Abstract

An image classification system is provided for determining a likely classification of an image using multiple machine learning models that share a base machine learning model. The image classification system may be a browser-based system on a user computing device that obtains multiple machine learning models over a network from a remote system once, stores the models locally in the image classification system, and uses the models multiple times without needing to subsequently request the machine learning models again from the remote system. The image classification system may therefore determine likely a classification associated with an image by running the machine learning models on a user computing device.

IPC Classes  ?

• G06V 10/764 - Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
• G06V 10/82 - Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
• G06N 3/02 - Neural networks

Abstract

An image classification system is provided for determining a likely classification of an image using multiple machine learning models that share a base machine learning model. The image classification system may be a browser-based system on a user computing device that obtains multiple machine learning models over a network from a remote system once, stores the models locally in the image classification system, and uses the models multiple times without needing to subsequently request the machine learning models again from the remote system. The image classification system may therefore determine likely a classification associated with an image by running the machine learning models on a user computing device.

Abstract

Disclosed are systems and methods for handling categorical field values in machine learning applications, and particularly neural networks. Categorical field values are generally transformed into vectors prior to being passed to a neural network. However, low-dimensionality vectors limit the ability of the network to understand correlations between contextually, semantically, or characteristically similar values. High-dimensionality vectors, in contrast, can overwhelm neural networks, causing the network to seek correlations with respect to individual dimensional values, which correlations may be illusory. The present disclosure relates to a hierarchical neural network that includes a main network as well as one or more auxiliary networks. Categorical field values are processed in an auxiliary network, to reduce a dimensionality of the value before being processed by the main network. This enables contextual, semantic, and characteristic correlations to be identified without overwhelming the network as a whole.

IPC Classes  ?

• G06N 3/045 - Combinations of networks
• G06N 3/08 - Learning methods
• G06Q 20/40 - Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check of credit lines or negative lists
• H04L 9/40 - Network security protocols

Abstract

A network -based service is disclosed for disambiguating search queries based on a location-based clustering of search results corresponding to the query. In one embodiment, a user may submit a query for travel items, such as hotel accommodations. The service can determine, based on the query, an initial set of hotels providing such accommodations. The service can then cluster the hotels according to their geographic positions. If the service identifies multiple clusters, the user can be prompted to select a specific cluster to receive additional information regarding the cluster. Illustratively, if a user submits a query for a hotel in "Springfield," the service may ask the user to select among multiple clusters of results, each corresponding to a different city named "Springfield" in which results have been located.

IPC Classes  ?

• G06F 16/903 - Querying
• G06F 16/9038 - Presentation of query results
• G06F 16/906 - Clustering; Classification

Abstract

A network-based service may be provided for facilitating queries for a number of items, such as travel services. A user may submit a query including criteria for determining relevant items. The network-based service may be configured to generate opaque search results in response to queries, revealing sufficient detail to allow the user to confirm the relevance of the travel item, but insufficient detail to disclose the identity of the supplier. Based on the submitted query or other information, the network-based service may associate a persona with the user, such as the persona of a business or leisure traveler. The network-based service may use the persona to determine a travel item or items best suited to the user, by identifying the travel item attributes that correlate positively with preferences of the persona. The network-based service may then generate an opaque search result that reveals travel item attributes desired by the persona.

IPC Classes  ?

• G06Q 50/12 - Hotels or restaurants

Abstract

A network-based service may be provided for facilitating queries for a number of items, such as travel services. A user may submit a query including criteria for determining one or more relevant items. Based on the submitted query, the network- based service may present the user with information regarding the actions of other similar users of the network-based service, such as searches performed by the other users. Based on this information, the user may elect to supplement the current query to conform to the actions for other users. In some embodiments, actions by other users may be based at least in part on a category of the querying user. By presenting actions of similar users, a current user may be enabled to select the most relevant query terms for identifying a desired item.

IPC Classes  ?

• G06F 16/24 - Querying
• G06Q 10/02 - Reservations, e.g. for tickets, services or events
• G06Q 50/14 - Travel agencies

Abstract

A computer-implemented method for identifying travel items associated with travel advisories during an on-line search for travel items involves in response to a search query, determining, on a per-travel item basis, whether a travel advisory pertains to each travel item of a plurality of travel items available for acquisition. A first travel item associated with a travel item parameter and to which the travel advisory applies is identified and a second travel item that is not associated with the travel item parameter and to which the travel advisory does not apply is identified. A display shows the plurality of travel items available for acquisition and an indication, on a per-travel item basis, of whether a travel advisory pertains to each travel item, as results of a search query, wherein the results indicate that the travel advisory pertains to the first travel item but does not pertain to the second travel item.

IPC Classes  ?

• G06Q 50/14 - Travel agencies

Abstract

A user of a personal computing device may identify multiple travel items of interest displayed for purchase by a network-based travel service. However, the user may not desire to purchase the travel items in a single purchase. Upon making a travel purchase that includes a first travel item that does not include a second travel item corresponding to the first travel item, the user may employ the network travel package service to search for second travel items corresponding to the first travel item, select from discounted travel packages including the first travel item and second travel items returned from the search, and purchase a discounted travel package. Beneficially, multiple travel items may be purchased separately while obtaining a package discount as compared to the stand-alone price of the travel items.

IPC Classes  ?

• G06Q 50/14 - Travel agencies