Amazon Technologies, Inc.

Miriam Farber

Gerard Guy Medioni

Manoj Aggarwal

Igor Kviatkovsky

Roman Goldenberg

George Leifman

Nadav Israel Bhonker

Lior Zamir

Alon Shoshan

A machine learning system to determine an identity of a user is trained using triplets of ad hoc synthetic data and actual data. The data may comprise multimodal images of a hand. Each triplet comprises an anchor, a positive, and a negative image. Synthetic triplets for different synthesized identities are generated on an ad hoc basis and provided as input during training of the machine learning system. The machine learning system uses a pairwise label-based loss function, such as a triplet loss function during training. Synthetic triplets may be generated to provide more challenging training data, to provide training data for categories that are underrepresented in the actual data, and so forth. The system uses substantially less memory during training, and the synthetic triplets need not be retained further reducing memory use. Ongoing training is supported as new actual triplets become available, and may be supplemented by additional synthetic triplets.

System for training of recognition system using ad hoc training data

A scanner acquires a set of images of a hand of a user to facilitate identification. These images may vary, due to changes in relative position, pose, lighting, obscuring objects such as a sleeve, and so forth. A first neural network determines output data comprising a spatial mask and a feature map for individual images in the set. The output data for two or more images is combined to provide aggregate data that is representative of the two or more images. The aggregate data may then be processed using a second neural network, such as convolutional neural network, to determine an embedding vector. The embedding vector may be stored and associated with a user account. At a later time, images acquired from the scanner may be processed to produce an embedding vector that is compared to the stored embedding vector to identify a user at the scanner.

System for determining embedding from multiple inputs

Described are systems and methods for training machine learning models of an ensemble of models that are de-correlated. For example, two or more machine learning models may be concurrently trained (e.g., co-trained) while adding a decorrelation component to one or both models that decreases the pairwise correlation between the outputs of the models. Unlike traditional approaches, in accordance with the disclosed implementations, only the negative results need to be decorrelated.

Concurrent ensemble model training for open sets

Disclosed are systems and method for training an ensemble of machine learning models with a focus on feature engineering. For example, the training of the models encourages each machine learning model of the ensemble to rely on a different set of input features from the training data samples used to train the machine learning models of the ensemble. However, instead of telling each model explicitly which features to learn, in accordance with the disclosed implementations, ML models of the ensemble may be trained sequentially, with each new model trained to disregard input features learned by previously trained ML models of the ensemble and learn based on other features included in the training data samples.

Sequential ensemble model training for open sets

Images of a hand may be used to identify users. Quality, detail, and so forth of these images may vary. An image is processed to determine a first spatial mask. A first neural network comprising many layers uses the first spatial mask at a first layer and a second spatial mask at a second layer to process images and produce an embedding vector representative of features in the image. The first spatial mask provides information about particular portions of the input image, and is determined by processing the image with an algorithm such as an orientation certainty level (OCL) algorithm. The second spatial mask is determined using unsupervised training and represents weights of particular portions of the input image as represented at the second layer. The use of the masks allows the first neural network to learn to use or disregard particular portions of the image, improving overall accuracy.

Growing community of inventors

Haifa, Israel

Miriam Farber