Daniel W. Franks’s research while affiliated with New York University and other places

Low-density social networks can be common in animal societies, even among species generally considered to be highly social. Social network analysis is commonly used to analyse animal societal structure, but edge weight (strength of association between two individuals) estimation methods designed for dense networks can produce biased measures when applied to low-density networks. Frequentist methods suffer when data availability is low, because they contain an inherent flat prior that will accept any possible edge weight value, and contain no uncertainty in their output. Bayesian methods can accept alternative priors, so can provide more reliable edge weights that include a measure of uncertainty, but they can only reduce bias when sensible prior values are selected. Currently, neither accounts for zero-inflation, so they produce edge weight estimates biased towards stronger associations than the true social network, which can be seen through diagnostic plots of data quality against output estimate. We address this by adding zero-inflation to the model, and demonstrate the process using group-based data from a population of male African savannah elephants. We show that the Bayesian approach performs better than the frequentist to reduce the bias caused by these problems, though the Bayesian requires careful consideration of the priors. We recommend the use of a Bayesian framework, but with a conditional prior that allows the modelling of zero-inflation. This reflects the fact that edge weight derivation is a two-step process: i) probability of ever interacting, and ii) frequency of interaction for those who do. Additional conditional priors could be added where the biology requires it, for example in a society with strong community structure, such as female elephants in which kin structure would create additional levels of social clustering. Although this approach was inspired by reducing bias observed in sparse networks, it could have value for networks of all densities.

Premise Automated disease, weed, and crop classification with computer vision will be invaluable in the future of agriculture. However, existing model architectures like ResNet, EfficientNet, and ConvNeXt often underperform on smaller, specialised datasets typical of such projects. Methods We address this gap with informed data collection and the development of a new convolutional neural network architecture, PhytNet. Utilising a novel dataset of infrared cocoa tree images, we demonstrate PhytNet's development and compare its performance with existing architectures. Data collection was informed by spectroscopy data, which provided useful insights into the spectral characteristics of cocoa trees. Cocoa was chosen as a focal species due to the diverse pathology of its diseases, which pose significant challenges for detection. Results ResNet18 showed some signs of overfitting, while EfficientNet variants showed distinct signs of overfitting. By contrast, PhytNet displayed excellent attention to relevant features, almost no overfitting, and an exceptionally low computation cost of 1.19 GFLOPS. Conclusions We show that PhytNet is a promising candidate for rapid disease or plant classification and for precise localisation of disease symptoms for autonomous systems. We also show that the most informative light spectra for detecting cocoa disease are outside the visible spectrum and that efforts to detect disease in cocoa should be focused on local symptoms, rather than the systemic effects of disease.

Understanding how and why menopause has evolved is a long-standing challenge across disciplines. Females can typically maximize their reproductive success by reproducing for the whole of their adult life. In humans, however, women cease reproduction several decades before the end of their natural lifespan1,2. Although progress has been made in understanding the adaptive value of menopause in humans3,4, the generality of these findings remains unclear. Toothed whales are the only mammal taxon in which menopause has evolved several times⁵, providing a unique opportunity to test the theories of how and why menopause evolves in a comparative context. Here, we assemble and analyse a comparative database to test competing evolutionary hypotheses. We find that menopause evolved in toothed whales by females extending their lifespan without increasing their reproductive lifespan, as predicted by the ‘live-long’ hypotheses. We further show that menopause results in females increasing their opportunity for intergenerational help by increasing their lifespan overlap with their grandoffspring and offspring without increasing their reproductive overlap with their daughters. Our results provide an informative comparison for the evolution of human life history and demonstrate that the same pathway that led to menopause in humans can also explain the evolution of menopause in toothed whales.

The mean relatedness of an individual to others in its social group may change with age. Such kinship dynamics, which are expected to influence sex- and age-specific social behaviours, are sensitive to demography. Existing models of kinship dynamics, however, have focused chiefly on the influences of population-level demographic parameters (e.g., sex-specific dispersal rate and mating pattern). Here, we extend these models to explore the effects of local group size on kinship dynamics and the resulting patterns of selection for sex- and age-specific helping and harming. We show that individuals' average relatedness to others is higher in smaller than larger groups, and consequently, selective pressure for helping or harming is generally stronger in smaller than larger groups. Moreover, relatedness changes faster with age in smaller groups, particularly at younger ages. These patterns favour more extreme social traits in smaller groups and especially so at earlier life history stages. In particular we highlight that, while social systems characterized by bi-sexual philopatry with non-local mating (e.g., whales) are known to favour a shift from more helpful to more harmful behaviour as females age, a trend that has been invoked to explain the evolution of menopause and post-reproductive helping, the timing of this shift is predicted to be sensitive to local group size, potentially favouring earlier female reproductive cessation in smaller groups. Our study generates new insights into the effects of local group properties on kinship dynamics, and suggests that such effects might also help to explain widespread variation in age-linked social traits, such as the timing of menopause in social mammals.

Plant pathogens can decimate crops and render the local cultivation of a species unprofitable. In extreme cases this has caused famine and economic collapse. Timing is vital in treating crop diseases, and the use of computer vision for precise disease detection and timing of pesticide application is gaining popularity. Computer vision can reduce labour costs, prevent misdiagnosis of disease, and prevent misapplication of pesticides. Pesticide misapplication is both financially costly and can exacerbate pesticide resistance and pollution. Here, we review the application and development of computer vision and machine learning methods for the detection of plant disease. This review goes beyond the scope of previous works to discuss important technical concepts and considerations when applying computer vision to plant pathology. We present new case studies on adapting standard computer vision methods and review techniques for acquiring training data, the use of diagnostic tools from biology, and the inspection of informative features. In addition to an in‐depth discussion of convolutional neural networks (CNNs) and transformers, we also highlight the strengths of methods such as support vector machines and evolved neural networks. We discuss the benefits of carefully curating training data and consider situations where less computationally expensive techniques are advantageous. This includes a comparison of popular model architectures and a guide to their implementation.

Animal social networks are often constructed from point estimates of edge weights. In many contexts, edge weights are inferred from observational data, and the uncertainty around estimates can be affected by various factors. Though this has been acknowledged in previous work, methods that explicitly quantify uncertainty in edge weights have not yet been widely adopted and remain undeveloped for many common types of data. Furthermore, existing methods are unable to cope with some of the complexities often found in observational data, and do not propagate uncertainty in edge weights to subsequent statistical analyses. We introduce a unified Bayesian framework for modelling social networks based on observational data. This framework, which we call BISoN , can accommodate many common types of observational social data, can capture confounds and model effects at the level of observations and is fully compatible with popular methods used in social network analysis. We show how the framework can be applied to common types of data and how various types of downstream statistical analyses can be performed, including non‐random association tests and regressions on network properties. Our framework opens up the opportunity to test new types of hypotheses, make full use of observational datasets, and increase the reliability of scientific inferences. We have made both an R package and example R scripts available to enable adoption of the framework.

Understanding the evolution of menopause presents a long-standing scientific challenge1,2,3-why should females cease ovulation prior to the end of their natural lifespan? In human societies, intergenerational resource transfers, for example, food sharing and caregiving, are thought to have played a key role in the evolution of menopause, providing a pathway by which postreproductive females can boost the fitness of their kin.4,5,6 To date however, other late-life contributions that postreproductive females may provide their kin have not been well studied. Here, we test the hypothesis that postreproductive female resident killer whales (Orcinus orca) provide social support to their offspring by reducing the socially inflicted injuries they experience. We found that socially inflicted injuries, as quantified by tooth rake marks, are lower for male offspring in the presence of their postreproductive mother. In contrast, we find no evidence that postreproductive mothers reduce rake marking in their daughters. Similarly, we find no evidence that either reproductive mothers or grandmothers (reproductive or postreproductive) reduce socially inflicted injuries in their offspring and grandoffspring, respectively. Moreover, we find that postreproductive females have no effect on reducing the rake marks for whales in their social unit who are not their offspring. Taken together, our results highlight that directing late-life support may be a key pathway by which postreproductive females transfer social benefits to their male offspring.

Age-related changes in the patterns of local relatedness (kinship dynamics) can be a significant selective force shaping the evolution of life history and social behaviour. In humans and some species of toothed whales, average female relatedness increases with age, which can select for a prolonged post-reproductive lifespan in older females due to both costs of reproductive conflict and benefits of late-life helping of kin. Killer whales (Orcinus orca) provide a valuable system for exploring social dynamics related to such costs and benefits in a mammal with an extended post-reproductive female lifespan. We use more than 40 years of demographic and association data on the mammal-eating Bigg's killer whale to quantify how mother-offspring social relationships change with offspring age and identify opportunities for late-life helping and the potential for an intergenerational reproductive conflict. Our results suggest a high degree of male philopatry and female-biased budding dispersal in Bigg's killer whales, with some variability in the dispersal rate for both sexes. These patterns of dispersal provide opportunities for late-life helping particularly between mothers and their adult sons, while partly mitigating the costs of mother-daughter reproductive conflict. Our results provide an important step towards understanding why and how menopause has evolved in Bigg's killer whales.

Accurate measures of lifespan and age-specific mortality are important both for understanding life-history evolution and informing conservation and population management strategies. The most accurate data to estimate lifespan are from longitudinal studies, but for many species - especially those such as toothed whales that are wide-ranging and live in difficult-to-access environments - these longitudinal data are not available. However, other forms of age-structured data - such as from mass-strandings - are available for many toothed species, and using these data to infer patterns of age-specific mortality and lifespan remains an important outstanding challenge. Here we develop and test a Bayesian mortality model to derive parameters of a mortality function from age-structured data while accounting for potential error introduced to these data by mistakes in age estimation, sampling biases and population growth. We then searched the literature to assemble a database of 269 published age-structured toothed whale datasets. We applied our mortality model to derive lifespan estimates for 32 species of female and 33 species of male toothed whale. We also use our model to characterise sex differences in lifespan in toothed whales. Our mortality model allows us to curate the most complete and accurate collection of toothed whale lifespans to date.

A major goal in evolutionary biology is to elucidate common principles that drive human and other animal societies towards war or peace. One idea proposed to account for differences between human societies is the "democratic peace" hypothesis, which suggests autocracies are more warlike than democracies because leaders can pursue fights for private gain. Given that animals can make autocratic or democratic collective decisions, we consider whether the logic of democratic peace may apply across taxa. We adapt the classic Hawk-Dove model to consider conflict decisions by groups rather than individuals, finding support for the democratic peace hypothesis without mechanisms involving complex human institutions. We suggest that the degree to which collective decisions are shared may explain variation in the intensity of intergroup conflict in nature.