Dataset Information

Machine Learning as Ecology.

ABSTRACT: Machine learning methods have had spectacular success on numerous problems. Here we show that a prominent class of learning algorithms - including Support Vector Machines (SVMs) - have a natural interpretation in terms of ecological dynamics. We use these ideas to design new online SVM algorithms that exploit ecological invasions, and benchmark performance using the MNIST dataset. Our work provides a new ecological lens through which we can view statistical learning and opens the possibility of designing ecosystems for machine learning.

SUBMITTER: Howell O

PROVIDER: S-EPMC7779242 | biostudies-literature | 2020 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:In the age of machine learning, building programs that take advantage of the speed and specificity of algorithm development can greatly aid efforts to quantify and interpret changes in animal behavior in response to abiotic environmental factors, like temperature. For both endotherms and ectotherms, temperature can affect everything from daily energy budgets to nesting behaviors. For instance, in birds environmental temperature plays a key role in shaping parental incubation behavior and temperatures experienced by embryos. Recent research indicates that temperatures experienced by embryos affect viability and are important in shaping fitness-related traits in young birds, sparking renewed interest in relationships among environmental factors, parental incubation behavior, and incubation temperature. Incubation behavior of birds can be monitored non-invasively by placing thermal probes into the nest and analyzing temperature fluctuations that occur as parents attend and leave the nest (on- and off-bouts, respectively). When other measures of temperature (e.g., ambient air or operative temperature) are collected simultaneously with incubation temperature it is possible to compare shifts in behavior with environmental changes. To improve analysis of incubation behavior using these large thermal data sets we developed a program, NestIQ, that uses machine learning to guide parameter optimization allowing it to track the behavior of diverse species. NestIQ's algorithm was tested using six species incubating in lab or field scenarios, that exhibit unique incubation patterns. This stand-alone and open source software is operated through a graphical user interface (i.e., no user programming is required) that provides important behavioral and thermal output statistics. Further, measures of environmental temperature can be imported alongside nest temperature into the program, which then reports various attributes of environmental temperature during shifts in parental behavior. This program will improve the ability of avian ecologists to interpret a critical parental care behavior that can be used across diverse incubation scenarios and species. Although specifically designed for quantifying avian incubation, NestIQ has the potential for broader applications, including basking and nesting behaviors of non-avian reptiles in relation to ambient temperature.

Dataset Information

Machine Learning as Ecology.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure