Pub Date : 2019-02-28DOI: 10.1017/9781108587792.002
{"title":"Law Enforcement for Wildlife Conservation","authors":"","doi":"10.1017/9781108587792.002","DOIUrl":"https://doi.org/10.1017/9781108587792.002","url":null,"abstract":"","PeriodicalId":260699,"journal":{"name":"Artificial Intelligence and Conservation","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128477762","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1900-01-01DOI: 10.1017/9781108587792.010
D. Sheldon, Kevin Winner, Debora Sujono
Population sizes and demographic rates are commonly estimated by fitting probabilistic population models to observational data. However, these models expose a gap in our probabilistic inference toolkit. They often contain count variables to represent unknown population sizes, and, until recently there was no known algorithm to exactly compute a College of Information and Computer Sciences, University of Massachusetts Amherst b Department of Computer Science, Mount Holyoke College 2 D. Sheldon, K. Winner, and D. Sujono the likelihood in models with latent count variables. This chapter summarizes recent advances in the AI research area of probabilistic inference motivated by this gap. We describe how a standard probabilistic inference algorithm, the forward algorithm, can be adapted to use probability generating functions as its internal representation of probability distributions. This leads to the first exact and efficient algorithms for these models. The new algorithms apply to a broad class of population models, are faster than existing approximate algorithms, and lead to improved behavior for estimating parameters of animal populations.
人口规模和人口比率通常是通过将概率人口模型拟合到观测数据来估计的。然而，这些模型暴露了我们概率推理工具箱中的一个缺口。它们通常包含计数变量来表示未知的人口规模，并且，直到最近，还没有已知的算法来精确计算信息与计算机科学学院，马萨诸塞大学阿默斯特分校计算机科学系，Mount Holyoke学院2 D. Sheldon, K. Winner和D. Sujono具有潜在计数变量的模型中的可能性。本章总结了基于这一差距的概率推理的人工智能研究领域的最新进展。我们描述了一个标准的概率推理算法，即前向算法，如何适应使用概率生成函数作为其概率分布的内部表示。这导致了这些模型的第一个精确和有效的算法。新算法适用于广泛的种群模型，比现有的近似算法更快，并且改进了估计动物种群参数的行为。
{"title":"Probabilistic Inference with Generating Functions for Animal Populations","authors":"D. Sheldon, Kevin Winner, Debora Sujono","doi":"10.1017/9781108587792.010","DOIUrl":"https://doi.org/10.1017/9781108587792.010","url":null,"abstract":"Population sizes and demographic rates are commonly estimated by fitting probabilistic population models to observational data. However, these models expose a gap in our probabilistic inference toolkit. They often contain count variables to represent unknown population sizes, and, until recently there was no known algorithm to exactly compute a College of Information and Computer Sciences, University of Massachusetts Amherst b Department of Computer Science, Mount Holyoke College 2 D. Sheldon, K. Winner, and D. Sujono the likelihood in models with latent count variables. This chapter summarizes recent advances in the AI research area of probabilistic inference motivated by this gap. We describe how a standard probabilistic inference algorithm, the forward algorithm, can be adapted to use probability generating functions as its internal representation of probability distributions. This leads to the first exact and efficient algorithms for these models. The new algorithms apply to a broad class of population models, are faster than existing approximate algorithms, and lead to improved behavior for estimating parameters of animal populations.","PeriodicalId":260699,"journal":{"name":"Artificial Intelligence and Conservation","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123916446","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 1900-01-01DOI: 10.1017/9781108587792.011
Yexiang Xue, C. Gomes
Big data is becoming increasingly important for monitoring large-scale and complex spatial and temporal ecological processes to support informed decision-making in computational sustainability [Gomes 2009]. Historically, data collection methods for conservation typically involved many hours of field work performed by professional researchers, which could not scale up to meet today’s conservation goals. Citizen science programs, on the other hand, engage the general public in the data collection process. With the active participation of thousands of volunteers, citizen science programs have been very successful at collecting rich datasets for conservation, enabling the possibility of conducting ecological surveys across multiple years and at large continental scales. Over the last few years, there have been several successful citizen science programs. For example, in biology, citizen scientists help with bird and arthropod research using eBird [Sullivan et al. 2009] and BugGuide [www.bugguide.net.]. In environmental studies, citizen scientists contribute to monitoring coral bleaching conditions [Marshall et al. 2012]. To attract more people to contribute data, the success of citizen science programs relies on tapping into the intrinsic motivations of citizen scientists. The citizen scientists’ contribution cycle is shown in Figure 11.1. The self-fulfillment brought by participation while contributing to science keeps citizen scientists engaged, in return accelerating the whole contribution cycle. In this chapter, we illustrate effective ways to engage citizen scientists in data collection for conservation (Figure 11.2). Our approaches are motivated by our collaborations with two citizen science programs: the eBird and Nchioto projects. eBird is a citizen science program of the Cornell Lab of Ornithology which engages the general public in bird conservation. To understand the distribution and migration of birds, eBird enlists bird watchers to identify
大数据在监测大规模和复杂的时空生态过程以支持计算可持续性中的明智决策方面变得越来越重要[Gomes 2009]。从历史上看，保护的数据收集方法通常涉及由专业研究人员进行的许多小时的实地工作，这无法扩大规模以满足今天的保护目标。另一方面，公民科学项目让公众参与到数据收集过程中来。在成千上万志愿者的积极参与下，公民科学项目已经非常成功地收集了丰富的保护数据集，使得在多年和大大陆尺度上进行生态调查成为可能。在过去的几年里，有几个成功的公民科学项目。例如，在生物学领域，公民科学家使用eBird [Sullivan et al. 2009]和BugGuide [www.bugguide.net.]来帮助鸟类和节肢动物的研究。在环境研究中，公民科学家为监测珊瑚白化状况做出了贡献[Marshall et al. 2012]。为了吸引更多的人贡献数据，公民科学项目的成功依赖于挖掘公民科学家的内在动机。公民科学家的贡献周期如图11.1所示。参与科学带来的自我实现使公民科学家保持参与，反过来加快了整个贡献周期。在本章中，我们说明了让公民科学家参与保护数据收集的有效方法(图11.2)。我们的方法源于我们与两个公民科学项目的合作:eBird和Nchioto项目。eBird是康奈尔鸟类学实验室的一个公民科学项目，它让公众参与到鸟类保护中来。为了了解鸟类的分布和迁徙，eBird邀请鸟类观察者来识别
{"title":"Engaging Citizen Scientists in Data Collection for Conservation","authors":"Yexiang Xue, C. Gomes","doi":"10.1017/9781108587792.011","DOIUrl":"https://doi.org/10.1017/9781108587792.011","url":null,"abstract":"Big data is becoming increasingly important for monitoring large-scale and complex spatial and temporal ecological processes to support informed decision-making in computational sustainability [Gomes 2009]. Historically, data collection methods for conservation typically involved many hours of field work performed by professional researchers, which could not scale up to meet today’s conservation goals. Citizen science programs, on the other hand, engage the general public in the data collection process. With the active participation of thousands of volunteers, citizen science programs have been very successful at collecting rich datasets for conservation, enabling the possibility of conducting ecological surveys across multiple years and at large continental scales. Over the last few years, there have been several successful citizen science programs. For example, in biology, citizen scientists help with bird and arthropod research using eBird [Sullivan et al. 2009] and BugGuide [www.bugguide.net.]. In environmental studies, citizen scientists contribute to monitoring coral bleaching conditions [Marshall et al. 2012]. To attract more people to contribute data, the success of citizen science programs relies on tapping into the intrinsic motivations of citizen scientists. The citizen scientists’ contribution cycle is shown in Figure 11.1. The self-fulfillment brought by participation while contributing to science keeps citizen scientists engaged, in return accelerating the whole contribution cycle. In this chapter, we illustrate effective ways to engage citizen scientists in data collection for conservation (Figure 11.2). Our approaches are motivated by our collaborations with two citizen science programs: the eBird and Nchioto projects. eBird is a citizen science program of the Cornell Lab of Ornithology which engages the general public in bird conservation. To understand the distribution and migration of birds, eBird enlists bird watchers to identify","PeriodicalId":260699,"journal":{"name":"Artificial Intelligence and Conservation","volume":"355 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121630423","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：