Deep learning‐based methods for individual recognition in small birds

Methods in Ecology and Evolution, Год журнала: 2020, Номер 11(9), С. 1072 - 1085

Опубликована: Июль 26, 2020

Abstract Individual identification is a crucial step to answer many questions in evolutionary biology and mostly performed by marking animals with tags. Such methods are well‐established, but often make data collection analyses time‐consuming, or limit the contexts which can be collected. Recent computational advances, specifically deep learning, help overcome limitations of collecting large‐scale across contexts. However, one bottlenecks preventing application learning for individual need collect identify hundreds thousands individually labelled pictures train convolutional neural networks (CNNs). Here we describe procedures automating training data, generating datasets, CNNs allow birds. We apply our three small bird species, sociable weaver Philetairus socius, great tit Parus major zebra finch Taeniopygia guttata , representing both wild captive first show how images automated, allowing construction datasets consisting per individual. Second, CNN uniquely re‐identify each new images. Third, illustrate general applicability studies animal showing that trained birds collected differ from ones originally used CNNs. Finally, present potential solution solve issues incoming individuals. Overall, work demonstrates feasibility applying state‐of‐the‐art tools birds, laboratory wild. These techniques made possible approaches efficient data. The ability conduct recognition without requiring external markers visually identified human observers represents advance over current methods.

Язык: Английский

---

Meta-analysis challenges a textbook example of status signalling and demonstrates publication bias

eLife, Год журнала: 2018, Номер 7

Опубликована: Ноя. 13, 2018

The status signalling hypothesis aims to explain within-species variation in ornamentation by suggesting that some ornaments signal dominance status. Here, we use multilevel meta-analytic models challenge the textbook example of this hypothesis, black bib male house sparrows (Passer domesticus). We conducted a systematic review, and obtained primary data from published unpublished studies test whether rank is positively associated with size across studies. Contrary previous studies, overall effect (i.e. mean) was small uncertain. Furthermore, found several biases literature further question support available for hypothesis. discuss explanations including pleiotropic, population- context-dependent effects. Our findings call reconsidering established evolutionary behavioural ecology, should stimulate renewed interest understanding ornamental traits.

Язык: Английский

---

The role of replication studies in ecology

Ecology and Evolution, Год журнала: 2020, Номер 10(12), С. 5197 - 5207

Опубликована: Май 22, 2020

Abstract Recent large‐scale projects in other disciplines have shown that results often fail to replicate when studies are repeated. The conditions contributing this problem also present ecology, but there not been any equivalent replication projects. Here, we survey ecologists' understanding of and opinions about studies. majority ecologists our sample considered be important (97%), prevalent enough (91%), worth funding even given limited resources (61%), suitable for publication all journals (62%). However, is a disconnect between enthusiasm the prevalence direct literature which much lower (0.023%: Kelly 2019) than participants' median estimate 10%. This may explained by obstacles participants identified including difficulty conducting publishing them. We conclude offering suggestions how replications could better integrated into ecological research.

Язык: Английский

---

Early-life social environment predicts social network position in wild zebra finches

Proceedings of the Royal Society B Biological Sciences, Год журнала: 2019, Номер 286(1897), С. 20182579 - 20182579

Опубликована: Фев. 27, 2019

Early-life experience can fundamentally shape individual life-history trajectories. Previous research has suggested that exposure to stress during development causes differences in social behaviour later life. In captivity, juvenile zebra finches exposed elevated corticosterone levels were less socially choosy and more central their networks when compared untreated siblings. These extended other aspects of life, with 'stress-exposed' juveniles switching learning strategies males faithfully father's song. However, while this body suggests the impacts early-life could be profound, it remains unknown whether such effects are strong enough expressed under natural conditions. Here, we collected data on associations Australian desert after experimentally manipulating brood sizes. Juveniles from enlarged broods experienced heightened sibling competition, predicted they would express similar patterns stress-treated birds captive study by having more, but differentiated, relationships. We show striking support for consequences developmental network positions, our wild replicating same results 9 out 10 predictions previously tested captivity. Chicks raised foraged greater numbers conspecifics 'choosy' network. Our confirm range variation sufficient predict individuals' trajectories theory highlighting potential importance conditions behaviour.

Язык: Английский

---

Computational Reproducibility in The Wildlife Society's Flagship Journals

Journal of Wildlife Management, Год журнала: 2020, Номер 84(5), С. 1012 - 1017

Опубликована: Март 17, 2020

ABSTRACT Scientific progress depends upon the accumulation of empirical knowledge via reproducible methodology. Although reproducibility is a main tenet scientific method, recent studies have highlighted widespread failures in adherence to this ideal. The goal study was gauge level computational reproducibility, or ability obtain same results using data and analytic methods as original publication, field wildlife science. We randomly selected 80 papers published Journal Wildlife Management Society Bulletin between 1 June 2016 2018. Of those that were suitable for review ( n = 74), we attempted from online repositories directly authors. Forty‐two authors did not respond our requests, further unable 13 other studies. 19 which able complete analysis, judged mostly fully reproducible. conclude with publicly available shared request largely reproducible, but remain concerned about difficulty obtaining recently papers. recommend increased data‐sharing, organization documentation, communication, training advance sciences. © 2020 Authors. by Wiley Periodicals, Inc. on behalf Society.

Язык: Английский

---

Reproducible research into human chemical communication by cues and pheromones: learning from psychology's renaissance

Philosophical Transactions of the Royal Society B Biological Sciences, Год журнала: 2020, Номер 375(1800), С. 20190262 - 20190262

Опубликована: Апрель 19, 2020

Despite the lack of evidence that ‘putative human pheromones' androstadienone and estratetraenol ever were pheromones, almost 60 studies have claimed ‘significant' results. These are quite possibly false positives can be best seen as potential examples ‘reproducibility crisis', sadly common in rest life biomedical sciences, which has many instances whole fields based on positives. Experiments effects olfactory cues behaviour also at risk because they look for subtle but use small sample sizes. Research chemical communication, much it falling within psychology, would benefit from vigorously adopting proposals made by psychologists to enable better, more reliable science, with an emphasis enhancing reproducibility. A key change is adoption study pre-registration and/or Registered Reports will reduce publication bias. As we mammals, communication important other likely our humans may new approaches needed reliably demonstrate them. This article part Theo Murphy meeting issue ‘Olfactory humans’.

Язык: Английский

---

Scrutinizing assortative mating in birds

PLoS Biology, Год журнала: 2019, Номер 17(2), С. e3000156 - e3000156

Опубликована: Фев. 21, 2019

It is often claimed that pair bonds preferentially form between individuals resemble one another. Such assortative mating appears to be widespread throughout the animal kingdom. Yet it unclear whether apparent ubiquity of arises primarily from mate choice ("like attracts like"), which can constrained by same-sex competition for mates; spatial or temporal separation; observer, reporting, publication, search bias. Here, based on a conventional literature search, we find compelling meta-analytical evidence size-assortative in birds (r = 0.178, 95% CI 0.142–0.215, 83 species, 35,591 pairs). However, our analyses reveal this effect vanishes gradually with increased control confounding factors. Specifically, size decreased 42% when used previously unpublished data nine long-term field studies, i.e., free reporting and publication bias 0.103, 0.074–0.132, eight 16,611 Moreover, those data, effectively disappeared both partners were measured independent observers separately space time (mean r 0.018, −0.016–0.057). Likewise, also found no direct experimental test mutual captive populations Zebra finches −0.020, −0.148–0.107, 1,414 These results highlight importance generating unbiased conclusions suggest reported overestimated may not driven preferred mates.

Язык: Английский

---

Wild zebra finches that nest synchronously have long‐term stable social ties

Journal of Animal Ecology, Год журнала: 2019, Номер 90(1), С. 76 - 86

Опубликована: Авг. 13, 2019

Many animals live and breed in colonies, yet, with just a few exceptions, the value of social bonds between colony members has rarely been examined. Social ties are important for group coordination at collective tasks, can facilitate synchronized reproduction among members. Synchronized turn amplify benefits coloniality, such as foraging predator avoidance. We conducted field study to investigate whether individuals replicated colonies is linked strength their bond, these strong maintained beyond reproductive period. PIT-tagged wild zebra finches (Taeniopygia guttata), monitoring over two consecutive years. then used network analysis characterize birds population. show that more timing (and same colony) had significantly stronger both during after than expected by chance. Our long-term sampling also revealed breeders were carried across reveals correspondence breeding structure colony. This suggests synchrony pairs not simple process based on opportunity, but mechanism underpinned complex sociality, which could be other behavioural contexts. The maintenance cross-contextual years have extensive consequences overall life history addition playing key role dynamics colonial breeders.

Язык: Английский

---

Insights found in century-old writings on animal behaviour and some cautions for today

Animal Behaviour, Год журнала: 2020, Номер 164, С. 241 - 249

Опубликована: Март 28, 2020

Язык: Английский

---

Evolution of female coloration: What have we learned from birds in general and blue tits in particular

Advances in the study of behavior, Год журнала: 2020, Номер unknown, С. 123 - 202

Опубликована: Янв. 1, 2020

Язык: Английский

---