Frankie Cho — Conservation AI & Biodiversity Research Fellow

I build AI and optimisation tools to address biodiversity and climate challenges.

Optimisation · machine learning · decisions under uncertainty

Research

Selected projects

Screenshot of the Sumatran tiger game: after two surveys find no tiger, belief bars show 3% extant and 97% extinct, with Survey, Manage and Surrender actions to choose from

Conservation Under Uncertainty: POMDP Games

Browser-based games that teach decision-making under partial observability, from a Sumatran tiger demo to weeds, fisheries, finches, sea otters and wolves. You act on beliefs rather than the true state, manage budgets, and can compare your choices with optimal policies.

Play

JavaScript · POMDPs · Education

Decision table for a Great Barrier Reef case study: ecosystem value in A$ million for non-targeted and spatially targeted abatement under two states of nature, with expected values of A$255M and A$258M

Forthcoming

vira: Risk-Averse Value of Information

An R package that calculates risk-averse value of information (EVPI, EVSI and EVPXI) under CRRA and CARA utility. It elicits risk preferences with adaptive Bayesian discrete-choice questions, and is designed for applied ecologists without an economics background.

Docs

R · Value of information · Expected utility

Maps of Victoria, Australia comparing a robust conservation solution (1% target) with a partially robust solution (25% target)

Robust Systematic Conservation Prioritization

An R package for conservation planning under uncertainty — generating solutions robust to climate projections, species distribution models, and ecosystem service estimates.

GitHub CRAN

R · C++

Map of Great Britain showing tree planting under the expected-value strategy, with a scatter of broadleaf against conifer net present value

PNAS 2025

Risk Modelling for UK Tree Planting

Decision-support framework for resilient tree-planting strategies under compounding climate and economic uncertainties, informing national net-zero land use policy.

Presented at EAERE 2022. Media coverage: The Guardian, Phys.org, and others.

GitHub Paper

R · Matlab · CPLEX · SQL

Projected koala conservation outcomes against the policy target and total cost under inflexible to flexible strategies

Conservation Letters 2025

Flexible Koala Conservation Decisions

Stochastic optimization model showing that adaptive conservation strategies can substantially reduce costs and mitigate climate-driven extinction risk for koalas on private land.

GitHub Paper

Julia · R · Gurobi

VESDIO world map of the geographic source of impact, with China highlighted

VESDIO — Supply Chain Ecosystem Risk

A web app for quantifying how ecosystem service disruptions propagate through global supply chains, helping businesses assess and disclose nature-related financial risk.

GitHub

Python · Dash · EXIOBASE

Simulated landscape returns without and with spatially clustered shocks

Monte Carlo Landscape Decisions

Monte Carlo experiments showing when risk-averse (CVaR) land-use portfolios outperform expected-value planning under spatially correlated catastrophic shocks such as wildfire.

Presented at BIOECON 2023.

Julia · R

Top of a workflow infographic for mapping vegetation lost to Typhoon Mangkhut in Hong Kong using Sentinel-2 imagery

Remote Sensing of Post-Typhoon Effects

Geospatial analysis of typhoon-induced landscape change using satellite imagery, combining Google Earth Engine and ArcGIS to map recovery patterns.

ESRI China Young Scholars Award 2019 — Second Runner-Up.

View Project

ArcGIS · Google Earth Engine · Stata

Distribution of respondents' attribute weights for culture, family, housing, jobs and transportation

ahpsurvey — Analytic Hierarchy Process for R

CRAN package for the Analytic Hierarchy Process, with tools for aggregation, consistency checking, and sensitivity analysis of survey preference data.

35K+ downloads. Cited in multiple peer-reviewed articles.

GitHub CRAN

Maps of Great Britain showing woodland planting under four decision-making approaches

PNAS 2024

Natural Capital and Land Use Policy

Natural capital and optimisation approaches to designing land-use and nature-recovery policy in Great Britain, showing that flat-rate per-hectare payments deliver far less value than policies that optimally price activities, environmental change or ecosystem service flows.

Figure: Bateman et al. 2024, PNAS, CC BY 4.0.

PNAS Phil Trans B People & Nature

Natural capital · Optimisation · Policy design

Maps of New South Wales local government areas showing conservation tender budget allocation under different planning approaches

Conservation Letters 2026

People in Private-Land Planning

A review of 66 conservation planning studies on how landholders are included, and a koala tender case study where optimal investments depend more on landholder preferences than on climate.

Figures: Archibald et al. 2026; Williams et al. 2024. CC BY 4.0.

Cons. Letters ERL

Conservation planning · Optimisation

3-D surface plot of the expected value of perfect information against the probability of a state and a utility gap

Methods Ecol Evol 2024

Why Collecting More Data Often Isn't Worth It

Explains why value-of-information analyses often find that extra data barely improves management decisions, reconciling this with ecologists' intuition and making the method less of a black box.

Figure: Holden et al. 2024, Methods Ecol Evol, CC BY 4.0.

Paper

Value of information · Decision theory

Diagram of the telecoupling framework with sending, receiving and spillover systems

BioScience 2025

Telecoupling and the Forest Carbon Market

Argues that global links between buyers and sellers of forest carbon credits hide social and ecological harms, and recommends prioritising social impact, comprehensive ecological outcomes and transparent investment.

Figure: Williams et al. 2025, BioScience, CC BY-NC 4.0.

Paper

Telecoupling · Carbon markets

Join my group

Open positions

3 open PhD & Honours positions in AI for conservation

See all opportunities

Students

Current student projects

5 students currently in the group

Publications

Selected publications

Journal articles

2026C. Archibald, B. Williams, A. Tulloch, … F.H.T. Cho, et al. Systematic Conservation Planning for Private Land Can No Longer Overlook People (opens in new tab). Conservation Letters 19(5), e70066.
2025F.H.T. Cho, P. Agnolucci, I.J. Bateman, et al. Resilient tree-planting strategies for carbon dioxide removal under compounding climate and economic uncertainties (opens in new tab). Proceedings of the National Academy of Sciences 122(10), e2320961122.
2025F.H.T. Cho, B.A. Williams, C.L. Archibald, et al. Flexible Climate Adaptation Can Substantially Reduce Conservation Costs and Mitigate Risk (opens in new tab). Conservation Letters 18(1), e13084.
2025B.A. Williams, J.-P. Metzger, F.H.T. Cho, et al. Telecoupling cannot be ignored for the forest-based carbon market (opens in new tab). BioScience 75(1), 61-67.
2024I.J. Bateman, A. Binner, E.T. Addicott, … F.H.T. Cho, et al. How to make land use policy decisions: Integrating science and economics to deliver connected climate, biodiversity, and food objectives (opens in new tab). Proceedings of the National Academy of Sciences 121(49), e2407961121.
2024B. Day, M. Mancini, I.J. Bateman, … F.H.T. Cho, et al. Natural capital approaches for the optimal design of policies for nature recovery (opens in new tab). Philosophical Transactions of the Royal Society B: Biological Sciences 379.
2024M.H. Holden, M.D. Akinlotan, A.D. Binley, … F.H.T. Cho, et al. Why shouldn't I collect more data? Reconciling disagreements between intuition and value of information analyses (opens in new tab). Methods in Ecology and Evolution 15(9), 1580-1592.
2024B.A. Williams, C.L. Archibald, J. Brazill-Boast, … F.H.T. Cho, et al. Optimal investments in private land conservation depend more on landholder preferences than climate change (opens in new tab). Environmental Research Letters 19(12), 124047.
2023I.J. Bateman, K. Anderson, A. Argles, … F.H.T. Cho, et al. A review of planting principles to identify the right place for the right tree for ‘net zero plus’ woodlands: Applying a place-based natural capital framework for sustainable, efficient and equitable (SEE) decisions (opens in new tab). People and Nature 5(2), 271-301.
2021X. Li, W.Y. Chen, F.Z.Y. Hu, … F.H.T. Cho, et al. Homebuyers’ heterogeneous preferences for urban green–blue spaces: A spatial multilevel autoregressive analysis (opens in new tab). Landscape and Urban Planning 216, 104250.
2021X. Li, W.Y. Chen, F.H.T. Cho, et al. Bringing the vertical dimension into a planar multilevel autoregressive model: A city-level hedonic analysis of homebuyers' utilities and urban river attributes (opens in new tab). Science of the Total Environment 772, 145547.
2021W.Y. Chen, F.H.T. Cho. Understanding China’s transition to environmental information transparency: citizens’ protest attitudes and choice behaviours (opens in new tab). Journal of Environmental Policy & Planning 23(3), 275-301.
2021F.H.T. Cho, W.Y. Chen, J. Hua. Validating Citizens’ Preferences for Restoring Urban Riverscape: Discrete Choice Experiment versus Analytical Hierarchy Process (opens in new tab). Journal of Water Resources Planning and Management 147(7), 05021005.
2021J. Hua, W.Y. Chen, I. Liekens, … F.H.T. Cho, et al. Partial attribute attendance in environmental choice experiments: A comparative case study between Guangzhou (China) and Brussels (Belgium) (opens in new tab). Journal of Environmental Management 285, 112107.
2020X. Li, W.Y. Chen, F.H.T. Cho. 3-D spatial hedonic modelling: Environmental impacts of polluted urban river in a high-rise apartment market (opens in new tab). Landscape and Urban Planning 203, 103883.
2019W.Y. Chen, F.H.T. Cho. Environmental information disclosure and societal preferences for urban river restoration: Latent class modelling of a discrete-choice experiment (opens in new tab). Journal of Cleaner Production 231, 1294-1306.
2018L. Ran, M. Tian, N. Fang, … F.H.T. Cho, et al. Riverine carbon export in the arid to semiarid Wuding River catchment on the Chinese Loess Plateau (opens in new tab). Biogeosciences 15(12), 3857-3871.

Preprints & software

2026F.H.T. Cho, J. Hanson. robust.prioritizr: Robust Systematic Conservation Prioritization (opens in new tab). EcoEvoRxiv (preprint).
2026X. Yang, F.H.T. Cho, I. Chadès. Windowed AK-MDP (opens in new tab). arXiv preprint arXiv:2609.13676.

Full list on Google Scholar (opens in new tab)

About

About me

I'm a Research Fellow in Data Science and AI at Monash University. I build optimisation and AI methods that help people make better decisions for nature: where to protect, restore or plant, and how to adapt those plans as climate, markets and knowledge change.

My work sits where ecology, economics and computer science meet. I combine mixed-integer and stochastic optimisation, sequential decision-making (Markov decision processes and reinforcement learning) and spatial data science to turn uncertain ecological and economic evidence into decisions that hold up across many possible futures.

This research has supported decisions on both sides of the world: national tree planting for net zero in the UK, with decision support for Defra, and koala conservation on private land in New South Wales, with the NSW Biodiversity Conservation Trust. It has been published in PNAS, Conservation Letters and Environmental Research Letters. I also turn methods into tools people use, including the robust.prioritizr (opens in new tab) and ahpsurvey (opens in new tab) R packages. Previously, as a nature and biodiversity consultant, I built analytics that help companies assess nature-related risk across their supply chains.

I trained in geography and environmental economics at the University of Hong Kong, then completed a joint PhD at the University of Exeter and the University of Queensland, including an Alan Turing Institute enrichment placement. Along the way I've also worked as a GIS analyst and a front-end software engineer, so I care about building tools that are fast, usable and open.

I'm keen to collaborate with ecologists, conservation agencies, optimisation and AI researchers, and organisations facing hard trade-offs between land use, biodiversity and climate. If you have a difficult decision problem, I'd love to hear about it.