Methods and systems for forecasting demand for items across multiple channels are disclosed. In some implementations, multi-channel demand forecasting may be performed on a per-item, per-location basis, by selectively generating item-location forecasts for each item and location within a supply chain for each channel, or disaggregating a chain level forecast on a per-item basis to each location. Particular selection of an appropriate model, and selective training of models, allows for efficient computation of such forecasts across a large supply chain with thousands of locations and hundreds of thousands, or millions, of items for which forecasts are generated.

An online concierge system uses a machine learning click through rate model to select promoted items based on user embeddings, item embeddings, and search query embeddings. Embeddings obtained by an embedding model may be used as inputs to the click through rate model. The embedding model may be trained using different actions to score the strength of a customer interaction with an item. For example, a customer purchasing an item may be a stronger signal than a customer placing an item in a shopping cart, which in turn may be a stronger signal than a customer clicking on an item. The online concierge system generates a ranking of candidate promoted items based on the search query and using the click through rate model. Based on the ranking, the online concierge system displays promoted items along with the organic search results to the customer.

A data generation system can include a computing device that is configured to receive a request to generate a training dataset for an attribute and identify a set of item identifiers from an item database based on an engagement indication. The computing device is further configured to, for each item identifier of the set of item identifiers, obtain a query list including queries resulting in an engagement between the corresponding item identifier and a user and, in response to a portion of queries of the query list including the attribute being above a threshold, assign the corresponding item identifier to the training dataset for the attribute. The computing device is also configured to store the training dataset for the attribute in a training dataset database.

An apparatus for performing product matching may include: a processor configured to receive a search as product item for searching for a catalog product that matches a target product; classify the target product in a product taxonomy tree including a plurality of taxonomy nodes, by identifying a taxonomy node to which the target product belongs, among the taxonomy nodes; obtain product data associated with the target product from an internal source and an external source and measure data quality; extract attributes from the product data based on a machine learning model; validate the attribute in response to the attribute corresponding to defined mandatory attribute of the taxonomy node to which the target product belongs, and provide a search result based on the validated attributes; re-rank searched results governed by relevancy score and display matched set of products to the customer above defined re-rank confidence score.

A product catalog service allows merchants to create and store product catalogs indicating products that are available from the merchants. Transaction data from point-of-sale (POS) devices of a plurality of merchants can be received. A first catalog of a first merchant can be obtained from a first POS device of a first merchant. The first catalog contains items to be sold by the first merchant. Using the transaction data and the catalog, a second catalog of items to be sold by a second merchant can be generated or updated and transmitted to the second merchant. Items can be compared using item attributes. Item descriptions in the second catalog can be based on item descriptions in the first catalog.

Provided is an information provision system which determines preference of a user based on a dialogue with the user utilizing an Internet shopping site and determines a recommended commodity based on the determined preference. An information provision server in an information provision system receives utterance (input) by a user utilizing an Internet shopping site via a user terminal and provides a response thereto onto the user terminal, thereby performing control so as to have a conversation with the user. Further, based on the input by the user and user attributes, the information provision server determines a user type of the user and when a recommended commodity is presented to the user, determines the recommended commodity from among commodities purchased by other user whose user type is the same as the user type of the user.

A method for returning a results page responsive to a user search query, such as a search query on a web site, may include receiving a search query from a user, determining, responsive to the query, a set of relevant products from a plurality of product listings based on a similarity of the user query to the respective product listings, retrieving inventory information respective of each of the relevant products, the inventory information comprising one or more available fulfillment channels respective of each of the relevant products, ranking the relevant products with respect to each other according to the inventory information, and returning, to the user, a search result comprising a list of the relevant products, ordered according to the ranking.

Aspects described herein may allow for generating a customized price rating using a machine learning algorithm. This may have the effect of improving the display of information about merchants by including customized, personalized price ratings that better reflect the tastes and preferences of a user or group of users. According to some aspects, these and other benefits may be achieved by using a machine learning model, trained to receive input corresponding to both user data and merchant data and output an indication of customized price rating(s) for the merchant that is specific to one or more users, and then to generate information about the merchant for display that includes the customized price rating(s).

An online platform generates a playlist of clips of a lecture accessed by a plurality of users of the online platform. The online platform receives a recording of the lecture, and receives a plurality of events captured during a time period corresponding to the lecture. Each captured event is associated with a time stamp corresponding to a time at which a user performed an activity while listening to the lecture. The online platform clusters the captured events based on the time stamps, and generates one or more clips of the recording of the lecture from the clustered events. The online platform generates a playlist including the clips of the lecture.

A system, method, and computer-readable medium for facilitating treatment of a vehicle damaged in a crash utilizes near real-time monitoring of the market value of a vehicle type that includes the damaged vehicle, its vehicle parts, and/or one or more arranged vehicle treatments to approximate an extent of damage to the damaged vehicle and determine a prospective vehicle treatment facility for treating the damaged vehicle. In particular, the system continually monitors the market value of a vehicle, its vehicle parts, and/or a prescribed vehicle treatment to calculate a treatment complexity level, e.g., vehicle repair, total loss; for treating the damaged vehicle.