Amazon.com

About: Amazon.com is a company organization based out in Seattle, Washington, United States. It is known for research contribution in the topics: Service (business) & Service provider. The organization has 13363 authors who have published 17317 publications receiving 266589 citations.

Optimized training of linear machine learning models

Abstract: An indication of a data source to be used to train a linear prediction model is obtained. The model is to generate predictions using respective parameters assigned to a plurality of features derived from observation records of the data source. The parameter values are stored in a parameter vector. During a particular learning iteration of the training phase of the model, one or more features for which parameters are to be added to the parameter vector are identified. In response to a triggering condition, parameters for one or more features are removed from the parameter vector based on an analysis of relative contributions of the features represented in the parameter vector to the model's predictions. After the parameters are removed, at least one parameter is added to the parameter vector.

Speech model retrieval in distributed speech recognition systems

Abstract: Features are disclosed for managing the use of speech recognition models and data in automated speech recognition systems. Models and data may be retrieved asynchronously and used as they are received or after an utterance is initially processed with more general or different models. Once received, the models and statistics can be cached. Statistics needed to update models and data may also be retrieved asynchronously so that it may be used to update the models and data as it becomes available. The updated models and data may be immediately used to re-process an utterance, or saved for use in processing subsequently received utterances. User interactions with the automated speech recognition system may be tracked in order to predict when a user is likely to utilize the system. Models and data may be pre-cached based on such predictions.

Fire disturbance in Amazonian blackwater floodplain forests

Abstract: Background: Large fire scars were detected in floodplain forests of the middle Rio Negro in dry years of the 1990s, using satellite data. Aim: To relate fire years and river level anomalies to the Tropical Pacific Southern Oscillation Index (SOI); to measure fire damage and post-fire succession rate. Methods: We analysed the relationship between the SOI and water levels for 1968–2010. In Landsat images of the 1990s we determined fire scar ages. Using QuickBird images, we measured forest cover loss in 36 of these scars, covering 873 ha. This was validated by field measurements of tree mortality in 15 scars. As a metric of post-fire succession, we compared change in the Enhanced Vegetation Index (EVI) in 10 large floodplain fire scars and 12 terra firme slash-and-burn fallows. Results: SOI explained 32% of the variation in annual low water level. Forest cover loss in the 36 burn scars was 88% ± 8% (mean ± SD), range 67–98%. Post-fire tree mortality was 91%, ±8%, range 75–100%. Correlation between cover loss...

Communication-efficient distributed SGD with Sketching

Abstract: Large-scale distributed training of neural networks is often limited by network bandwidth, wherein the communication time overwhelms the local computation time. Motivated by the success of sketching methods in sub-linear/streaming algorithms, we introduce Sketched SGD, an algorithm for carrying out distributed SGD by communicating sketches instead of full gradients. We show that Sketched SGD has favorable convergence rates on several classes of functions. When considering all communication -- both of gradients and of updated model weights -- Sketched SGD reduces the amount of communication required compared to other gradient compression methods from $\mathcal{O}(d)$ or $\mathcal{O}(W)$ to $\mathcal{O}(\log d)$, where $d$ is the number of model parameters and $W$ is the number of workers participating in training. We run experiments on a transformer model, an LSTM, and a residual network, demonstrating up to a 40x reduction in total communication cost with no loss in final model performance. We also show experimentally that Sketched SGD scales to at least 256 workers without increasing communication cost or degrading model performance.

Facilitating a supply of used items

Abstract: Techniques are described for facilitating a supply of used items on behalf of an entity that desires to acquire such used items, such as a merchant that sells used items to customers. The used items are obtained based on interactions with recipients who receive items that are sold or otherwise provided by one or more merchants (or other suppliers of items), such as by supplying appropriate information and materials to the recipients along with the items. In some situations, when an item is delivered or otherwise supplied to a recipient by a merchant, the merchant also supplies information to the recipient that indicates an offer from a used item acquirer entity to re-acquire the item from the recipient (such as after the recipient has used the item) and shipping materials for use in returning the item to an indicated destination if the recipient accepts the offer.