Hi, I’m Jake

I’m a research scientist working at the intersection of AI, data science, and animal behavior.

My research focuses on modeling and theory, using computational models, Bayesian statistics, and machine learning to understand animal behavior.

My current work explores decision making and inference: how animals use uncertain information to weigh alternatives and select actions. I develop models that connect internal beliefs, environmental context, and action selection to understand how choices emerge.

I collaborate with neuroscientists, biologists, physicists, and engineers, connecting theory with data from experiments and observations in the lab and in the wild.

My goal is to turn computational insights into scientific understanding, using AI and mathematical models to go beyond prediction: testing hypotheses and revealing the mechanisms behind animal behavior.

I completed my PhD with Iain Couzin at the Max Planck Institute of Animal Behavior and the University of Konstanz. My doctoral research developed deep learning and computer vision methods for measuring and modeling animal behavior.

My research Curriculum vitae
Jake Graving outdoors
Jake GravingResearch Scientist
Max Planck Institute of Animal Behavior CASCB, University of Konstanz Department of Biology, University of Konstanz
Konstanz, Germany

The geometry of choice

Move the targets to explore how their geometry shapes a decision. Each line is a repeat of the same model, starting with equal preference for each target.

Starting point
Closer togetherFar apart
More competitionMore cooperation
SlowerFaster
NoneStronger

Each path starts with equal weights for each target. Random fluctuations can tip the decision toward a target.

How the model works

The agent moves toward a weighted average of the target directions. When those directions are similar, the weights tend to stay balanced. As the agent approaches and the directions spread apart, balanced weights become unstable. Small random fluctuations can then grow, and the agent commits to one target. This loss of stability is a bifurcation.

The cooperation threshold is the angle where pairwise coupling changes sign: closer bearings cooperate and wider bearings compete. It defaults to 90 degrees. The cosine kernel varies smoothly as cos(angle) − cos(threshold). The step kernel uses +1 at or below the threshold and −1 above it. Both are scaled by the same coupling strength. The learning rate controls the belief update per step; zero freezes learning. Internal noise is the per-step standard deviation of a pairwise logit difference; zero gives deterministic dynamics. The model parameters are chosen to show the mechanism; they are not fitted to animal data. Noise acts on the internal weights, with no added movement noise. Moving the targets into an asymmetric arrangement can favor one outcome, even though the starting weights remain equal.

By playing, you allow YouTube (Google) to process data and use cookies. Privacy.

How a swarm moves

Collective behavior 2025

Testing how individual decisions give rise to collective motion in locusts.

About this study

This study revealed that locust swarming behavior emerges from internal decision-making processes rather than alignment with nearby individuals. I collaborated with colleagues to analyze behavioral data from controlled lab and immersive virtual reality experiments, developing a Bayesian model to test competing mechanistic hypotheses of collective motion. Our results challenged classical alignment-based theories, showing that group coordination emerges from individual decision-making informed by neural representations of neighboring locusts.

Publication · Perspective · Feature Article · Model Code

Citation

Sayin, S., Couzin-Fuchs, E., Petelski, I., Günzel, Y., Salahshour, M., Lee, C.-Y., Graving, J.M., Li, L., Deussen, O., Couzin, I.D., et al. (2025). The behavioral mechanisms governing collective motion in swarming locusts. Science, 387(6737), 995–1000. https://doi.org/10.1126/science.adq7832

By playing, you allow YouTube (Google) to process data and use cookies. Privacy.

Tracking groups with drones

Computer vision · Field research 2023

Reconstructing individual trajectories and environmental context with computer vision.

About this study

This work introduced a drone- and computer vision–based system for high-resolution tracking of animal groups in the wild. I helped build and refine a scalable pipeline that combines drone video with deep learning and computer vision algorithms to detect, track, and reconstruct 3D movement and environmental context for wild animals. Working with an interdisciplinary team, we enabled large-scale, non-invasive behavioral analysis in natural settings — advancing tools for behavioral ecology and conservation.

Code on GitHub · Publication · Press Release · Feature Article · HerdHover Project

Citation

Koger, B., Deshpande, A., Kerby, J.T., Graving, J.M., Costelloe, B.R., & Couzin, I.D. (2023). Quantifying the movement, behaviour and environmental context of group-living animals using drones and computer vision. Journal of Animal Ecology, 92(7), 1498–1515. https://doi.org/10.1111/1365-2656.13904

View full-resolution research figure

By playing, you allow YouTube (Google) to process data and use cookies. Privacy.

The mechanics of moving together

Collective behavior 2020

How fish coordinate their tailbeats to exploit the vortices produced by their neighbors.

About this study

This paper identified a schooling strategy where fish synchronize their tailbeats to exploit hydrodynamic advantages, matching behavior previously seen in bio-inspired robots. I contributed to the analysis pipeline, applying deep learning–based pose estimation to track fish movement and test predictions from robotic models. The findings showed that real fish use similar vortex-phase matching strategies to gain energetic benefits during coordinated swimming.

Publication · Press Release

Citation

Li, L., Nagy, M., Graving, J.M., Bak-Coleman, J., Xie, G., & Couzin, I.D. (2020). Vortex phase matching as a strategy for schooling in robots and in fish. Nature Communications, 11, Article 5408. https://doi.org/10.1038/s41467-020-19086-0

Measuring animal posture

Animal pose estimation 2019

Tracking body landmarks across species with deep learning, using DeepPoseKit.

About this study

DeepPoseKit is an open-source software toolkit for fast, accurate animal pose estimation using a deep learning–based, multi-scale model. I led the development of DeepPoseKit to make deep learning–based posture tracking more accessible and efficient for behavioral researchers. The toolkit supports fast, accurate pose estimation across species — even with small training datasets — and has been adopted by labs working in both lab and field contexts.

Code on GitHub · Publication · Feature Article

Quanta Magazine · Nature Methods · Nature News & Views

Citation

Graving, J.M., Chae, D., Naik, H., Li, L., Koger, B., Costelloe, B.R., Couzin, I.D. (2019). DeepPoseKit, a software toolkit for fast and robust animal pose estimation using deep learning. eLife, 8, e47994. https://doi.org/10.7554/eLife.47994

Tracking birds’ social lives

Behavioral tracking 2018

Using lightweight barcodes and computer vision to map individual movements and social networks.

About this study

This system uses barcode-equipped backpacks to enable automated, individual-level tracking of birds for behavioral research. As part of a collaborative project, I helped design and implement the computer vision pipeline used to detect and track barcoded zebra finches. The system provides continuous identity, position, and orientation data, supporting long-term studies of behavior and social networks at high resolution.

Publication

Citation

Alarcón-Nieto, G., Graving, J.M., Klarevas-Irby, J.A., Maldonado-Chaparro, A.A., Mueller, I., & Farine, D.R. (2018). An automated barcode tracking system for behavioural studies in birds. Methods in Ecology and Evolution, 9(6), 1536–1547. https://doi.org/10.1111/2041-210X.13005

By playing, you allow YouTube (Google) to process data and use cookies. Privacy.

Finding a way in the dark

Sensory ecology 2016–2017

Exploring how whip spiders navigate using touch and smell.

About this study

Our studies revealed that whip spiders use tactile and olfactory cues — not vision — to navigate complex environments with impressive spatial precision. I led experimental design and data collection across both lab and field settings, collaborating with colleagues to combine radio telemetry, behavioral tracking, and sensory manipulations. Together, we showed that antenniform legs provide essential multisensory input for goal-directed navigation in the absence of visual cues.

Journal of Experimental Biology (2017) · Journal of Comparative Physiology A (2017) · Frontiers in Behavioral Neuroscience (2016)

Citations

Graving, J.M., Bingman, V.P., Hebets, E.A., & Wiegmann, D.D. (2017). Development of site fidelity in the nocturnal amblypygid, Phrynus marginemaculatus. Journal of Comparative Physiology A, 203, 313–328. https://doi.org/10.1007/s00359-017-1169-5

Bingman, V.P., Graving, J.M., Hebets, E.A., & Wiegmann, D.D. (2017). Importance of the antenniform legs, but not vision, for homing by the neotropical whip spider Paraphrynus laevifrons. Journal of Experimental Biology, 220, 885–890. https://doi.org/10.1242/jeb.149823

Wiegmann, D.D., Hebets, E.A., Gronenberg, W., Graving, J.M., & Bingman, V.P. (2016). Amblypygids: Model organisms for the study of arthropod navigation mechanisms in complex environments? Frontiers in Behavioral Neuroscience, 10:47. https://doi.org/10.3389/fnbeh.2016.00047

Animal tracks, circular distributions, and Bayesian model summaries.

Making sense of direction

In progress

Developing Bayesian circular statistics to model directional data and quantify uncertainty in animal behavior.

In collaboration with James Foster · University of Konstanz.

Optimal decision making

In progress

Exploring how movement, beliefs, and spatial geometry shape animal decisions.

In collaboration with Iain Couzin · MPI-AB / University of Konstanz.

3D reconstruction of zebras moving across natural terrain.

Movement decisions in the wild

In progress

Studying how Grévy’s and plains zebras combine terrain, social context, and internal state when choosing where to move.

In collaboration with Blair Costelloe · Mara Elephant Project.

Three-dimensional honeybee nest maps comparing cell contents, model predictions, and prediction likelihood.

Inside the honeybee nest

In progress

Modeling how brood and food stores are arranged across three-dimensional nests as colonies develop.

In collaboration with Michael Smith · Auburn University.

Publications on Google Scholar