Abstract: Systems and methods for query processing are described. Embodiments of the present disclosure identify a target phrase in an original query, wherein the target phrase comprises a phrase to be replaced in the original query; replace the target phrase with a mask token to obtain a modified query; generate an alternative query based on the modified query using a masked language model (MLM), wherein the alternative query includes an alternative phrase in place of the target phrase that is consistent with a context of the target phrase; and retrieve a search result based on the alternative query.

Abstract: Systems and methods for query processing are described. Embodiments of the present disclosure identify an original query; generate a plurality of expanded queries by generating a plurality of additional phrases based on the original query using a causal language model (CLM) and augmenting the original query with each of the plurality of additional phrases, respectively; and provide a plurality of images in response to the original query, wherein the plurality of images are associated with the plurality of expanded queries, respectively.

Abstract: Systems and methods for image exploration are provided. One aspect of the systems and methods includes identifying a set of images; reducing the set of images to obtain a representative set of images that is distributed throughout the set of images by removing a neighbor image based on a proximity of the neighbor image to an image of the representative set of images; arranging the representative set of images in a grid structure using a self-sorting map (SSM) algorithm; and displaying a portion of the representative set of images based on the grid structure.

Abstract: Understanding emojis in the context of online experiences is described. In at least some embodiments, text input is received and a vector representation of the text input is computed. Based on the vector representation, one or more emojis that correspond to the vector representation of the text input are ascertained and a response is formulated that includes at least one of the one or more emojis. In other embodiments, input from a client machine is received. The input includes at least one emoji. A computed vector representation of the emoji is used to look for vector representations of words or phrases that are close to the computed vector representation of the emoji. At least one of the words or phrases is selected and at least one task is performed using the selected word(s) or phrase(s).

Abstract: Methods, systems, and computer programs are presented for adding new features to a network service. A method includes receiving an image depicting an object of interest. A category set is determined for the object of interest and an image signature is generated for the image. Using the category set and the image signature, the method identifies a set of publications within a publication database and assigns a rank to each publication. The method causes presentation of the ranked list of publications at a computing device from which the image was received.

Abstract: In various example embodiments, a system and method for determining a spam publication using a spam detection system are presented. The spam detection system receives, from a device, an image of an item and an item attribute for the item. Additionally, the spam detection system extracts an image attribute based on the received image, and compares the item attribute and the image attribute. Moreover, the spam detection system calculates a confidence score based on the comparison. Furthermore, the spam detection system determines that the item attribute is incorrect based on the confidence score transgressing a predetermined threshold. In response to the determination that the item attribute is incorrect, the spam detection system causes presentation, on a display of the device, of a notification.

Abstract: Systems and methods for image processing are described. Embodiments of the present disclosure receive a first image depicting a scene and a second image that includes a style; segment the first image to obtain a first segment and a second segment, wherein the first segment has a shape of an object in the scene; apply a style transfer network to the first segment and the second image to obtain a first image part, wherein the first image part has the shape of the object and the style from the second image; combine the first image part with a second image part corresponding to the second segment to obtain a combined image; and apply a lenticular effect to the combined image to obtain an output image.

Abstract: Systems, methods, and computer program products for identifying a relevant candidate product in an electronic marketplace. Embodiments perform a visual similarity comparison between candidate product image visual content and input query image visual content, process formal and informal natural language user inputs, and coordinate aggregated past user interactions with the marketplace stored in a knowledge graph. Visually similar items and their corresponding product categories, aspects, and aspect values can determine suggested candidate products without discernible delay during a multi-turn user dialog. The user can then refine the search for the most relevant items available for purchase by providing responses to machine-generated prompts that are based on the initial search results from visual, voice, and/or text inputs. An intelligent online personal assistant can thus guide a user to the most relevant candidate product more efficiently than existing search tools.

Abstract: Systems and methods for image processing, and specifically for generating object-agnostic image representations, are described. Embodiments of the present disclosure receive a training image including a foreground object and a background, remove the foreground object from the training image to obtain a modified training image, inpaint a portion of the modified training image corresponding to the foreground object to obtain an inpainted training image, encode the training image and the inpainted training image using a machine learning model to obtain an encoded training image and an encoded inpainted training image, and update parameters of the machine learning model based on the encoded training image and the encoded inpainted training image.

Abstract: Methods, systems, and computer programs are presented for adding new features to a network service. A method includes receiving an image depicting an object of interest. A category set is determined for the object of interest and an image signature is generated for the image. Using the category set and the image signature, the method identifies a set of publications within a publication database and assigns a rank to each publication. The method causes presentation of the ranked list of publications at a computing device from which the image was received.

Abstract: Techniques for generating a novel image using tokenized image representations are disclosed. In some embodiments, a method of generating the novel image includes generating, via a first machine learning model, a first sequence of coded representations of a first image having one or more features; generating, via a second machine learning model, a second sequence of coded representations of a sketch image having one or more edge features associated with the one or more features; predicting, via a third machine learning model, one or more subsequent coded representations based on the first sequence of coded representations and the second sequence of coded representations; and based on the subsequent coded representations, generating, via the third machine learning model, a first portion of a reconstructed image having one or more image attributes of the first image, and a second portion of the reconstructed image associated with the one or more edge features.

Abstract: Various embodiments use a neural network to analyze images for aspects that characterize the images, to present locations of those aspects on the images, and, additionally, to permit a user to interact with those locations on the images. For example, a user may interact with a visual cue over one of those locations to modify, refine, or filter the results of a visual search, performed on a publication corpus, that uses an input image (e.g., one captured using a mobile device) as a search query.

Abstract: Systems, methods, and computer program products for identifying a candidate product in an electronic marketplace based on a visual comparison between candidate product image visual content and input query image visual content. Embodiments generate and store descriptive image signatures from candidate product images or selected portions of such images. A subsequently calculated visual similarity measure serves as a visual search result score for the candidate product in comparison to an input query image. Any number of images of any number of candidate products may be analyzed, such as for items available for sale in an online marketplace. Image analysis results are stored in a database and made available for subsequent automated on-demand visual comparisons to an input query image. The embodiments enable substantially real time visual based product searching of a potentially vast catalog of items.

Abstract: Systems, methods, and computer program products for identifying a relevant candidate product in an electronic marketplace. Embodiments perform a visual similarity comparison between candidate product image visual content and input query image visual content, process formal and informal natural language user inputs, and coordinate aggregated past user interactions with the marketplace stored in a knowledge graph. Visually similar items and their corresponding product categories, aspects, and aspect values can determine suggested candidate products without discernible delay during a multi-turn user dialog. The user can then refine the search for the most relevant items available for purchase by providing responses to machine-generated prompts that are based on the initial search results from visual, voice, and/or text inputs. An intelligent online personal assistant can thus guide a user to the most relevant candidate product more efficiently than existing search tools.

Abstract: In various example embodiments, a system and method for determining a spam publication using a spam detection system are presented. The spam detection system receives, from a device, an image of an item and an item attribute for the item. Additionally, the spam detection system extracts an image attribute based on the received image, and compares the item attribute and the image attribute. Moreover, the spam detection system calculates a confidence score based on the comparison. Furthermore, the spam detection system determines that the item attribute is incorrect based on the confidence score transgressing a predetermined threshold. In response to the determination that the item attribute is incorrect, the spam detection system causes presentation, on a display of the device, of a notification.

Abstract: Various methods and systems for determining a dominant object of a query and employing the dominant object to provide enhanced search services are discussed. A query is segmented into a set of n_grams. Entity extraction and resolution (EER) methods are employed to determine implicit and explicit aspects for each n_gram. N_grams that include explicit aspects are pruned from the set of n_grams and a pruned set of candidate n_grams is generated from the non-pruned n_grams. Knowledge graphs are employed to generate a ranked list of associated categories for each candidate n_gram. A ranked list of categories associated with the un-segmented query is generated based on knowledge graphs. The candidate n_gram with the highest ranked associated category that is also a highly ranked category associated with the un-segmented query is selected as the dominant object of the query. Enhanced search results are provided based on the determined dominant object.

Abstract: Systems and methods for product retrieval are described. One or more aspects of the systems and methods include receiving a query that includes a text description of a product associated with a brand; identifying the product based on the query by comparing the text description to a product embedding of the product, wherein the product embedding is based on a brand embedding of the brand; and displaying product information for the product in response to the query, wherein the product information includes the brand.

Abstract: Systems and methods for image processing are configured. Embodiments of the present disclosure identify target style attributes and target structure attributes for a composite image; generate a matrix of composite feature tokens based on the target style attributes and the target structure attributes, wherein subsequent feature tokens of the matrix of composite feature tokens are sequentially generated based on previous feature tokens of the matrix of composite feature tokens according to a linear ordering of the matrix of composite feature tokens; and generate the composite image based on the matrix of composite feature tokens, wherein the composite image includes the target style attributes and the target structure attributes.

Abstract: Systems, methods, and computer program products for identifying a candidate product in an electronic marketplace based on a visual comparison between candidate product image visual content and input query image visual content. Embodiments generate and store descriptive image signatures from candidate product images or selected portions of such images. A subsequently calculated visual similarity measure serves as a visual search result score for the candidate product in comparison to an input query image. Any number of images of any number of candidate products may be analyzed, such as for items available for sale in an online marketplace. Image analysis results are stored in a database and made available for subsequent automated on-demand visual comparisons to an input query image. The embodiments enable substantially real time visual based product searching of a potentially vast catalog of items.

Abstract: Systems and methods for image processing are described. Embodiments of the present disclosure identify a plurality of candidate concepts in a knowledge graph (KG) that correspond to an image tag of an image; generate an image embedding of the image using a multi-modal encoder; generate a concept embedding for each of the plurality of candidate concepts using the multi-modal encoder; select a matching concept from the plurality of candidate concepts based on the image embedding and the concept embedding; and generate association data between the image and the matching concept.