Oregon State University

About: Oregon State University is a education organization based out in Corvallis, Oregon, United States. It is known for research contribution in the topics: Population & Gene. The organization has 28192 authors who have published 64044 publications receiving 2634108 citations. The organization is also known as: Oregon Agricultural College & OSU.

Drosophila suzukii (Diptera: Drosophilidae): Invasive Pest of Ripening Soft Fruit Expanding its Geographic Range and Damage Potential

Abstract: Spotted wing drosophila, Drosophila suzukii Matsumura, a native of eastern and southeastern Asia, is a pest of small and stone fruits. First detected in California in 2008, the insect is now found across the Pacific Coast states. Its penchant for attacking healthy, ripening fruit (as opposed to overripe and rotting fruit favored by other so-called “vinegar flies”) makes it a potential economic threat to a host of soft- and thin-skinned fruit crops including cherry, raspberry, blackberry, blueberry, strawberry, peach, plums, pluots, nectarines, juice grape, table grape, and wine grape. Coordinated research projects to determine host preference, seasonal phenology, biology, and management options are taking place among entomologists in Washington, Oregon, and California. A description of the pest and initial findings on its biology, life history, known and expected geographic range, management and monitoring techniques, and economic considerations are presented and discussed.

Benchmarking Neural Network Robustness to Common Corruptions and Perturbations

Abstract: In this paper we establish rigorous benchmarks for image classifier robustness. Our first benchmark, ImageNet-C, standardizes and expands the corruption robustness topic, while showing which classifiers are preferable in safety-critical applications. Then we propose a new dataset called ImageNet-P which enables researchers to benchmark a classifier's robustness to common perturbations. Unlike recent robustness research, this benchmark evaluates performance on common corruptions and perturbations not worst-case adversarial perturbations. We find that there are negligible changes in relative corruption robustness from AlexNet classifiers to ResNet classifiers. Afterward we discover ways to enhance corruption and perturbation robustness. We even find that a bypassed adversarial defense provides substantial common perturbation robustness. Together our benchmarks may aid future work toward networks that robustly generalize.