University of New South Wales

About: University of New South Wales is a education organization based out in Sydney, New South Wales, Australia. It is known for research contribution in the topics: Population & Poison control. The organization has 51197 authors who have published 153634 publications receiving 4880608 citations. The organization is also known as: UNSW & UNSW Australia.

Impacts of shrub encroachment on ecosystem structure and functioning: towards a global synthesis

Abstract: Encroachment of woody plants into grasslands has generated considerable interest among ecologists. Syntheses of encroachment effects on ecosystem processes have been limited in extent and confined largely to pastoral land uses or particular geographical regions. We used univariate analyses, meta-analysis and structural equation modelling to test the propositions that (1) shrub encroachment does not necessarily lead to declines in ecosystem functions and (2) shrub traits influence the functional outcome of encroachment. Analyses of 43 ecosystem attributes from 244 case studies worldwide showed that some attributes consistently increased with encroachment (e.g. soil C, N), and others declined (e.g. grass cover, pH), but most exhibited variable responses. Traits of shrubs were associated with significant, though weak, structural and functional outcomes of encroachment. Our review revealed that encroachment had mixed effects on ecosystem structure and functioning at global scales, and that shrub traits influence the functional outcome of encroachment. Thus, a simple designation of encroachment as a process leading to functionally, structurally or contextually degraded ecosystems is not supported by a critical analysis of existing literature. Our results highlight that the commonly established link between shrub encroachment and degradation is not universal.

Monte-Carlo Planning in Large POMDPs

Abstract: This paper introduces a Monte-Carlo algorithm for online planning in large POMDPs. The algorithm combines a Monte-Carlo update of the agent's belief state with a Monte-Carlo tree search from the current belief state. The new algorithm, POMCP, has two important properties. First, Monte-Carlo sampling is used to break the curse of dimensionality both during belief state updates and during planning. Second, only a black box simulator of the POMDP is required, rather than explicit probability distributions. These properties enable POMCP to plan effectively in significantly larger POMDPs than has previously been possible. We demonstrate its effectiveness in three large POMDPs. We scale up a well-known benchmark problem, rocksample, by several orders of magnitude. We also introduce two challenging new POMDPs: 10 x 10 battleship and partially observable PacMan, with approximately 1018 and 1056 states respectively. Our Monte-Carlo planning algorithm achieved a high level of performance with no prior knowledge, and was also able to exploit simple domain knowledge to achieve better results with less search. POMCP is the first general purpose planner to achieve high performance in such large and unfactored POMDPs.