Shripad Tuljapurkar · Stanford University · PhdFit
Resume-aware faculty matching
Find professors who actually fit you
Upload your resume. PhdFit's six research agents compare your background with faculty profiles, recent publications, lab focus, and outreach opportunities, then rank professors with evidence you can review.
Resume-aware matchingEvidence from recent researchOutreach drafts included
Advisor Match Brief Resume uploaded
Resume signals5 found
NLPHCIFairnessGraph MLPyTorch
Six-agent analysisRunning
Axel · Candidate AnalystResume parsed
Nova · Professor ResearcherRecent papers
Echo · Match ExplainerEvidence linked
Atlas · Strategy AdvisorFit ranked
Lyra · Outreach WriterDraft queued
Juno · Meeting Prep CompilerNext step ready
SC
Dr. Sarah Chen
Stanford · NLP & interpretability
91
fit
Fairness + NLP projects overlap with her recent work.
Strong methods fit: model evaluation and interpretability.
Clear outreach angle for current trustworthy AI research.
Outreach angle
Ask how her lab is extending interpretability methods into fairness audits for real-world AI systems.
Shripad Tuljapurkar is the Dean and Virginia Morrison Professor of Population Studies in the Department of Biology at Stanford University. His research interests include demography of both animals and humans, as well as theoretical ecology and evolution. He is affiliated with several programs including Bio-X, CEHG, ChEM-H, Hopkins Marine Station, Woods Institute, and Wu Tsai Neuro. His work focuses on understanding population dynamics through the lenses of evolution and ecology, contributing to the fields of demography and theoretical biology.
We show how transient responses to pulse disturbances accumulate to determine a press disturbances. For stage-structured models, this cumulative change is given by a new Transient Response Matrix (TRM) that can be readily computed. TRM also yields the second derivatives of population growth rate with respect to matrix elements Much work in ecology has developed methods to help predict how natural populations respond to disturbances, but analyses of pulse (one-time) and press (persistent) disturbances have been largely disconnected. Here we show how transient responses to pulse disturbances accumulate to determine the long-term response to press disturbances. For stage-structured models, this cumulative change is given by a new Transient Response Matrix (TRM) that can be readily computed. Strikingly, the TRM also yields the second derivatives of population growth rate with respect to matrix elements. Thus there is an intimate but unexpected relationship between nonlinear selection pressures on demographic rates, and the transient dynamics of populations. This relationship yields a strong correlation between nonlinear selection and generation time across 439 unique plant and animal species (2690 population models). We also show that the TRM is directly related to Cohen's cumulative distance measure for populations converging to stability. Finally, we indicate how our method is generalizable to stochastic disturbances, and to equilibria in nonlinear models. The code for running the analysis is attached and the setup is as follows: Data file for Figure 2 and 4: mpm1.csv - This is matrix used for Figure 2. Data files for Figure 5: final_tree_corrected_newversion v2.tre full animal and plant data v2.csv match_list_uncorrected_newversion v2.csv These files contain phylogenetically corrected tree for the data from Compadre and Comadre databases. We use these to then regress the traits generation time against dominant eigenvalue of the Transient Response Matrix (J0). Please load all files in the same working directory to run the analysis in the Rscript titled "code for pulse and press paper figures.R"
Zenodo (CERN European Organization for Nuclear Research) · 2026-03-07
articleOpen access
We show how transient responses to pulse disturbances accumulate to determine a press disturbances. For stage-structured models, this cumulative change is given by a new Transient Response Matrix (TRM) that can be readily computed. TRM also yields the second derivatives of population growth rate with respect to matrix elements Much work in ecology has developed methods to help predict how natural populations respond to disturbances, but analyses of pulse (one-time) and press (persistent) disturbances have been largely disconnected. Here we show how transient responses to pulse disturbances accumulate to determine the long-term response to press disturbances. For stage-structured models, this cumulative change is given by a new Transient Response Matrix (TRM) that can be readily computed. Strikingly, the TRM also yields the second derivatives of population growth rate with respect to matrix elements. Thus there is an intimate but unexpected relationship between nonlinear selection pressures on demographic rates, and the transient dynamics of populations. This relationship yields a strong correlation between nonlinear selection and generation time across 439 unique plant and animal species (2690 population models). We also show that the TRM is directly related to Cohen's cumulative distance measure for populations converging to stability. Finally, we indicate how our method is generalizable to stochastic disturbances, and to equilibria in nonlinear models. The code for running the analysis is attached and the setup is as follows: Data file for Figure 2 and 4: mpm1.csv - This is matrix used for Figure 2. Data files for Figure 5: final_tree_corrected_newversion v2.tre full animal and plant data v2.csv match_list_uncorrected_newversion v2.csv These files contain phylogenetically corrected tree for the data from Compadre and Comadre databases. We use these to then regress the traits generation time against dominant eigenvalue of the Transient Response Matrix (J0). Please load all files in the same working directory to run the analysis in the Rscript titled "code for pulse and press paper figures.R"
We utilized a long-term study of Weddell seals to compute lifetime reproductive success (LRS) distributions using a theoretical approach and an empirical approach. These comparisons are often difficult to achieve among natural populations but are important for disentangling sources of variation in lifetime measures. We performed three independent analyses: the first compared LRS and age-at-death distributions among populations living under different environmental conditions, the second focused on populations differing in individual heterogeneity (i.e., high or low reproductive strategy), and the third compared a theoretical LRS distribution with an empirical LRS distribution. Iceberg conditions increased the probability of LRS = 0 while individual heterogeneity had little influence on LRS distributions. Age-at-death distributions were also highly skewed and only early life mortality was affected by environmental condition. Our theoretical LRS distribution was strikingly similar to the empirical distribution estimated from females experiencing natural intrinsic trait variability and extrinsic environmental variability. We also examine inequality measures which show that that females must be “lucky” to survive past maturity and of those who do, only 74% breed. These findings contribute to ongoing research that reveals how diversity in lifetime outcomes is largely governed by chance alone.
Understanding why population growth rates vary through time is central to ecology, evolution, and conservation. In structured populations, such variation arises from both environmentally-driven fluctuations in vital rates and intrinsic transient dynamics generated by changes in population structure. Despite long-standing recognition of these processes, existing approaches do not provide an exact and general partitioning of their relative contributions. Here, we develop a mathematically rigorous framework that decomposes variation in realized population growth rate into contributions from fluctuations in vital rates and from transient deviations in population structure. Building on stochastic population theory, we derive a first-order decomposition under stationary environmental variation, yielding analytical expressions for the variance components associated with each source. This framework clarifies how damping rate, life-history speed, and covariance among vital rates shape temporal variability in growth. We complement the analytical results with a simulation-based procedure that allows the decomposition to be estimated from time series of population projection matrices without requiring observations of past population structure. Applying the method to empirical case studies spanning plants and animals with contrasting generation times, we show that short-lived species exhibit variability dominated by vital-rate fluctuations, whereas long-lived species can exhibit substantial contributions from transient population structure when vital-rate variation is sufficiently large. Our approach provides a unifying and exact link between stochastic demography and transient dynamics, offering a powerful tool for comparing life histories, testing ecological hypotheses, and evaluating how environmental variability propagates through population structure to influence population growth.
Our ability to define the causes of aging could enable targeted interventions to extend healthspan. Classical evolutionary models based on individual age have provided critical insights into empirical trajectories of aging; however, gaps remain. We argue that technological advances in data capture, resolution, and scale present a rich opportunity to shed light on heterogeneity in patterns of aging. Computational and data analysis advances have produced expanded theoretical models that explicitly address details of the underlying biology, introducing variables and dynamics that go beyond 'age' itself. We argue that by incorporating richer biological detail to create more integrative predictive models, we can gain insight into expected future distributions of aging within populations, and better understand the molecular and demographic context in which selection has given rise to variability in aging. We provide an overview of existing models that address heterogeneity, and outline future directions and applications that would advance this key area in aging and biomedical research.
Abstract When should you claim a pension (such as social security) in the US? Much current advice focuses on the increase in the annual pension benefit as individuals delay claiming a pension, an increase due to the actuarial fairness of the system. However, individuals face two risks: whether they live to the planned claiming age, and how long they live after they claim. By isolating and measuring these two separate demographic risks, we quantify and describe the effect of demographic uncertainty in ages at death on claiming decisions, without the complexities of additional factors. We show that for both individuals and couples, the coefficient of variation of the (combined) lifetime benefit, calculated as the ratio of the standard deviation to the average, increases as individuals/couples delay claiming, indicating a rising risk relative to the benefit. Using a simple utility analysis that only considers lifetime pension benefits, we find a conditional optimal portfolio under different risk aversion levels, suggesting that individuals and couples should consider both average benefit and risk when deciding on a claiming age, as with other investment decisions. This focused approach highlights the important role of lifespan uncertainty in Social Security claiming decisions. We illustrate the flexibility and generality of our approach by applying it to income-based subgroups and deriving analytical solutions under the Gompertz mortality model, though the framework readily extends to other subgroup definitions and mortality assumptions. Our approach highlights one essential component in making a decision, and can be used in conjunction with factors such as wealth and consumption to make a more comprehensive analysis.
Traditional agricultural landscapes are vital reservoirs of biocultural heritage and agrobiodiversity, yet traditional farming systems and their unique crop landraces face increasing marginalization and genetic erosion. Using northwest Himalaya as a case study, we examine the ecological resilience and genetic diversity of an understudied traditional crop, black pea (scientific name unclear), alongside barley ( Hordeum vulgare ), and compare them to the introduced cash crop, green pea ( Pisum sativum L. ). Participatory field experiments with local farmers revealed that traditional crops outperform introduced varieties in survival and reproduction traits across sites. To our knowledge, we generate the first whole-genome sequencing data for black peas. Clustering and nutritional analyses highlight black pea’s genetic richness and dietary potential. Our findings underscore the importance of integrating traditional ecological knowledge with ecological science to sustain agrobiodiversity, enhance climate resilience, and promote sustainable food systems. We provide insights for global agrifood innovations and socioecological stability in fragile mountain ecosystems.
The number and age of kin determine the companionship and support individuals provide or receive. Over recent decades, fertility and mortality rates have changed considerably, with varying speeds across countries. We investigate the changes in kinship networks in response to time-varying demographic rates, with a focus on the speed of change. We start with stylized demographic trajectories to determine the separate effects of fertility and mortality. First, we find that differences in the number of living kin depend strongly on the speed of fertility decline. In a fast fertility transition (as in China), a 65-year-old could have 20% fewer daughters than a 70-year-old in a specific year. However, in a slow transition (as in India), this difference is only 7%. Second, the speed of fertility decline has large effects on the mean and variability of the ages of kin. Third, a cohort perspective provides valuable insight into the changes in the number and age of kin. Fourth, we show how changes in the age pattern of mortality affect kinship for individuals at different ages. We use these conclusions to examine and understand kin dynamics based on empirical demographic data from four illustrative countries (Thailand, Indonesia, Ghana, and Nigeria).
bioRxiv (Cold Spring Harbor Laboratory) · 2025-05-04 · 1 citations
preprintOpen access
Abstract In ecology and evolutionary biology, understanding the relationship between vital rates ( e.g. , survival, development, reproduction) and population growth is essential to elucidate how life history strategies are shaped by natural selection. However, the established demographic methods to decipher the relationship between vital rates and population growth often analyse only the linear changes in population fitness as a result of changes in vital rates, thus simplifying the complexities of said relationships. To overcome the widespread linearity simplification, here we introduce the second-order elasticities of mean population fitness, the S-elasticity . The S-elasticity quantifies how changes in one or more vital rates can produce a second-order change in mean fitness. We provide a systematic mathematical framework behind the S-elasticity, revealing its ability to identify the convex and concave responses of mean fitness to perturbations of vital rates. Through structured population models, we demonstrate the distinct roles of linear and nonlinear mean fitness responses, and their combination, enabling to characterise local concavity/convexity of the mean population fitness function. We illustrate the application and the differences of S-elasticities and their biological meanings using matrix population models of the armadillo ( Dasupys novemcinctus ) and Pyne’s plum ( Astragallus bibullatus ). These two case studies showcase how the S-elasticity provides key insights into mean fitness responses to perturbations on demographic process and their correlations. We discuss the improvements that the S-elasticity provides for species management and our understanding of how natural populations cope with environmental change.
