Abstract: Provided are computing systems, methods, and platforms that train query processing models, such as large language models, to perform query intent classification tasks by using retrieval augmentation and multi-stage distillation. Unlabeled training examples of queries may be obtained, and a set of the training examples may be augmented with additional feature annotations to generate augmented training examples. A first query processing model may annotate the retrieval augmented queries to generate inferred labels for the augmented training examples. A second query processing model may be trained on the inferred labels, distilling the query processing model that was trained with retrieval augmentation into a non-retrieval augmented query processing model. The second query processing model may annotate the entire set of unlabeled training examples. Another stage of distillation may train a third query processing model using the entire set of unlabeled training examples without retrieval augmentation.

Abstract: Provided are computing systems, methods, and platforms that train query processing models, such as large language models, to perform query intent classification tasks by using retrieval augmentation and multi-stage distillation. Unlabeled training examples of queries may be obtained, and a set of the training examples may be augmented with additional feature annotations to generate augmented training examples. A first query processing model may annotate the retrieval augmented queries to generate inferred labels for the augmented training examples. A second query processing model may be trained on the inferred labels, distilling the query processing model that was trained with retrieval augmentation into a non-retrieval augmented query processing model. The second query processing model may annotate the entire set of unlabeled training examples. Another stage of distillation may train a third query processing model using the entire set of unlabeled training examples without retrieval augmentation.

Abstract: A method for determining a salient region of an image is disclosed. For a plurality of different saliency cue functions, a single saliency value is calculated for each pixel in a plurality of adjacent pixels in an image using the saliency cue function, wherein one of the saliency cue functions is based on whether the pixel is in a region of the image whose colors contrast with the region's background and another of the saliency cue functions is based on a foreground and background color models of the image. A classifier is used to calculate a combined single saliency value for each pixel based on the single saliency values for the pixel. The salient region of the pixels is determined with a subwindow search based on the combined single saliency values.

Abstract: A method for determining a salient region of an image is disclosed. For a plurality of different saliency cue functions, a single saliency value is calculated for each pixel in a plurality of adjacent pixels in an image using the saliency cue function, wherein one of the saliency cue functions is based on whether the pixel is in a region of the image whose colors contrast with the region's background and another of the saliency cue functions is based on a foreground and background color models of the image. A classifier is used to calculate a combined single saliency value for each pixel based on the single saliency values for the pixel. The salient region of the pixels is determined with a subwindow search based on the combined single saliency values.

Abstract: Each training image in a collection of training images is associated with a corresponding mask. A set of training images is selected from the collection as being a match for an input image, based at least in part on a comparison of the input image to each training image in the collection. An output mask is determined from the associated masks of the set of training images. One or more boundaries are determined for an object depicted in the input image using the output mask.

Abstract: A fashion preference of a user is determined based on a user's interaction with a plurality of fashion product content items that individually depict a corresponding fashion product. A recommendation is made to a user of a fashion product based at least in part on the fashion preference of the user.