Abstract: Examples enable pallet tag tracking and cluster voting for more accurate pallet tag management using images of a selected pallet. A tag manager tracks a pallet through multiple images of the pallet to ensure the same pallet appears in every image. If the pallet tag is absent from all the images, a tag missing confidence score is generated that indicates the degree of confidence that the tag is missing from the pallet and not merely out of view. The score is used to prioritize handling of pallet tag missing exceptions. If the pallet tag is present in the images, optical character recognition (OCR) results for each tag image are aggregated into a tag cluster with a confidence score calculated for each result. A pallet tag identification (ID) number is predicted based on the result having the highest confidence score to ensure the pallet tag ID is complete and accurate.

Abstract: A system comprises one or more processors and one or more non-transitory computer-readable storage devices storing computing instructions configured to run on the one or more processors and cause the one or more processors to perform operations comprising: receiving a set of images of respective different portions of a label on a package. The operations also can include: determining, using contrast, one or more respective locations in the images of the set of images that are associated with the respective different portions of the label on the package. The operations further can include: providing, for display on a user interface, a flattened reconstruction of the respective different portions of the label on the package based at least in part on the one or more respective locations that have been determined. Other embodiments and features are also disclosed.

Abstract: A system is provided including a processor and a non-transitory computer-readable medium storing computing instructions that, when executed on the processor, cause the processor to perform operations including: obtaining a plurality of query-item pairs associated with a query, wherein the plurality of query-item pairs associated with the query are initially ranked at least partially based on user engagement metrics; obtaining, for each query-item pair of the plurality of query-item pairs associated with the query, first temporal-behavioral features, second temporal-behavioral features, and context-aware features, wherein the context-aware features include a vertical of the query; generating, using a Gradient Boosted Decision Tree (GBDT) model, respective query-item pair GBDT output values based on respective inputs for each query-item pair of the plurality of query-item pairs associated with the query, wherein the respective inputs for each query-item pair of the plurality of query-item pairs associated with

Abstract: Described in detail herein is a nested blockchain system. The nested blockchain system can include distributed nodes in a network. A node is configured to generate a master cryptographically verifiable distributed ledger represented by a sequence of blocks. The node spawns a first sub cryptographically verifiable ledger represented by a first genesis block containing a hash value referencing the first block generated in the master cryptographically verifiable ledger, in response to a first event. The node purges the first sub cryptographically verifiable ledger, in response to a second event.

Abstract: A method can include computing, a dimension of a product using a machine learning model, a dimension of a product. The computing can include, transforming preprocessed data into numerical values using a vectorize layer. The computing can also include, condensing the numerical values using an embedding layer. The predicting can further include, processing the numerical values, as condensed, using a global average pooling layer. The predicting can also include, concatenating the numerical values, as processed into a layer. The predicting can further include processing the layer using a first dense layer. The predicting can additionally include, standardizing the layer, as processed, using a batch normalization layer. The predicting can also include, transforming the layer, as standardized, using an activation function. The predicting can additionally include, regularizing the layer, as transformed, using a dropout function.

Abstract: Example implementations related to next state prediction and action selection are disclosed. In an example, an initial state is estimated based on a set of priors including at least one engagement opportunity. A diminishing effect of the at least one engagement opportunity is determined using a non-linear transformation and a set of likelihood estimates for the initial state is generated using a Bayesian steady state filter model that receives the initial state and the diminishing effect of the at least one engagement opportunity. A next state is predicted based on the initial state and the set of likelihood estimates.

Abstract: Systems and methods for providing item recommendations to increase retention rates of customers are disclosed. In some embodiments, a disclosed method includes: receiving, from a computing device, a recommendation request for recommending items to a customer; determining, based on the recommendation request, an anchor item to be displayed to the customer via a user interface; determining relevance scores between the anchor item and a plurality of candidate items; determining retention scores between the anchor item and the plurality of candidate items; generating, using at least one machine learning model, a ranked list of recommended items based on the plurality of candidate items, the relevance scores and the retention scores; and transmitting to the computing device the ranked list of recommended items to be displayed to the customer with the anchor item on the user interface.

Abstract: A system comprising one or more processors and one or more non-transitory computer-readable storage devices storing computing instructions, when executed on the one or more processors, cause the one or more processors to perform operations comprising: generating a respective similarity score pair comprising an item and one or more potential matches to the item for each item in a set of items; determining a respective page number similarity score between each item and one or more of the one or more respective sub-categories of each item for each item in the set of items; re-ranking one or more remaining respective potential matches; and inserting one or more hyperlinks into an anchor webpage comprising the item, wherein the one or more hyperlinks are configured to navigate a user from the anchor webpage to one or more potential match webpages, as re-ranked. Other embodiments are disclosed herein.

Abstract: A method can be implemented via execution of computing instructions configured to run at a processor and stored at a non-transitory computer-readable media. The method can include: determining one or more virtual categories associated with a search query by a user. The method can also include: classifying one or more search results for the search query into one or more category groups based on the one or more virtual categories. The method can further include: generating a respective user interface portion for each group of the one or more category groups. The method can additionally include: generating a user interface comprising the respective user interface portion for each group of the one or more category groups. The method can also include: transmitting, via a computer network, instructions to display the user interface on a user device for the user. Other embodiments are disclosed.

Abstract: Systems and methods for providing personalized item recommendations with personalized title are disclosed. In some embodiments, each user and each item of a website is represented as an embedding in a same heterogeneous graph. Based on an anchor embedding representing an anchor item in the heterogeneous graph, a set of items are determined. At least one feature aspect is selected from a set of feature aspects based on at least one implicit intent of a user, each feature aspect characterizing a respective aspect of features common to the set of items. From the set of items, a ranked list of items are generated as recommended items based on the at least one feature aspect. A recommendation title describing the at least one feature aspect is generated for the recommended items. The recommended items and the recommendation title are transmitted to be displayed together with the anchor item to the user.

Abstract: This application relates to apparatus and methods for automatically generating item information, such as item descriptions, and providing the item information to customers. For example, the embodiments may generate and provide personalized item descriptions to customers during conversational interactions in speech-based systems. In some examples, the embodiments determine entities (e.g., attributes) from item information, and apply trained machine learning processes to the extracted entities to generate textual data, such as item descriptions. For example, a computing device may apply a trained natural language processing, such as a trained transformer-based machine learning technique, to the extracted entities to generate the item descriptions. In some examples, the computing device applies post processing techniques to the generated textual data. The generated textual data may include descriptive phrases that are user friendly to customers in an e-commerce system.

Abstract: Systems and methods of route optimization are disclosed. An adjustment engine receives a plurality of driver-trip pairs each including a driver selected from a plurality of drivers and a trip selected from a plurality of trips and determines a weight for each of the driver-trip pairs. The weight is determined based on at least one driver-trip feature, at least one driver feature, and at least one trip feature. A route optimization engine selects a set of optimal driver-trip pairs based on a bipartite matching process including the weight for each of the driver trip pairs. The set of optimal driver-trip pairs minimizes total estimated time of arrival to a pickup location for each trip in the plurality of trips. Trip data is transmitted to a corresponding device associated with a corresponding driver paired with each trip in the set of optimal driver-trip pairs.

Abstract: Examples relate to updating data items in a data store. In some examples, a batch of data items are received at a data update node associated with the data store. In accordance with a determination that the batch of data items corresponds to a real-time data update, for each data item, the data item in the data store are updated and supplemental data indicating a historical generation time of the data item is stored in a local map of the data update node. In accordance with a determination that the batch of data items corresponds to a full data refresh, a subset of the batch of data items are adaptively updated based on the supplemental data stored in the local map.

Abstract: Systems and methods for selecting items from a pool of items based on comparisons of composite embeddings are disclosed. A first initial embedding is obtained from a first database for each item in a pool of items and a second initial embedding is obtained from a second database for each item in the pool of items. The first initial embedding is generated using a first embedding model and the second initial embedding is generated using a second embedding model. A first composite embedding is generated for each item in the pool of items comprising the first initial embedding and the second initial embedding; and compare the first composite embedding for each item in the pool of items to a first anchor embedding, wherein the first anchor embedding comprises a first initial anchor embedding generated using the first embedding model and a second initial anchor embedding generated using the second embedding model.

Abstract: A solution for automated food selection includes: an order processing component operable to receive an order identifying an item; a selection component comprising: a first hyperspectral sensor operable to sense a reflection from the item and produce a sensor output based at least on the reflection; and a picking mechanism; a control component operable to: based at least on identification of the item, select a hyperspectral profile from a set of hyperspectral profiles; compare the sensor output with the selected hyperspectral profile; and based at least on the comparison, determine whether to select the item for fulfillment of the order, wherein the picking mechanism is operable to divert the item to a selection output based at least on a determination to select the item for fulfillment of the order; and a transport component operable to transport the item to an order storage zone.

Abstract: Systems and methods including one or more processors and one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and perform: receiving input information for generating one or more routes for one or more vehicles, the input information including one or more group classifiers; processing the input information based on the one or more group classifiers; analyzing, using one or more solving engines, the input information to generate the one or more routes for the one or more vehicles; selecting, from each of the one or more solving engines, a vehicle route from the one or more routes that satisfies a threshold; and transmitting the vehicle route to a dispatcher to facilitate coordinating, by the dispatcher, operation of a vehicle from the one or more vehicles along the vehicle route. Other embodiments are disclosed herein.

Abstract: Systems and methods for attribute recommendation are disclosed. Transaction data related a user is received and attribute recommendations for the user are generated based on the transaction data. The attribute recommendations are generated by a variational inference model configured using a transaction matrix and a loyalty matrix. A set of N recommendations is generated by ranking the generated attribute recommendations based on a combined transaction score and loyalty score and a user interface is generated including the set of N recommendations.

Abstract: A method including obtaining real-time observed orders per minute (OPM) data. The method also can include training a prediction model to make a real-time OPM prediction for a current time period, based on the real-time observed OPM data over past time steps based on lagged time steps in a moving average. The method additionally can include determining, in real-time, whether a demand surge exists based on the real-time observed OPM data and the real-time OPM prediction, to generate a first surge modifier. The method further can include when the demand surge exists, generating, in real-time, a sub-hour revenue per click (RPC) prediction for a first sub-hour time interval. The method additionally can include determining, in real-time, the first surge modifier for the first sub-hour time interval. The method further can include uploading, in real-time, the first surge modifier to a dayparting system of a search engine to bypass existing time intervals and modifiers. Other embodiments are described.

Abstract: In some examples, a system may be configured to, for at least a first user of the plurality of users, implement a first set of operations that generate, for each of a first set of item types, attribute value data. Additionally, the system may implement a second set of operations that generate, for each of a second set of item types identified in catalogue data, clique data. Moreover, the system may, for the at least first user, implement a third set of operations that generate preference dependency data. Further, the system may, for the at least first user, based on the preference dependency data, the clique data, the attribute value data, generate, for each item type of a set of item types, output data including an affinity value for each item type of the first set of item types.