Abstract: A computer, including a processor and a memory, the memory including instructions to be executed by the processor to train a quantile neural network to input an image and output a lower quantile (LQ) prediction, a median quantile (MQ) prediction and an upper quantile (UQ) prediction corresponding to an object in the image, wherein an LQ loss, an MQ loss and a UQ loss are determined for the LQ prediction, the MQ prediction and the UQ prediction respectively and wherein the LQ loss, the MQ loss and the UQ loss are combined to form a base layer loss and output the quantile neural network.

Abstract: A computer, including a processor and a memory, the memory including instructions to be executed by the processor to train a quantile neural network to input an image and output a lower quantile (LQ) prediction, a median quantile (MQ) prediction and an upper quantile (UQ) prediction corresponding to an object in the image, wherein an LQ loss, an MQ loss and a UQ loss are determined for the LQ prediction, the MQ prediction and the UQ prediction respectively and wherein the LQ loss, the MQ loss and the UQ loss are combined to form a base layer loss and output the quantile neural network.

Abstract: In various example embodiments, a mapping system and method for generating product listings for machine sensed and user specified criteria are presented. In example embodiments, sensor data about an object, and user characteristic information are received. Physical characteristics are extracted from the sensor data and mapped with the user characteristic information and related characteristics to create mapped characteristics. Based on the mapped characteristics, item listings are identified, ranked and presented to the user. The user can subsequently refine the search criteria by adding, subtracting or reweighing the characteristics.