US 2022/0035878 A1Published application

Framework for optimization of machine learning architectures

AssigneeIntel Corporation

InventorsAnthony Sarah, Daniel Cummings, Juan Pablo Munoz, Tristan Webb

Life of this patent

Estimating the term…

Loading patent…

Abstract

The present disclosure is related to framework for automatically and efficiently finding machine learning (ML) architectures that are optimized to one or more specified performance metrics and/or hardware platforms. This framework provides ML architectures that are applicable to specified ML domains and are optimized for specified hardware platforms in significantly less time than could be done manually and in less time than existing ML model searching techniques. Furthermore, a user interface is provided that allows a user to search for different ML architectures based on modified search parameters, such as different hardware platform aspects and/or performance metrics. Other embodiments may be described and/or claimed.

Claims26

Dependent claims

An apparatus for identifying machine learning (ML) architectures, the apparatus comprising:

a population initializer configured to initialize a set of candidate ML architectures based on an ML configuration;
an architecture search type engine connected to the population initializer, the architecture search type engine configured to determine a search strategy for searching the initialized population;
a multi-objective candidate generator (MOCG) connected to the architecture search type engine, the MOCG configured to:
search the initialized set of candidate ML architectures according to the search strategy, and
determine, based on the search, a set of optimal ML architectures from the set of candidate ML architectures, the set of optimal ML architectures satisfy one or more of a set of ML parameters and hardware platform information (HPI) included in the ML configuration;
and
a performance metric evaluator connected to the MOCG, the performance metric evaluator configured to evaluate performance of each optimal ML architecture in the set of optimal ML architectures.

4 more elements

One or more non-transitory computer readable media (NTCRM) comprising instructions for operating a machine learning architecture search interface (MLASI), wherein execution of the instructions by one or more processors of a compute node is to cause the compute node to:

serve the MLASI to a client device;
and
in response to receipt of an individual machine learning (ML) configuration from the client device via the MLASI:
determine a set of candidate ML architectures based on a super network indicated by the individual ML configuration,
perform multi-objective optimization on the set of candidate ML architectures to determine a set of optimal ML architectures that satisfy one or more of a combination of ML parameters and hardware platform information (HPI) included in the ML configuration,
determine performance metrics for each optimal ML architecture in the set of optimal ML architectures, and
indicate, via the MLASI, the set of optimal ML architectures and the determined performance metrics for each optimal ML architecture.

4 more elements

Dependencies are inferred from the claim text.

Facts

Application: 17/505,568Filed 19 Oct 2021
Published: 3 Feb 2022
Term ends: Not estimatedThe source records don’t support an estimate.

Classification15 codes

G06F 16/953
Retrieval from the web: querying, e.g. by the use of web search engines
G06N 3/082
Learning methods: modifying the architecture, e.g. adding, deleting or silencing nodes or connections
G06N 3/045
Architecture, e.g. interconnection topology: combinations of networks
G06N 3/0464
Architecture, e.g. interconnection topology: convolutional networks [CNN, ConvNet]

Family

US 12,367,249 B2Related patent
US 2022/0035878 A1This patent

From the US filing history. Foreign family members aren’t listed.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key