Machine learning goes wild: Using data from captive individuals to infer wildlife behaviours

Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration

Observing animals in the wild often poses extreme challenges, but animal-borne accelerometers are increasingly revealing unobservable behaviours. Automated machine learning streamlines behaviour identification from the substantial datasets generated during multi-animal, long-term studies; however, the accuracy of such models depends on the qualities of the training data. We examined how data processing influenced the predictive accuracy of random forest (RF) models, leveraging the easily observed domestic cat (Felis catus) as a model organism for terrestrial mammalian behaviours. Nine indoor domestic cats were equipped with collar-mounted tri-axial accelerometers, and behaviours were recorded alongside video footage. From this calibrated data, eight datasets were derived with (i) additional descriptive variables, (ii) altered frequencies of acceleration data (40 Hz vs. a mean over 1 s) and (iii) standardised durations of different behaviours. These training datasets were used to generate RF models that were validated against calibrated cat behaviours before identifying the behaviours of five free-ranging tag-equipped cats. These predictions were compared to those identified manually to validate the accuracy of the RF models for free-ranging animal behaviours. RF models accurately predicted the behaviours of indoor domestic cats (F-measure up to 0.96) with discernible improvements observed with post-data-collection processing. Additional variables, standardised durations of behaviours and higher recording frequencies improved model accuracy. However, prediction accuracy varied with different behaviours, where high-frequency models excelled in identifying fast-paced behaviours (e.g. locomotion), whereas lower-frequency models (1 Hz) more accurately identified slower, aperiodic behaviours such as grooming and feeding, particularly when examining free-ranging cat behaviours. While RF modelling offered a robust means of behaviour identification from accelerometer data, field validations were important to validate model accuracy for free-ranging individuals. Future studies may benefit from employing similar data processing methods that enhance RF behaviour identification accuracy, with extensive advantages for investigations into ecology, welfare and management of wild animals.

Advances in biologging can identify nuanced energetic costs and gains in predators

Foraging is a key driver of animal movement patterns, with specific challenges for predators which must search for mobile prey. These patterns are increasingly impacted by global changes, principally in land use and climate. Understanding the degree of flexibility in predator foraging and social strategies is pertinent to wildlife conservation under global change, including potential top-down effects on wider ecosystems. Here we propose key future research directions to better understand foraging strategies and social flexibility in predators. In particular, rapid continued advances in biologging technology are helping to record and understand dynamic behavioural and movement responses of animals to environmental changes, and their energetic consequences. Data collection can be optimised by calibrating behavioural interpretation methods in captive settings and strategic tagging decisions within and between social groups. Importantly, many species' social systems are increasingly being found to be more flexible than originally described in the literature, which may be more readily detectable through biologging approaches than behavioural observation. Integrating the effects of the physical landscape and biotic interactions will be key to explaining and predicting animal movements and energetic balance in a changing world.

Supervised versus unsupervised approaches to classification of accelerometry data

Sophisticated animal-borne sensor systems are increasingly providing novel insight into how animals behave and move. Despite their widespread use in ecology, the diversity and expanding quality and quantity of data they produce have created a need for robust analytical methods for biological interpretation. Machine learning tools are often used to meet this need. However, their relative effectiveness is not well known and, in the case of unsupervised tools, given that they do not use validation data, their accuracy can be difficult to assess. We evaluated the effectiveness of supervised (n = 6), semi-supervised (n = 1), and unsupervised (n = 2) approaches to analyzing accelerometry data collected from critically endangered California condors (Gymnogyps californianus). Unsupervised K-means and EM (expectation-maximization) clustering approaches performed poorly, with adequate classification accuracies of <0.8 but very low values for kappa statistics (range: -0.02 to 0.06). The semi-supervised nearest mean classifier was moderately effective at classification, with an overall classification accuracy of 0.61 but effective classification only of two of the four behavioral classes. Supervised random forest (RF) and k-nearest neighbor (kNN) machine learning models were most effective at classification across all behavior types, with overall accuracies >0.81. Kappa statistics were also highest for RF and kNN, in most cases substantially greater than for other modeling approaches. Unsupervised modeling, which is commonly used for the classification of a priori-defined behaviors in telemetry data, can provide useful information but likely is instead better suited to post hoc definition of generalized behavioral states. This work also shows the potential for substantial variation in classification accuracy among different machine learning approaches and among different metrics of accuracy. As such, when analyzing biotelemetry data, best practices appear to call for the evaluation of several machine learning techniques and several measures of accuracy for each dataset under consideration.

Year-round behavioural time budgets of common woodpigeons inferred from acceleration data using machine learning

Abstract

Accelerometers capture rapid changes in animal motion, and the analysis of large quantities of such data using machine learning algorithms enables the inference of broad animal behaviour categories such as foraging, flying, and resting over long periods of time. We deployed GPS-GSM/GPRS trackers with tri-axial acceleration sensors on common woodpigeons (Columba palumbus) from Hesse, Germany (forest and urban birds) and from Lisbon, Portugal (urban park). We used three machine learning algorithms, Random Forest, Support Vector Machine, and Extreme Gradient Boosting, to classify the main behaviours of the birds, namely foraging, flying, and resting and calculated time budgets over the breeding and winter season. Woodpigeon time budgets varied between seasons, with more foraging time during the breeding season than in winter. Also, woodpigeons from different sites showed differences in the time invested in foraging. The proportion of time woodpigeons spent foraging was lowest in the forest habitat from Hesse, higher in the urban habitat of Hesse, and highest in the urban park in Lisbon. The time budgets we recorded contrast to previous findings in woodpigeons and reaffirm the importance of considering different populations to fully understand the behaviour and adaptation of a particular species to a particular environment. Furthermore, the differences in the time budgets of Woodpigeons from this study and previous ones might be related to environmental change and merit further attention and the future investigation of energy budgets.

Significance statement

In this study we took advantage of accelerometer technology and machine learning methods to investigate year-round behavioural time budgets of wild common woodpigeons (Columba palumbus). Our analysis focuses on identifying coarse-scale behaviours (foraging, flying, resting) using various machine learning algorithms. Woodpigeon time budgets varied between seasons and among sites. Particularly interesting is the result showing that urban woodpigeons spend more time foraging than forest conspecifics. Our study opens an opportunity to further investigate and understand how a successful bird species such as the woodpigeon copes with increasing environmental change and urbanisation. The increase in the proportion of time devoted to foraging might be one of the behavioural mechanisms involved but opens questions about the costs associated to such increase in terms of other important behaviours.

Enhanced Classification of Dog Activities with Quaternion-Based Fusion Approach on High-Dimensional Raw Data from Wearable Sensors

The employment of machine learning algorithms to the data provided by wearable movement sensors is one of the most common methods to detect pets' behaviors and monitor their well-being. However, defining features that lead to highly accurate behavior classification is quite challenging. To address this problem, in this study we aim to classify six main dog activities (standing, walking, running, sitting, lying down, and resting) using high-dimensional sensor raw data. Data were received from the accelerometer and gyroscope sensors that are designed to be attached to the dog's smart costume. Once data are received, the module computes a quaternion value for each data point that provides handful features for classification. Next, to perform the classification, we used several supervised machine learning algorithms, such as the Gaussian naïve Bayes (GNB), Decision Tree (DT), K-nearest neighbor (KNN), and support vector machine (SVM). In order to evaluate the performance, we finally compared the proposed approach's F-score accuracies with the accuracy of classic approach performance, where sensors' data are collected without computing the quaternion value and directly utilized by the model. Overall, 18 dogs equipped with harnesses participated in the experiment. The results of the experiment show a significantly enhanced classification with the proposed approach. Among all the classifiers, the GNB classification model achieved the highest accuracy for dog behavior. The behaviors are classified with F-score accuracies of 0.94, 0.86, 0.94, 0.89, 0.95, and 1, respectively. Moreover, it has been observed that the GNB classifier achieved 93% accuracy on average with the dataset consisting of quaternion values. In contrast, it was only 88% when the model used the dataset from sensors' data.

The Future of Artificial Intelligence in Monitoring Animal Identification, Health, and Behaviour

Simple Summary

Due to climate change and human interference, many species are now without habitats and on the brink of extinction. Zoos and other conservation spaces allow for non-human animal preservation and public education about endangered species and ecosystems. Monitoring the health and well-being of animals in care, while providing species-specific environments, is critical for zoo and conservation staff. In order to best provide such care, keepers and researchers need to gather as much information as possible about individual animals and species as a whole. This paper focuses on existing technology to monitor animals, providing a review on the history of technology, including recent technological advancements and current limitations. Subsequently, we provide a brief introduction to our proposed novel computer software: an artificial intelligence software capable of unobtrusively and non-invasively tracking individuals’ location, estimating position, and analyzing behaviour. This innovative technology is currently being trained with orangutans at the Toronto Zoo and will allow for mass data collection, permitting keepers and researchers to closely monitor individual animal welfare, learn about the variables impacting behaviour and provide additional enrichment or interventions accordingly.

Abstract

With many advancements, technologies are now capable of recording non-human animals’ location, heart rate, and movement, often using a device that is physically attached to the monitored animals. However, to our knowledge, there is currently no technology that is able to do this unobtrusively and non-invasively. Here, we review the history of technology for use with animals, recent technological advancements, current limitations, and a brief introduction to our proposed novel software. Canadian tech mogul EAIGLE Inc. has developed an artificial intelligence (AI) software solution capable of determining where people and assets are within public places or attractions for operational intelligence, security, and health and safety applications. The solution also monitors individual temperatures to reduce the potential spread of COVID-19. This technology has been adapted for use at the Toronto Zoo, initiated with a focus on Sumatran orangutans (Pongo abelii) given the close physical similarity between orangutans and humans as great ape species. This technology will be capable of mass data collection, individual identification, pose estimation, behaviour monitoring and tracking orangutans’ locations, in real time on a 24/7 basis, benefitting both zookeepers and researchers looking to review this information.

Red Foxes in the Filing Cabinet: Günter Tembrock's Image Collection and Media Use in Mid-Century Ethology

This paper considers the epistemic career of visual media in ethology in the mid-20th century. Above all, ethologists claimed close contact with research animals and drew scientific evidence from these human-animal communities, particularly in public relations. However, if we look into the toolboxes of comparative behavioral biologists, it becomes evident that scientifically valid research results were primarily obtained by experimenting with model images. These visual specimens tell a technical story of the methodological requirements in behavioral science necessary to bridge everyday observations between the laboratory and the field. By neutralizing individual traces of animal bodies as well as their observers, they prompted the abstraction of ethological hypotheses. The case study of East-German biologist Günter Tembrock (1918-2011), who maintained his own collection of newspaper clippings, drawings, photographs, and films, offers a new perspective on the methodological development of this field. Furthermore, this article contributes to a scholarly discussion geared toward expanding the spaces of ethological research. My analysis of the image collections of the Forschungsstätte für Tierpsychologie presents the archive as a relevant site of study in the history of ethology.

Spatial Ecology and Diel Activity of European Wildcat (Felis silvestris) in a Protected Lowland Area in Northern Greece

Simple Summary

The European wildcat is a species of conservation concern protected across its range in Europe, where it occurs in five discontinuous populations. The Balkan population has received little attention, making it difficult to assess whether the ecological traits reported for other populations apply also to this population. This hampers the development of targeted conservation measures. The present study reports the first findings on the spatial ecology and daily activity pattern of wildcats in a human modified landscape in Greece, using cutting edge data loggers attached to wildcat collars. In Greece, wildcat home range sizes are within the range of those reported for other populations. Male wildcats are active primarily at night and near dawn and dusk, as is typical for the species. However, the activity of some females varied from this pattern in late spring, in ways expected for wildcats, caring for offspring at a den. Overall, our findings help fill the ecological knowledge gap of the species in Greece and suggest that lowland agricultural areas with patches of natural habitats may have a significant role in the future conservation of the species.

Abstract

The Balkan populations of the European wildcat are among the least studied. This study reports the first findings on the spatial ecology and activity pattern of the wildcat in Greece and compares them to those of better studied northern populations. We fitted five wildcats (two males, three females) with collars containing GPS and accelerometer loggers (E-obs 1A) and collected data from fall to early summer. All animals moved within a mosaic of lowland agricultural fields, woodland patches, riparian forests and wetlands near the banks of a lake. The trapping rate was the highest reported for the species. The home range sizes, estimated using Brownian bridge movement models, ranged from 0.94 to 3.08 km² for females and from 1.22 to 4.43 km² for males. Based on overall dynamic body acceleration (ODBA) values estimated from the accelerometer data, the diel activity of male wildcats followed the species’ typical nocturnal pattern with crepuscular peaks. Female activity varied seasonally, at times being cathemeral. We found only weak effects of environmental variables on wildcat activity, and no significant difference in the activity in open versus forested areas. Our findings suggest that human modified landscapes can play a significant role in the conservation of this typically forest-associated species.

Using Machine Learning for Remote Behaviour Classification-Verifying Acceleration Data to Infer Feeding Events in Free-Ranging Cheetahs

Behavioural studies of elusive wildlife species are challenging but important when they are threatened and involved in human-wildlife conflicts. Accelerometers (ACCs) and supervised machine learning algorithms (MLAs) are valuable tools to remotely determine behaviours. Here we used five captive cheetahs in Namibia to test the applicability of ACC data in identifying six behaviours by using six MLAs on data we ground-truthed by direct observations. We included two ensemble learning approaches and a probability threshold to improve prediction accuracy. We used the model to then identify the behaviours in four free-ranging cheetah males. Feeding behaviours identified by the model and matched with corresponding GPS clusters were verified with previously identified kill sites in the field. The MLAs and the two ensemble learning approaches in the captive cheetahs achieved precision (recall) ranging from 80.1% to 100.0% (87.3% to 99.2%) for resting, walking and trotting/running behaviour, from 74.4% to 81.6% (54.8% and 82.4%) for feeding behaviour and from 0.0% to 97.1% (0.0% and 56.2%) for drinking and grooming behaviour. The model application to the ACC data of the free-ranging cheetahs successfully identified all nine kill sites and 17 of the 18 feeding events of the two brother groups. We demonstrated that our behavioural model reliably detects feeding events of free-ranging cheetahs. This has useful applications for the determination of cheetah kill sites and helping to mitigate human-cheetah conflicts.

Limitations of using surrogates for behaviour classification of accelerometer data: refining methods using random forest models in Caprids

BACKGROUND

Animal-attached devices can be used on cryptic species to measure their movement and behaviour, enabling unprecedented insights into fundamental aspects of animal ecology and behaviour. However, direct observations of subjects are often still necessary to translate biologging data accurately into meaningful behaviours. As many elusive species cannot easily be observed in the wild, captive or domestic surrogates are typically used to calibrate data from devices. However, the utility of this approach remains equivocal.

METHODS

Here, we assess the validity of using captive conspecifics, and phylogenetically-similar domesticated counterparts (surrogate species) for calibrating behaviour classification. Tri-axial accelerometers and tri-axial magnetometers were used with behavioural observations to build random forest models to predict the behaviours. We applied these methods using captive Alpine ibex (Capra ibex) and a domestic counterpart, pygmy goats (Capra aegagrus hircus), to predict the behaviour including terrain slope for locomotion behaviours of captive Alpine ibex.

RESULTS

Behavioural classification of captive Alpine ibex and domestic pygmy goats was highly accurate (> 98%). Model performance was reduced when using data split per individual, i.e., classifying behaviour of individuals not used to train models (mean ± sd = 56.1 ± 11%). Behavioural classifications using domestic counterparts, i.e., pygmy goat observations to predict ibex behaviour, however, were not sufficient to predict all behaviours of a phylogenetically similar species accurately (> 55%).

CONCLUSIONS

We demonstrate methods to refine the use of random forest models to classify behaviours of both captive and free-living animal species. We suggest there are two main reasons for reduced accuracy when using a domestic counterpart to predict the behaviour of a wild species in captivity; domestication leading to morphological differences and the terrain of the environment in which the animals were observed. We also identify limitations when behaviour is predicted in individuals that are not used to train models. Our results demonstrate that biologging device calibration needs to be conducted using: (i) with similar conspecifics, and (ii) in an area where they can perform behaviours on terrain that reflects that of species in the wild.

Deep learning-based pose estimation for African ungulates in zoos

The description and analysis of animal behavior over long periods of time is one of the most important challenges in ecology. However, most of these studies are limited due to the time and cost required by human observers. The collection of data via video recordings allows observation periods to be extended. However, their evaluation by human observers is very time-consuming. Progress in automated evaluation, using suitable deep learning methods, seems to be a forward-looking approach to analyze even large amounts of video data in an adequate time frame.In this study, we present a multistep convolutional neural network system for detecting three typical stances of African ungulates in zoo enclosures which works with high accuracy. An important aspect of our approach is the introduction of model averaging and postprocessing rules to make the system robust to outliers.Our trained system achieves an in-domain classification accuracy of >0.92, which is improved to >0.96 by a postprocessing step. In addition, the whole system performs even well in an out-of-domain classification task with two unknown types, achieving an average accuracy of 0.93. We provide our system at https://github.com/Klimroth/Video-Action-Classifier-for-African-Ungulates-in-Zoos/tree/main/mrcnn_based so that interested users can train their own models to classify images and conduct behavioral studies of wildlife.The use of a multistep convolutional neural network for fast and accurate classification of wildlife behavior facilitates the evaluation of large amounts of image data in ecological studies and reduces the effort of manual analysis of images to a high degree. Our system also shows that postprocessing rules are a suitable way to make species-specific adjustments and substantially increase the accuracy of the description of single behavioral phases (number, duration). The results in the out-of-domain classification strongly suggest that our system is robust and achieves a high degree of accuracy even for new species, so that other settings (e.g., field studies) can be considered.

Wildlife Monitoring on the Edge: A Performance Evaluation of Embedded Neural Networks on Microcontrollers for Animal Behavior Classification

Monitoring animals’ behavior living in wild or semi-wild environments is a very interesting subject for biologists who work with them. The difficulty and cost of implanting electronic devices in this kind of animals suggest that these devices must be robust and have low power consumption to increase their battery life as much as possible. Designing a custom smart device that can detect multiple animal behaviors and that meets the mentioned restrictions presents a major challenge that is addressed in this work. We propose an edge-computing solution, which embeds an ANN in a microcontroller that collects data from an IMU sensor to detect three different horse gaits. All the computation is performed in the microcontroller to reduce the amount of data transmitted via wireless radio, since sending information is one of the most power-consuming tasks in this type of devices. Multiples ANNs were implemented and deployed in different microcontroller architectures in order to find the best balance between energy consumption and computing performance. The results show that the embedded networks obtain up to 97.96% ± 1.42% accuracy, achieving an energy efficiency of 450 Mops/s/watt.

Behaviour Classification on Giraffes (Giraffa camelopardalis) Using Machine Learning Algorithms on Triaxial Acceleration Data of Two Commonly Used GPS Devices and Its Possible Application for Their Management and Conservation

Averting today's loss of biodiversity and ecosystem services can be achieved through conservation efforts, especially of keystone species. Giraffes (Giraffa camelopardalis) play an important role in sustaining Africa's ecosystems, but are 'vulnerable' according to the IUCN Red List since 2016. Monitoring an animal's behavior in the wild helps to develop and assess their conservation management. One mechanism for remote tracking of wildlife behavior is to attach accelerometers to animals to record their body movement. We tested two different commercially available high-resolution accelerometers, e-obs and Africa Wildlife Tracking (AWT), attached to the top of the heads of three captive giraffes and analyzed the accuracy of automatic behavior classifications, focused on the Random Forests algorithm. For both accelerometers, behaviors of lower variety in head and neck movements could be better predicted (i.e., feeding above eye level, mean prediction accuracy e-obs/AWT: 97.6%/99.7%; drinking: 96.7%/97.0%) than those with a higher variety of body postures (such as standing: 90.7-91.0%/75.2-76.7%; rumination: 89.6-91.6%/53.5-86.5%). Nonetheless both devices come with limitations and especially the AWT needs technological adaptations before applying it on animals in the wild. Nevertheless, looking at the prediction results, both are promising accelerometers for behavioral classification of giraffes. Therefore, these devices when applied to free-ranging animals, in combination with GPS tracking, can contribute greatly to the conservation of giraffes.

Assessment of Machine Learning Models to Identify Port Jackson Shark Behaviours Using Tri-Axial Accelerometers

Movement ecology has traditionally focused on the movements of animals over large time scales, but, with advancements in sensor technology, the focus can become increasingly fine scale. Accelerometers are commonly applied to quantify animal behaviours and can elucidate fine-scale (<2 s) behaviours. Machine learning methods are commonly applied to animal accelerometry data; however, they require the trial of multiple methods to find an ideal solution. We used tri-axial accelerometers (10 Hz) to quantify four behaviours in Port Jackson sharks (Heterodontus portusjacksoni): two fine-scale behaviours (<2 s)-(1) vertical swimming and (2) chewing as proxy for foraging, and two broad-scale behaviours (>2 s-mins)-(3) resting and (4) swimming. We used validated data to calculate 66 summary statistics from tri-axial accelerometry and assessed the most important features that allowed for differentiation between the behaviours. One and two second epoch testing sets were created consisting of 10 and 20 samples from each behaviour event, respectively. We developed eight machine learning models to assess their overall accuracy and behaviour-specific accuracy (one classification tree, five ensemble learners and two neural networks). The support vector machine model classified the four behaviours better when using the longer 2 s time epoch (F-measure 89%; macro-averaged F-measure: 90%). Here, we show that this support vector machine (SVM) model can reliably classify both fine- and broad-scale behaviours in Port Jackson sharks.