further operable to obtain the query plan by one of:
generating the query plan;
retrieving the query plan from memory of the database system;
looking up the query plan; and
receiving the query plan.

### Claim 10

A computer readable memory device that comprises:
a memory that stores operational instructions that, when executed by a database system including a plurality of computing device clusters, wherein a computing device cluster of the plurality of computing device clusters includes a plurality of computing devices, wherein a computing device of the plurality of computing devices includes a plurality of computing nodes, wherein a computing node of the plurality of computing nodes includes a plurality of processing core resources, causes a set of processing core resources of the computing device cluster to:
receive a query regarding a machine learning (ML) model, wherein the query includes an object name of the machine learning model and a requested ML purpose, wherein the ML model is stored as an object within memory of the database system based on the object name, wherein the object name identifies the ML model as a non-standard data object;
identify data for use by the ML model to achieve the requested ML purpose;
obtain a query plan for the query, wherein the query plan allocates computing resources of the database system to execute the ML model in light of the requested ML purpose;
retrieve the ML model from the memory of the database system based on the object name; and
execute the ML model using the identified data in accordance with the query plan to produce an ML result that corresponds to the requested ML purpose.

### Claim 11 (depends on 10)

The computer readable memory device of claim 10, wherein the instructions further cause the set of processing core resources to:
interpret the query as including a request for training the ML model, wherein the requested ML purpose is training, and where in the identified data is training data.

### Claim 12 (depends on 11)

The computer readable memory device of claim 11, wherein the instructions further cause the set of processing core resources to:
train the ML model by tuning parameters of the ML model.

### Claim 13 (depends on 11)

The computer readable memory device of claim 11, where in the training data is data of a dataset stored as a set of data segments in the memory of the database system.

### Claim 14 (depends on 10)

The computer readable memory device of claim 10, wherein the instructions further cause the set of processing core resources to:
interpret the query as including a request to execute the ML model, wherein the requested ML purpose is execution of the ML model on the identified data to produce an ML data result.

### Claim 15 (depends on 10)

The computer readable memory device of claim 10, wherein the instructions further cause the set of processing core resources to store the ML model by:
segmenting a plurality of operational instructions that constitute the ML model into a plurality of segments;
processing the plurality of segments in accordance with a long-term storage (LTS) protocol to produce a plurality of LTS segments; and
storing the plurality of LTS segments as the object in the memory of the database system.

### Claim 16 (depends on 15)

The computer readable memory device of claim 15, wherein the LTS protocol comprises one or more of:
dictionary compression;
data compression;
data deduplication;
data integrity; and
error correction encoding.

### Claim 17 (depends on 10)

The computer readable memory device of claim 10, wherein the instructions further cause the set of processing core resources to:
interpret the query and/or the object name to determine that the query is regarding storage of a non-standard data object of the database system;
determine that the non-standard data object is the ML model based on metadata associated with the object; and
modify, when required, the query plan to indicate that the query is regarding the non-standard data object, wherein standard data objects include datasets, query operations, and various types of query results, wherein a general non-standard data object is an executable algorithm.

### Claim 18 (depends on 10)

The computer readable memory device of claim 10, wherein the instructions further cause the set of processing core resources to obtain the query plan by one of:
generating the query plan;
retrieving the query plan from memory of the database system;
looking up the query plan; and
receiving the query plan.
