Workflow design as a research variable in computational materials discovery
- 1Department of Chemical and Biomolecular Engineering
- 2KAIST
Abstract
For decades, much of computational materials discovery has relied on campaign architectures whose high-level decision policies were written in advance by human researchers. Agentic AI changes this arrangement because choices of method, system, and next step can increasingly be made at runtime. The workflow can therefore become a deliberative trajectory rather than only a pre-authored graph. I argue that this shift creates a new layer of research design. The central questions are no longer only how capable scientific agents are, but which decision rights should be delegated, how frequently humans should intervene, how agents should be distributed within a laboratory, and how stochastic research trajectories should be audited and reproduced. I develop the engagement-cadence trade-off, laboratory organization as an experimental variable, and a distinction between trajectory and conclusion reproducibility. I also address the strongest practical objection, namely that reliability degrades with task horizon and verification may erase throughput gains. I close with a measurement agenda for studying human-agent research systems empirically.
Related
- Integration of Acoustic Ejection Mass Spectrometry Enables Efficient High Throughput Purification Workflow for Peptides in Discovery Space
- A Catalyst Discovery Workflow Coupling MLIP with MOBO for Nitrite-to-Hydroxylamine Electroreduction
- AGAPE (Computational G-Quadruplex Stabilization Prediction): The First Machine Learning Workflow for G-Quadruplex Stabilization Prediction
- IMPECCABLE: A Digital Twin Drug Discovery Workflow Comprising AI, Physics and Experiment & the Creation of Nanomolar-Potent Inhibitors
- Accelerating the discovery of nanoporous gas separation materials via Coulomb-enhanced classical density functional theory
BibTeX
@misc{kim2026workflow,
title = {{Workflow design as a research variable in computational materials discovery}},
author = {Jihan Kim},
howpublished = {ChemRxiv},
year = {2026},
month = aug,
doi = {10.26434/chemrxiv.15008076/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008076/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Workflow design as a research variable in computational materials discovery
AU - Jihan Kim
PY - 2026
DA - 2026/08/31
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008076/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008076/v1
AB - For decades, much of computational materials discovery has relied on campaign architectures whose high-level decision policies were written in advance by human researchers. Agentic AI changes this arrangement because choices of method, system, and next step can increasingly be made at runtime. The workflow can therefore become a deliberative trajectory rather than only a pre-authored graph. I argue that this shift creates a new layer of research design. The central questions are no longer only how capable scientific agents are, but which decision rights should be delegated, how frequently humans should intervene, how agents should be distributed within a laboratory, and how stochastic research trajectories should be audited and reproduced. I develop the engagement-cadence trade-off, laboratory organization as an experimental variable, and a distinction between trajectory and conclusion reproducibility. I also address the strongest practical objection, namely that reliability degrades with task horizon and verification may erase throughput gains. I close with a measurement agenda for studying human-agent research systems empirically.
ER -