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.

Dependence of hydropower energy generation on forests in the Amazon Basin at local and regional scales.

Abstract: Tropical rainforest regions have large hydropower generation potential that figures prominently in many nations’ energy growth strategies. Feasibility studies of hydropower plants typically ignore the effect of future deforestation or assume that deforestation will have a positive effect on river discharge and energy generation resulting from declines in evapotranspiration (ET) associated with forest conversion. Forest loss can also reduce river discharge, however, by inhibiting rainfall. We used land use, hydrological, and climate models to examine the local “direct” effects (through changes in ET within the watershed) and the potential regional “indirect” effects (through changes in rainfall) of deforestation on river discharge and energy generation potential for the Belo Monte energy complex, one of the world’s largest hydropower plants that is currently under construction on the Xingu River in the eastern Amazon. In the absence of indirect effects of deforestation, simulated deforestation of 20% and 40% within the Xingu River basin increased discharge by 4–8% and 10–12%, with similar increases in energy generation. When indirect effects were considered, deforestation of the Amazon region inhibited rainfall within the Xingu Basin, counterbalancing declines in ET and decreasing discharge by 6–36%. Under business-as-usual projections of forest loss for 2050 (40%), simulated power generation declined to only 25% of maximum plant output and 60% of the industry’s own projections. Like other energy sources, hydropower plants present large social and environmental costs. Their reliability as energy sources, however, must take into account their dependence on forests.

Data Augmentation using Pre-trained Transformer Models

Abstract: Language model based pre-trained models such as BERT have provided significant gains across different NLP tasks. In this paper, we study different types of transformer based pre-trained models such as auto-regressive models (GPT-2), auto-encoder models (BERT), and seq2seq models (BART) for conditional data augmentation. We show that prepending the class labels to text sequences provides a simple yet effective way to condition the pre-trained models for data augmentation. Additionally, on three classification benchmarks, pre-trained Seq2Seq model outperforms other data augmentation methods in a low-resource setting. Further, we explore how different pre-trained model based data augmentation differs in-terms of data diversity, and how well such methods preserve the class-label information.

Automated detection of associations between search criteria and item categories based on collective analysis of user activity data

Abstract: A web site or other database access system (30) provides access to a database (38) in which items are arranged within item categories, such as browse categories of a hierarchical browse tree (36). Actions of users of the system are monitored and recorded to generate user activity data (42) reflective of searches, item selection actions, and possibly other types of actions. A correlation analysis component (44) collectively analyses the user activity to automatically identify associations between specific search criteria, such as specific search strings, and specific item categories. The results of the analysis are stored in a mapping table (40) that is used to suggest specific item categories on search results pages.

Charging an electronic device including traversing at least a portion of a path with an apparatus

Abstract: A robotic device can provide power management and other functionality for various electronic devices. An electronic device and the robotic device can communicate information such as device charge level, relative position and rate of power consumption and can determine an appropriate time at which the robotic device should cause the electronic device to be charged, either by charging the device directly or transporting the device to an appropriate charger. The robotic device can also cause other functions to be performed as well, such as to enable high-bandwidth data transfer or perform resource-intensive activities when in communication with the electronic device. The robotic device can transport the electronic device to specific locations or to specific users, or as otherwise appropriate.

Location-based Coupons and Mobile Devices

Abstract: Techniques for providing electronic coupons to users of mobile devices are described herein. The coupons may be redeemable at certain merchant locations and at certain times. Users may actively select coupons they wish to access from a mobile device. Alternatively, the users may opt in to a system that automatically presents potentially relevant coupons. The relevance of the coupons may be determined based on the distance of the mobile device from a merchant location and/or a time period during which the coupon is valid. The users may activate coupons by logging into a remote computing system from their respective mobile devices. Activated coupons may be transmitted as an electronic signal from a mobile device to a merchant computing device. Alternatively, a display screen of a mobile device may show a barcode or image of the coupon that is read by a machine or employee of the merchant.