King Abdullah University of Science and Technology

About: King Abdullah University of Science and Technology is a education organization based out in Jeddah, Saudi Arabia. It is known for research contribution in the topics: Catalysis & Membrane. The organization has 6221 authors who have published 22019 publications receiving 625706 citations. The organization is also known as: KAUST.

The future of evapotranspiration: global requirements for ecosystem functioning, carbon and climate feedbacks, agricultural management, and water resources.

Abstract: The fate of the terrestrial biosphere is highly uncertain given recent and projected changes in climate. This is especially acute for impacts associated with changes in drought frequency and intensity on the distribution and timing of water availability. The development of effective adaptation strategies for these emerging threats to food and water security are compromised by limitations in our understanding of how natural and managed ecosystems are responding to changing hydrological and climatological regimes. This information gap is exacerbated by insufficient monitoring capabilities from local to global scales. Here, we describe how evapotranspiration (ET) represents the key variable in linking ecosystem functioning, carbon and climate feedbacks, agricultural management, and water resources, and highlight both the outstanding science and applications questions and the actions, especially from a space-based perspective, necessary to advance them.

Breeding crops to feed 10 billion.

Abstract: Crop improvements can help us to meet the challenge of feeding a population of 10 billion, but can we breed better varieties fast enough? Technologies such as genotyping, marker-assisted selection, high-throughput phenotyping, genome editing, genomic selection and de novo domestication could be galvanized by using speed breeding to enable plant breeders to keep pace with a changing environment and ever-increasing human population.

Multi-Task Learning as Multi-Objective Optimization

Abstract: In multi-task learning, multiple tasks are solved jointly, sharing inductive bias between them. Multi-task learning is inherently a multi-objective problem because different tasks may conflict, necessitating a trade-off. A common compromise is to optimize a proxy objective that minimizes a weighted linear combination of per-task losses. However, this workaround is only valid when the tasks do not compete, which is rarely the case. In this paper, we explicitly cast multi-task learning as multi-objective optimization, with the overall objective of finding a Pareto optimal solution. To this end, we use algorithms developed in the gradient-based multi-objective optimization literature. These algorithms are not directly applicable to large-scale learning problems since they scale poorly with the dimensionality of the gradients and the number of tasks. We therefore propose an upper bound for the multi-objective loss and show that it can be optimized efficiently. We further prove that optimizing this upper bound yields a Pareto optimal solution under realistic assumptions. We apply our method to a variety of multi-task deep learning problems including digit classification, scene understanding (joint semantic segmentation, instance segmentation, and depth estimation), and multi-label classification. Our method produces higher-performing models than recent multi-task learning formulations or per-task training.

Monolayer MoSe2 grown by chemical vapor deposition for fast photodetection.

Abstract: Monolayer molybdenum disulfide (MoS2) has become a promising building block in optoelectronics for its high photosensitivity. However, sulfur vacancies and other defects significantly affect the electrical and optoelectronic properties of monolayer MoS2 devices. Here, highly crystalline molybdenum diselenide (MoSe2) monolayers have been successfully synthesized by the chemical vapor deposition (CVD) method. Low-temperature photoluminescence comparison for MoS2 and MoSe2 monolayers reveals that the MoSe2 monolayer shows a much weaker bound exciton peak; hence, the phototransistor based on MoSe2 presents a much faster response time (<25 ms) than the corresponding 30 s for the CVD MoS2 monolayer at room temperature in ambient conditions. The images obtained from transmission electron microscopy indicate that the MoSe exhibits fewer defects than MoS2. This work provides the fundamental understanding for the differences in optoelectronic behaviors between MoSe2 and MoS2 and is useful for guiding future designs...