Amazon Technologies, Inc.

Valentin Flunkert

Jan Alexander Gasthaus

Andrew Edward Caldwell

Andrew Kent Warfield

James Caleb Kirschner

Seth William Markle

Leon Thrane

Sainath Chowdary Mallidi

Weiwei Cheng

Tim Januschowski

Steffen Rochel

Enrico Sartorello

Yuyang Wang

Miroslav Miladinovic

Francois-Xavier Benoit Marie Aubet

David Jean Bernard Alfred Salinas

Islam Mohamed Hatem A Atta

Arjun Ashok

Jessie E Felix

Konstantinos Benidis

Christoph Bartenstein

David Salinas

Christian Uriel Carmona Perez

An object storage system includes mass storage devices that implement general storage for objects stored in the object storage system and additionally includes other storage devices, such as solid-state drives, that provide higher performance storage access. The object storage system implements a common access interface for accessing both accelerated access objects (who are eligible to have cached copies stored on the higher performance storage devices) and non-accelerated access objects stored in the general storage. The cache is fully managed by the service and no changes are required for client applications to receive accelerated access to objects that are classified as accelerated access objects per a customer configurable acceleration policy for the object or for a bucket in which the object is stored.

Object storage service with managed cache and common interface for accessing accelerated and non-accelerated access objects

Placement decisions may be made to place data in a multi-tenant cache. Usage of multi-tenant cache nodes for performing access requests may be obtained. Usage prediction techniques may be applied to the usage to determine placement decisions for data amongst the multi-tenant cache nodes. Placement actions for the data amongst at the multi-tenant cache nodes may be performed according to the placement decisions.

Making decisions for placing data in a multi-tenant cache

Systems and methods are described for detecting anomalies within data, such as time series data. In one example, unlabeled data, such as time series data, may be obtained. At least one data point, representing an artificial anomaly, may be inserted into the data. The data may then be divided into a number of different windows. The windows may have a fixed size and may at least partially overlap in time. The data contained within different windows may be compared, to each other and to the injected data point, to determine an anomaly score for individual windows. The anomaly score may indicate a likelihood that a given window contains an anomaly. In a specific example, a convolution neural network may be trained based on the data and inserted data points representing anomalies, where a contrastive loss function is used to represent different portions of the data in the neural network.

Neural contrastive anomaly detection

At an artificial intelligence system, a respective feature set is generated from individual text collections pertaining to an item, using a first machine learning model which is trained to perform character-level analysis. Using at least a portion of a second machine learning model, a score associated with a semantic criterion is generated for an item; the training input to the second model is based on the feature sets. A recommendation associated with the item is generated based on the score.

Artificial intelligence system using deep neural networks for pairwise character-level text analysis and recommendations

Techniques are described for a time series probabilistic forecasting framework that combines recurrent neural networks (RNNs) with a flexible, nonparametric representation of the output distribution. The representation is based on the nonparametric quantile function (instead of, for example, a parametric density function) and is trained by minimizing a continuous ranked probability score (CRPS) derived from the quantile function. Unlike methods based on parametric probability density functions and maximum likelihood estimation, the techniques described herein can flexibly adapt to different output distributions without manual intervention. Furthermore, the nonparametric nature of the quantile function provides a significant boost in the approach's robustness, making it more readily applicable to a wide variety of time series datasets.

Probabilistic forecasting with nonparametric quantile functions

At a network-accessible artificial intelligence service for time series predictions, a recurrent neural network model is trained using a plurality of time series of demand observations to generate demand forecasts for various items. A probabilistic demand forecast is generated for a target item using multiple executions of the trained model. Within the training set used for the model, the count of demand observations of the target item may differ from the count of demand observations of other items. A representation of the probabilistic demand forecast may be provided via a programmatic interface.

Recurrent neural network-based artificial intelligence system for time series predictions

Growing community of inventors

Berlin, Germany

Valentin Flunkert