Patents by Inventor Matthew Hagen
Matthew Hagen has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20240144339Abstract: A computer-implemented method for determining image similarity includes determining, by a first neural network, a first feature value associated with a first characteristic of a first product based on an image of the first product. The method also includes determining, by a second neural network, a second feature value associated with a second characteristic of the first product based on the image of the first product. The method further involves calculating a first vector space distance between the first feature value and a third feature value associated with the first characteristic of a second product, and calculating a second vector space distance between the second feature value and a fourth feature value associated with the second characteristic of the second product. Additionally, the method includes determining a similarity value based on the first vector space distance and the second vector space distance.Type: ApplicationFiled: January 5, 2024Publication date: May 2, 2024Inventors: Estelle Afshar, Matthew Hagen, Huiming Qu
-
Patent number: 11954572Abstract: A method for machine learning-based classification may include training a machine learning model with a full training data set, the full training data set comprising a plurality of data points, to generate a first model state of the machine learning model, generating respective embeddings for the data points in the full training data set with the first model state of the machine learning model, applying a clustering algorithm to the respective embeddings to generate one or more clusters of the embeddings, identifying outlier embeddings from the one or more clusters of the embeddings, generating a reduced training data set comprising the full training data set less the data points associated with the outlier embeddings, training the machine learning model with the reduced training data set to a second model state, and applying the second model state to one or more data sets to classify the one or more data sets.Type: GrantFiled: May 16, 2023Date of Patent: April 9, 2024Assignee: Home Depot Product Authority, LLCInventors: Matthew Hagen, Estelle Afshar, Huiming Qu, Ala Eddine Ayadi, Jiaqi Wang
-
Publication number: 20240070554Abstract: A method for machine learning-based classification may include training a machine learning model with a full training data set, the full training data set comprising a plurality of data points, to generate a first model state of the machine learning model, generating respective embeddings for the data points in the full training data set with the first model state of the machine learning model, applying a clustering algorithm to the respective embeddings to generate one or more clusters of the embeddings, identifying outlier embeddings from the one or more clusters of the embeddings, generating a reduced training data set comprising the full training data set less the data points associated with the outlier embeddings, training the machine learning model with the reduced training data set to a second model state, and applying the second model state to one or more data sets to classify the one or more data sets.Type: ApplicationFiled: May 16, 2023Publication date: February 29, 2024Inventors: Matthew Hagen, Estelle Afshar, Huiming Qu, Ala Eddine Ayadi, Jiaqi Wang
-
Patent number: 11907987Abstract: A computer-implemented method for determining image similarity includes determining, by a first neural network, a first feature value associated with a first characteristic of a first product based on an image of the first product. The method also includes determining, by a second neural network, a second feature value associated with a second characteristic of the first product based on the image of the first product. The method further involves calculating a first vector space distance between the first feature value and a third feature value associated with the first characteristic of a second product, and calculating a second vector space distance between the second feature value and a fourth feature value associated with the second characteristic of the second product. Additionally, the method includes determining a similarity value based on the first vector space distance and the second vector space distance.Type: GrantFiled: December 14, 2021Date of Patent: February 20, 2024Assignee: Home Depot Product Authority, LLCInventors: Estelle Afshar, Matthew Hagen, Huiming Qu
-
Patent number: 11687841Abstract: A method for machine learning-based classification may include training a machine learning model with a full training data set, the full training data set comprising a plurality of data points, to generate a first model state of the machine learning model, generating respective embeddings for the data points in the full training data set with the first model state of the machine learning model, applying a clustering algorithm to the respective embeddings to generate one or more clusters of the embeddings, identifying outlier embeddings from the one or more clusters of the embeddings, generating a reduced training data set comprising the full training data set less the data points associated with the outlier embeddings, training the machine learning model with the reduced training data set to a second model state, and applying the second model state to one or more data sets to classify the one or more data sets.Type: GrantFiled: June 5, 2020Date of Patent: June 27, 2023Assignee: HOME DEPOT PRODUCT AUTHORITY, LLCInventors: Matthew Hagen, Estelle Afshar, Huiming Qu, Ala Eddine Ayadi, Jiaqi Wang
-
Publication number: 20220253643Abstract: A computer-implemented method for determining image similarity includes determining, by a first neural network, a first feature value associated with a first characteristic of a first product based on an image of the first product. The method also includes determining, by a second neural network, a second feature value associated with a second characteristic of the first product based on the image of the first product. The method further involves calculating a first vector space distance between the first feature value and a third feature value associated with the first characteristic of a second product, and calculating a second vector space distance between the second feature value and a fourth feature value associated with the second characteristic of the second product. Additionally, the method includes determining a similarity value based on the first vector space distance and the second vector space distance.Type: ApplicationFiled: December 14, 2021Publication date: August 11, 2022Inventors: Estelle Afshar, Matthew Hagen, Huiming Qu
-
Patent number: 11200445Abstract: A computer-implemented method for determining image similarity includes determining, by a first neural network, a first feature value associated with a first characteristic of a first product based on an image of the first product. The method also includes determining, by a second neural network, a second feature value associated with a second characteristic of the first product based on the image of the first product. The method further involves calculating a first vector space distance between the first feature value and a third feature value associated with the first characteristic of a second product, and calculating a second vector space distance between the second feature value and a fourth feature value associated with the second characteristic of the second product. Additionally, the method includes determining a similarity value based on the first vector space distance and the second vector space distance.Type: GrantFiled: January 22, 2020Date of Patent: December 14, 2021Assignee: Home Depot Product Authority, LLCInventors: Estelle Afshar, Matthew Hagen, Huiming Qu
-
Publication number: 20210224582Abstract: A computer-implemented method for determining image similarity includes determining, by a first neural network, a first feature value associated with a first characteristic of a first product based on an image of the first product. The method also includes determining, by a second neural network, a second feature value associated with a second characteristic of the first product based on the image of the first product. The method further involves calculating a first vector space distance between the first feature value and a third feature value associated with the first characteristic of a second product, and calculating a second vector space distance between the second feature value and a fourth feature value associated with the second characteristic of the second product. Additionally, the method includes determining a similarity value based on the first vector space distance and the second vector space distance.Type: ApplicationFiled: January 22, 2020Publication date: July 22, 2021Inventors: Estelle Afshar, Matthew Hagen, Huiming Qu
-
Publication number: 20200387755Abstract: A method for machine learning-based classification may include training a machine learning model with a full training data set, the full training data set comprising a plurality of data points, to generate a first model state of the machine learning model, generating respective embeddings for the data points in the full training data set with the first model state of the machine learning model, applying a clustering algorithm to the respective embeddings to generate one or more clusters of the embeddings, identifying outlier embeddings from the one or more clusters of the embeddings, generating a reduced training data set comprising the full training data set less the data points associated with the outlier embeddings, training the machine learning model with the reduced training data set to a second model state, and applying the second model state to one or more data sets to classify the one or more data sets.Type: ApplicationFiled: June 5, 2020Publication date: December 10, 2020Inventors: Matthew Hagen, Estelle Afshar, Huiming Qu, Ala Eddine Ayadi, Jiaqi Wang
-
Publication number: 20060265272Abstract: A system and methods for determining an operational characteristic of a business process associated with a network. An operational characteristic of the business process is determined. An action is taken to modify a parameter associated with a change in the operational characteristic of the business process. The operational characteristic of the business process is re-determined.Type: ApplicationFiled: May 17, 2005Publication date: November 23, 2006Inventors: Patrick Bosa, Matthew Hagen, Thomas Pantelis
-
Publication number: 20050235890Abstract: A product-on-demand delivery system for agricultural product overcomes the problem of over pressuring near application units in order to achieve sufficient pressure to feed further application units. The invention provides two or more main hoppers, each hopper pressurized to a different level by a fan that charges the air nozzles on the hoppers. The near row application units have product delivery hoses that are coupled to a first hopper. The far row application units have longer product delivery hoses that are coupled to a second hopper. The fan charging the second hopper is configured to pressurize the second hopper to a higher pressure than the pressure of the first hopper. The invention provides separate product-on-demand delivery systems to better control the product flow to the row application units on the machine. A product-on-demand delivery system is also provided that can dispense two different products from two main hoppers operated at different pressures.Type: ApplicationFiled: April 26, 2004Publication date: October 27, 2005Inventors: Nathan Mariman, Matthew Hagen, Christopher Myers
-
Publication number: 20050028714Abstract: A pneumatic seed on demand delivery system comprises a frame having a main hopper and an individual planting unit. An air pump directs pressurized air to a nozzle located at the base of the main hopper. The seed in the main hopper is taken up by the air stream passing through the nozzle and is directed to an auxiliary hopper located on the planting units. The auxiliary hopper has sidewalls that are provided with an opening into which is inserted a removable screen for venting the interior of the auxiliary hopper. An exterior wall extends outward from the auxiliary hopper to protect the removable screen and is open at the bottom to vent air from the removable screen. The removable screen is trapped in the opening by a removable lid engaging the sidewalls of the auxiliary hopper.Type: ApplicationFiled: August 8, 2003Publication date: February 10, 2005Inventors: Matthew Hagen, James Lodico