pith. sign in

arxiv: 2605.00909 · v1 · submitted 2026-04-29 · 💻 cs.AI · cond-mat.mtrl-sci· cs.LG

Accelerating battery research with an AI interface between FINALES and Kadi4Mat

Giovanna Tosato (1) , Leon Merker (1 , 2 , 3) , Monika Vogler (3) , Michael Selzer (1) , Arnd Koeppe (1) ((1) Karlsruhe Institute of Technology , (2) Helmholtz Institute Ulm
show 1 more author
(3) Technical University of Munich)
This is my paper

Pith reviewed 2026-05-09 20:52 UTC · model grok-4.3

classification 💻 cs.AI cond-mat.mtrl-scics.LG
keywords sodium-ion batteriesformation protocolsBayesian optimizationinteroperabilityPareto frontactive learningbattery researchresearch data management
0
0 comments X

The pith

An interface between two research platforms lets an AI agent optimize sodium-ion battery formation protocols.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper shows how to connect the FINALES system, which plans and runs battery experiments, with the Kadi4Mat data platform so that an active-learning agent can pick the next tests. The agent balances two goals: making the formation step as short as possible while keeping the cells' end-of-life performance high. A sympathetic reader cares because formation is a long, expensive part of making sodium-ion cells, and cutting the number of trials needed speeds up research. The linked systems support teams at different sites by letting automated machines and human steps work from the same data.

Core claim

The paper establishes a framework that enables interoperability between the FINALES orchestration system and the Kadi4Mat research data management ecosystem. In this setup the active-learning agent inside Kadi4Mat applies multi-objective batched Bayesian optimization to select formation parameters on the POLiS MAP. The workflow iteratively explores the trade-off between formation duration and end-of-life performance and returns candidate solutions that approximate the Pareto front while using fewer experiments overall.

What carries the argument

the interoperability framework that links FINALES experiment planning with Kadi4Mat data management and hosts the multi-objective batched Bayesian optimization agent for guiding experiment selection across automated and human workflows

If this is right

  • Formation protocols can be found that shorten processing time while preserving high end-of-life performance.
  • The total number of experiments required to reach good solutions is reduced.
  • Coordinated work across automated systems and human-operated workflows at multiple research centers becomes feasible.
  • The same interoperability approach can be transferred to other optimization tasks in materials science and engineering.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Teams at separate labs could feed new results into the same optimization loop without manual file transfers.
  • The candidate protocols identified near the Pareto front could be used as starting points when scaling formation processes to larger cell formats.
  • Similar bridges between planning and data platforms might shorten development cycles for other battery chemistries.

Load-bearing premise

The optimization agent can reliably choose useful formation experiments from past results alone without needing extensive extra checks on its predictions.

What would settle it

A side-by-side test in which the protocols chosen by the agent show no better balance of formation time and end-of-life performance than randomly selected or standard protocols after the cells are fully cycled.

Figures

Figures reproduced from arXiv: 2605.00909 by 2, (2) Helmholtz Institute Ulm, 3), (3) Technical University of Munich), Arnd Koeppe (1) ((1) Karlsruhe Institute of Technology, Giovanna Tosato (1), Leon Merker (1, Michael Selzer (1), Monika Vogler (3).

Figure 1
Figure 1. Figure 1: The conceptual framework that interfaces the Kadi ecosystem with the FINALES workflow system for [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Visualization of the application results in the objective space. Batches 0, 15, and 16 lie on the Pareto front and [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Effect of cycling protocol parameters on EOL cycle. (a) Charge and discharge rate [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Maximum charge (a) and discharge (b) capacity over cycles of the four cells constituting batch 15, which is a [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: A representation of the overall operation of the MAP, including also the data management tools. Active [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: A representation of the process orchestrated by the OVERLORT workflow manager. Requests for new [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The dashboard visualization provides the scientists with an overview of the optimization study. The panels [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The reduced view of the knowledge graph for the optimization study focuses on the main components of [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The full knowledge graph from the FINALES perspective. [PITH_FULL_IMAGE:figures/full_fig_p017_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Maximum charge capacity over cycles of the four cells constituting respectively batch 16 (a) and 17 (b). [PITH_FULL_IMAGE:figures/full_fig_p022_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Effect of cycling protocol parameters on formation time (expressed in hours). Evaluation limited to batches 0 [PITH_FULL_IMAGE:figures/full_fig_p023_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Effect of cycling protocol parameters on EOL cycle after including Trial 17. (a) Charge and discharge rate [PITH_FULL_IMAGE:figures/full_fig_p023_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Effect of cycling protocol parameters on formation time after including Trial 17. (a) Charge and discharge [PITH_FULL_IMAGE:figures/full_fig_p023_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Visualization of the application results including Trial 17. For clarity, regions of the objective space [PITH_FULL_IMAGE:figures/full_fig_p024_14.png] view at source ↗
read the original abstract

The time-consuming formation process critically impacts the longevity of sodium-ion coin cells and End Of Life (EOL) performance. This study aims to optimize formation protocols for duration efficiency, targeting high-performance outcomes while minimizing the number of experiments to reduce resource consumption and accelerate discovery. Specifically, we consider two potentially competing objectives: minimizing formation time and maximizing EOL performance. Beyond this application focus, we also present a methodological contribution: a framework designed to enable interoperability between the FINALES and Kadi RDM ecosystems, which we employ to tackle our optimization problem. In this setup, the FINALES framework orchestrates experiment planning and execution on the POLiS MAP, while an active-learning agent implemented within Kadi4Mat guides experiment selection, using multi-objective batched Bayesian optimization to efficiently explore the parameter space. This interoperability enhancement enables coordinated, distributed collaboration across automated systems and human-operated workflows, bridging multiple research centers. Using this approach, we iteratively explore the trade-off between formation time and EOL performance and identify candidate solutions approximating the Pareto front. The resulting workflow demonstrates the capability of interoperable infrastructures to facilitate data-driven optimization in battery research, and establishes a transferable framework applicable to diverse materials science and engineering optimization tasks.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript describes an interoperability framework between the FINALES experiment orchestration platform and the Kadi4Mat research data management system. This interface supports a multi-objective batched Bayesian optimization agent that selects formation protocols for sodium-ion coin cells to minimize formation time while maximizing end-of-life performance. The authors state that the setup was used to iteratively explore the trade-off and identify candidate protocols approximating the Pareto front, while emphasizing its role in enabling distributed collaboration across automated and human-operated workflows.

Significance. The interoperability framework offers a practical methodological advance for integrating automated experiment systems with data platforms, which could facilitate collaborative, data-driven optimization in battery research and related fields. Credit is due for the deployed, cross-institutional workflow description. However, the absence of any quantitative results, surrogate validation, or optimization diagnostics means the claimed identification of Pareto-front candidates and acceleration of discovery remain prospective rather than demonstrated.

major comments (2)
  1. [Abstract] Abstract: The claim that the approach 'identify[s] candidate solutions approximating the Pareto front' is load-bearing for the central contribution yet unsupported by any reported protocols, performance metrics, hypervolume values, or convergence diagnostics.
  2. [Section 3] Section 3 (Optimization Agent): The multi-objective batched Bayesian optimization is described only at the workflow level; no details are given on the surrogate model formulation, acquisition function (e.g., EHVI or qEHVI), batch size, or any model-validation steps, preventing assessment of whether the agent reliably guides experiment selection.
minor comments (2)
  1. [Abstract] The acronym 'EOL' is introduced in the abstract without prior expansion, although it is defined later in the text.
  2. [Figure 1] Workflow diagrams (if present) would benefit from explicit arrows or labels showing data exchange directionality between FINALES and Kadi4Mat components.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive feedback on our manuscript. We have addressed the major comments by revising the abstract to accurately reflect the scope of our claims and by expanding the description of the optimization agent with additional technical details.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that the approach 'identify[s] candidate solutions approximating the Pareto front' is load-bearing for the central contribution yet unsupported by any reported protocols, performance metrics, hypervolume values, or convergence diagnostics.

    Authors: We concur that the abstract's phrasing implies completed identification of Pareto front candidates without the supporting data. Our work centers on the interoperability framework that enables such optimization. We will revise the abstract to emphasize the framework's role in facilitating the exploration of trade-offs and candidate identification, noting that specific quantitative outcomes are illustrative of the workflow's potential rather than fully validated results in this study. revision: yes

  2. Referee: [Section 3] Section 3 (Optimization Agent): The multi-objective batched Bayesian optimization is described only at the workflow level; no details are given on the surrogate model formulation, acquisition function (e.g., EHVI or qEHVI), batch size, or any model-validation steps, preventing assessment of whether the agent reliably guides experiment selection.

    Authors: We appreciate this observation and agree that more specifics are required. In the revised manuscript, Section 3 will be updated to detail the surrogate model as a multi-output Gaussian process, the use of the qEHVI acquisition function for batched multi-objective optimization, the chosen batch size, and the model validation approach including any hyperparameter tuning or cross-validation steps performed. revision: yes

Circularity Check

0 steps flagged

No circularity: descriptive framework without derivations or predictions

full rationale

The paper describes an interoperability setup between FINALES and Kadi4Mat, using multi-objective batched Bayesian optimization to explore formation time vs. EOL performance trade-offs in sodium-ion cells. No equations, derivations, fitted parameters, or first-principles results are presented. Claims concern workflow deployment and experiment selection guidance, not any prediction that reduces to inputs by construction. No self-citations serve as load-bearing uniqueness theorems or ansatzes. This is a standard non-circular engineering report on a deployed system.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the effectiveness of Bayesian optimization for experimental design and on the technical feasibility of the FINALES-Kadi4Mat interface; no new physical entities or ad-hoc constants are introduced.

axioms (1)
  • domain assumption Multi-objective batched Bayesian optimization can efficiently identify near-Pareto-optimal points in the formation-parameter space
    Invoked when the active-learning agent is said to guide experiment selection and approximate the Pareto front.

pith-pipeline@v0.9.0 · 5571 in / 1182 out tokens · 45143 ms · 2026-05-09T20:52:09.627244+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages

  1. [1]

    Current status and challenges for automotive battery production technologies.Nature Energy, 3(4):290–300, 2018

    Arno Kwade, Wolfgang Haselrieder, Ruben Leithoff, Armin Modlinger, Franz Dietrich, and Klaus Droeder. Current status and challenges for automotive battery production technologies.Nature Energy, 3(4):290–300, 2018

    work page 2018

  2. [2]

    PEM Der RWTH Aachen University, 2018

    Heiner Hans Heimes, Achim Kampker, Christoph Lienemann, Marc Locke, Christian Offermanns, Sarah Michaelis, and Ehsan Rahimzei.Lithium-ion battery cell production process. PEM Der RWTH Aachen University, 2018

    work page 2018

  3. [3]

    The effects of mechanical and thermal loads during lithium-ion pouch cell formation and their impacts on process time.Energy Technology, 8 (2):1900118, 2020

    Heiner Hans Heimes, Christian Offermanns, Ahmad Mohsseni, Hendrik Laufen, Uwe Westerhoff, Louisa Hoffmann, Philip Niehoff, Michael Kurrat, Martin Winter, and Achim Kampker. The effects of mechanical and thermal loads during lithium-ion pouch cell formation and their impacts on process time.Energy Technology, 8 (2):1900118, 2020

    work page 2020

  4. [4]

    Formation challenges of lithium-ion battery manufacturing.Joule, 3 (12):2884–2888, 2019

    David L Wood, Jianlin Li, and Seong Jin An. Formation challenges of lithium-ion battery manufacturing.Joule, 3 (12):2884–2888, 2019. 18 arXivTemplateA PREPRINT

    work page 2019

  5. [5]

    Modeling the performance and cost of lithium-ion batteries for electric-drive vehicles

    Paul A Nelson, Shabbir Ahmed, Kevin G Gallagher, and Dennis W Dees. Modeling the performance and cost of lithium-ion batteries for electric-drive vehicles. Technical report, Argonne National Laboratory (ANL), Argonne, IL (United States), 2019

    work page 2019

  6. [6]

    PhD thesis, Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2024, 2024

    Martin Florian Börner.Ein prozessbasiertes Modell zur Berechnung der Kosten von Lithium-Ionen-Batteriezellen. PhD thesis, Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2024, 2024

    work page 2024

  7. [7]

    Economic analysis of cnt lithium-ion battery manufacturing.Environmental Science: Nano, 2(5):463–476, 2015

    A Hakimian, S Kamarthi, S Erbis, KM Abraham, TP Cullinane, and JA Isaacs. Economic analysis of cnt lithium-ion battery manufacturing.Environmental Science: Nano, 2(5):463–476, 2015

    work page 2015

  8. [8]

    Current and future lithium-ion battery manufacturing

    Yangtao Liu, Ruihan Zhang, Jun Wang, and Yan Wang. Current and future lithium-ion battery manufacturing. IScience, 24(4), 2021

    work page 2021

  9. [9]

    From small batteries to big claims.Nature nanotechnology, pages 1–7, 2025

    Rares-George Scurtu, Alessandro Innocenti, Vanessa Scheck, Mario Maunz, Thomas Waldmann, Markus Hölzle, Alice Hoffmann, Peter Axmann, and Margret Wohlfahrt-Mehrens. From small batteries to big claims.Nature nanotechnology, pages 1–7, 2025

    work page 2025

  10. [10]

    Castelli, Francisco Fernando Ramirez, Johan Carlsson, Giovanni Pizzi, Simon Clark, Felix Hanke, Arghya Bhowmik, and Helge S

    Monika V ogler, Jonas Busk, Hamidreza Hajiyani, Peter Bjørn Jørgensen, Nehzat Safaei, Ivano E. Castelli, Francisco Fernando Ramirez, Johan Carlsson, Giovanni Pizzi, Simon Clark, Felix Hanke, Arghya Bhowmik, and Helge S. Stein. Brokering between tenants for an international materials acceleration platform.Matter, 6 (9):2647–2665, 2023. ISSN 2590-2385. doi:...

    work page doi:10.1016/j.matt.2023.07.016 2023

  11. [11]

    Autonomous battery optimization by deploying distributed experiments and simulations.Advanced Energy Materials, 14(46):2403263, 2024

    Monika V ogler, Simon Krarup Steensen, Francisco Fernando Ramírez, Leon Merker, Jonas Busk, Johan Martin Carlsson, Laura Hannemose Rieger, Bojing Zhang, François Liot, Giovanni Pizzi, et al. Autonomous battery optimization by deploying distributed experiments and simulations.Advanced Energy Materials, 14(46):2403263, 2024

    work page 2024

  12. [12]

    Autonomous discovery of battery electrolytes with robotic experimentation and machine learning.Cell Reports Physical Science, 1(12), 2020

    Adarsh Dave, Jared Mitchell, Kirthevasan Kandasamy, Han Wang, Sven Burke, Biswajit Paria, Barnabás Póczos, Jay Whitacre, and Venkatasubramanian Viswanathan. Autonomous discovery of battery electrolytes with robotic experimentation and machine learning.Cell Reports Physical Science, 1(12), 2020

    work page 2020

  13. [13]

    The evolution of materials acceleration platforms: toward the laboratory of the future with amanda.Journal of Materials Science, 56(29):16422–16446, 2021

    Jerrit Wagner, Christian G Berger, Xiaoyan Du, Tobias Stubhan, Jens A Hauch, and Christoph J Brabec. The evolution of materials acceleration platforms: toward the laboratory of the future with amanda.Journal of Materials Science, 56(29):16422–16446, 2021

    work page 2021

  14. [14]

    Combinatorial approach for fast screening of functional materials.Journal of Polymer Science Part B: Polymer Physics, 48(14): 1587–1593, 2010

    Harald Walter, Guillaume Basset, Tilman Beierlein, Adrian V on Mühlenen, and Giovanni Nisato. Combinatorial approach for fast screening of functional materials.Journal of Polymer Science Part B: Polymer Physics, 48(14): 1587–1593, 2010

    work page 2010

  15. [15]

    A mobile robotic chemist.Nature, 583(7815):237–241, 2020

    Benjamin Burger, Phillip M Maffettone, Vladimir V Gusev, Catherine M Aitchison, Yang Bai, Xiaoyan Wang, Xiaobo Li, Ben M Alston, Buyi Li, Rob Clowes, et al. A mobile robotic chemist.Nature, 583(7815):237–241, 2020

    work page 2020

  16. [16]

    Fuzhan Rahmanian, Jackson Flowers, Dan Guevarra, Matthias Richter, Maximilian Fichtner, Phillip Donnely, John M Gregoire, and Helge S Stein. Enabling modular autonomous feedback-loops in materials science through hierarchical experimental laboratory automation and orchestration.Advanced Materials Interfaces, 9(8):2101987, 2022

    work page 2022

  17. [17]

    Alexey Sanin, Jackson K Flowers, Tobias H Piotrowiak, Frederic Felsen, Leon Merker, Alfred Ludwig, Dominic Bresser, and Helge Sören Stein. Integrating automated electrochemistry and high-throughput characterization with machine learning to explore si-ge-sn thin-film lithium battery anodes.Advanced Energy Materials, 15(11): 2404961, 2025

    work page 2025

  18. [18]

    Dynamic cycling enhances battery lifetime.Nature Energy, 10(2):172–180, 2025

    Alexis Geslin, Le Xu, Devi Ganapathi, Kevin Moy, William C Chueh, and Simona Onori. Dynamic cycling enhances battery lifetime.Nature Energy, 10(2):172–180, 2025

    work page 2025

  19. [19]

    Prospects for managing end-of-life lithium-ion batteries: present and future.Interdisciplinary Materials, 1(3):417–433, 2022

    Xiao-Tong Wang, Zhen-Yi Gu, Edison Huixiang Ang, Xin-Xin Zhao, Xing-Long Wu, and Yichun Liu. Prospects for managing end-of-life lithium-ion batteries: present and future.Interdisciplinary Materials, 1(3):417–433, 2022

    work page 2022

  20. [20]

    Recycling end-of-life electric vehicle lithium-ion batteries.Joule, 3(11):2622–2646, 2019

    Mengyuan Chen, Xiaotu Ma, Bin Chen, Renata Arsenault, Peter Karlson, Nakia Simon, and Yan Wang. Recycling end-of-life electric vehicle lithium-ion batteries.Joule, 3(11):2622–2646, 2019

    work page 2019

  21. [21]

    End-of-life (eol) issues and options for electric vehicle batteries.Clean Technologies and Environmental Policy, 15(6):881–891, 2013

    Monsuru Olalekan Ramoni and Hong-Chao Zhang. End-of-life (eol) issues and options for electric vehicle batteries.Clean Technologies and Environmental Policy, 15(6):881–891, 2013

    work page 2013

  22. [22]

    Quantifying ev battery end-of-life through analysis of travel needs with vehicle powertrain models.Journal of Power Sources, 282:265–276, 2015

    Samveg Saxena, Caroline Le Floch, Jason MacDonald, and Scott Moura. Quantifying ev battery end-of-life through analysis of travel needs with vehicle powertrain models.Journal of Power Sources, 282:265–276, 2015. 19 arXivTemplateA PREPRINT

    work page 2015

  23. [23]

    Yager, Danielle Monteverde, Dave Baiocchi, Hyukjun Kwon, Shi- jing Sun, and Santosh K

    Linda Hung, Joyce A. Yager, Danielle Monteverde, Dave Baiocchi, Ha-Kyung Kwon, Shijing Sun, and Santosh Suram. Autonomous laboratories for accelerated materials discovery: a community survey and practical insights. Digital Discovery, 3:1273–1279, 2024. doi:10.1039/D4DD00059E. URL http://dx.doi.org/10.1039/ D4DD00059E

    work page doi:10.1039/d4dd00059e 2024

  24. [24]

    Self-Driving Laboratories for Chemistry and Materials Science.Chemical Reviews, 2024

    Gary Tom, Stefan P. Schmid, Sterling G. Baird, Yang Cao, Kourosh Darvish, Han Hao, Stanley Lo, Sergio Pablo-García, Ella M. Rajaonson, Marta Skreta, Naruki Yoshikawa, Samantha Corapi, Gun Deniz Akkoc, Felix Strieth-Kalthoff, Martin Seifrid, and Alán Aspuru-Guzik. Self-driving laboratories for chemistry and materials science.Chemical Reviews, 124(16):9633–...

    work page doi:10.1021/acs.chemrev.4c00055 2024

  25. [25]

    Role of the human-in-the-loop in emerging self-driving laboratories for heterogeneous catalysis.Nature Catalysis, 8(1):13–19, 2025

    Christoph Scheurer and Karsten Reuter. Role of the human-in-the-loop in emerging self-driving laboratories for heterogeneous catalysis.Nature Catalysis, 8(1):13–19, 2025

    work page 2025

  26. [26]

    Rose, Huada Lian, Alexis Geslin, Steven B

    Xiao Cui, Stephen Dongmin Kang, Sunny Wang, Justin A. Rose, Huada Lian, Alexis Geslin, Steven B. Tor- risi, Martin Z. Bazant, Shijing Sun, and William C. Chueh. Data-driven analysis of battery formation re- veals the role of electrode utilization in extending cycle life.Joule, 8(11):3072–3087, 2024. ISSN 2542-

    work page 2024

  27. [27]

    URL https://www.sciencedirect.com/science/ article/pii/S2542435124003532

    doi:https://doi.org/10.1016/j.joule.2024.07.024. URL https://www.sciencedirect.com/science/ article/pii/S2542435124003532

    work page doi:10.1016/j.joule.2024.07.024 2024

  28. [28]

    Data-driven approaches for the fast formation of lithium-ion pouch cells using operando gassing analysis.Journal of Power Sources, 613:234858, 2024

    Sandro Stock, Jonas Böhm, Manuel Ank, Philipp Rath, Markus Lienkamp, and Rüdiger Daub. Data-driven approaches for the fast formation of lithium-ion pouch cells using operando gassing analysis.Journal of Power Sources, 613:234858, 2024. ISSN 0378-7753. doi:https://doi.org/10.1016/j.jpowsour.2024.234858. URL https://www.sciencedirect.com/science/article/pii...

    work page doi:10.1016/j.jpowsour.2024.234858 2024

  29. [29]

    Benchmarking the performance of bayesian optimization across multiple experimental materials science domains.npj Computational Materials, 7(1):188, 2021

    Qiaohao Liang, Aldair E Gongora, Zekun Ren, Armi Tiihonen, Zhe Liu, Shijing Sun, James R Deneault, Daniil Bash, Flore Mekki-Berrada, Saif A Khan, et al. Benchmarking the performance of bayesian optimization across multiple experimental materials science domains.npj Computational Materials, 7(1):188, 2021

    work page 2021

  30. [30]

    Benchmarking the acceleration of materials discovery by sequential learning.Chemical science, 11(10):2696–2706, 2020

    Brian Rohr, Helge S Stein, Dan Guevarra, Yu Wang, Joel A Haber, Muratahan Aykol, Santosh K Suram, and John M Gregoire. Benchmarking the acceleration of materials discovery by sequential learning.Chemical science, 11(10):2696–2706, 2020

    work page 2020

  31. [31]

    From materials discovery to system optimization by integrating combinatorial electrochemistry and data science.Current Opinion in Electrochemistry, 35:101053, 2022

    Helge S Stein, Alexey Sanin, Fuzhan Rahmanian, Bojing Zhang, Monika V ogler, Jackson K Flowers, Leon Fischer, Stefan Fuchs, Nirmal Choudhary, and Lisa Schroeder. From materials discovery to system optimization by integrating combinatorial electrochemistry and data science.Current Opinion in Electrochemistry, 35:101053, 2022

    work page 2022

  32. [32]

    Lattice metamaterials with mesoscale motifs: exploration of property charts by bayesian optimization.Advanced Engineering Materials, 25(13):2300048, 2023

    Roman Kulagin, Patrick Reiser, Kyryl Truskovskyi, Arnd Koeppe, Yan Beygelzimer, Yuri Estrin, Pascal Friederich, and Peter Gumbsch. Lattice metamaterials with mesoscale motifs: exploration of property charts by bayesian optimization.Advanced Engineering Materials, 25(13):2300048, 2023

    work page 2023

  33. [33]

    Characterization of porous membranes using artificial neural networks.Acta Materialia, 253: 118922, 2023

    Yinghan Zhao, Patrick Altschuh, Jay Santoki, Lars Griem, Giovanna Tosato, Michael Selzer, Arnd Koeppe, and Britta Nestler. Characterization of porous membranes using artificial neural networks.Acta Materialia, 253: 118922, 2023

    work page 2023

  34. [34]

    Machine learning assisted design of experiments for solid state electrolyte lithium aluminum titanium phosphate.Frontiers in Materials, 9:821817, 2022

    Yinghan Zhao, Nikolas Schiffmann, Arnd Koeppe, Nico Brandt, Ethel C Bucharsky, Karl G Schell, Michael Selzer, and Britta Nestler. Machine learning assisted design of experiments for solid state electrolyte lithium aluminum titanium phosphate.Frontiers in Materials, 9:821817, 2022

    work page 2022

  35. [35]

    Jeela, G

    R.K. Jeela, G. Tosato, M. Ahmad, M. Wieler, A. Koeppe, B. Nestler, and D. Schneider. Enhancing solid oxide fuel cells development through bayesian active learning.Advanced Energy Materials, page 2501216, 2025. doi:10.1002/aenm.202501216

    work page doi:10.1002/aenm.202501216 2025

  36. [36]

    Balancing formation time and electrochemical performance of high energy lithium-ion batteries.Journal of Power Sources, 402:107–115, 2018

    Chengyu Mao, Seong Jin An, Harry M Meyer III, Jianlin Li, Marissa Wood, Rose E Ruther, and David L Wood III. Balancing formation time and electrochemical performance of high energy lithium-ion batteries.Journal of Power Sources, 402:107–115, 2018

    work page 2018

  37. [37]

    Lisa: A lithium-ion solid-state assistant using large language models for knowledge defragmentation in battery science and beyond.Materials Today Communications, 45:112380, 2025

    Yinghan Zhao, Anna-Lena Hansen, Anna Dahlhaus, Nico Brandt, Michael Selzer, Arnd Koeppe, Britta Nestler, Michael Knapp, and Helmut Ehrenberg. Lisa: A lithium-ion solid-state assistant using large language models for knowledge defragmentation in battery science and beyond.Materials Today Communications, 45:112380, 2025

    work page 2025

  38. [38]

    Kadi4mat: A research data infrastructure for materials science.Data Science Journal, 20:8–8, 2021

    Nico Brandt, Lars Griem, Christoph Herrmann, Ephraim Schoof, Giovanna Tosato, Yinghan Zhao, Philipp Zschumme, and Michael Selzer. Kadi4mat: A research data infrastructure for materials science.Data Science Journal, 20:8–8, 2021

    work page 2021

  39. [39]

    Kadistudio: Fair modelling of scientific research processes.Data Science Journal, 21:16–16, 2022

    Lars Griem, Philipp Zschumme, Matthieu Laqua, Nico Brandt, Ephraim Schoof, Patrick Altschuh, and Michael Selzer. Kadistudio: Fair modelling of scientific research processes.Data Science Journal, 21:16–16, 2022. 20 arXivTemplateA PREPRINT

    work page 2022

  40. [40]

    Ax: A platform for adaptive experimentation

    Miles Olson, Elizabeth Santorella, Louis C Tiao, Sait Cakmak, David Eriksson, Mia Garrard, Sam Daulton, Maximilian Balandat, Eytan Bakshy, Elena Kashtelyan, et al. Ax: A platform for adaptive experimentation. In AutoML 2025 ABCD Track, 2025

    work page 2025

  41. [41]

    Cids: 3.1, 2023

    Arnd Koeppe and The KadiAI Team. Cids: 3.1, 2023

    work page 2023

  42. [42]

    PhD thesis, Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2021, 2021

    Arnd Koeppe.Deep learning in the finite element method. PhD thesis, Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2021, 2021

    work page 2021

  43. [43]

    Robotic cell assembly to accelerate battery research

    Bojing Zhang, Leon Merker, Alexey Sanin, and Helge S Stein. Robotic cell assembly to accelerate battery research. Digital Discovery, 1(6):755–762, 2022

    work page 2022

  44. [44]

    Apples to apples: shift from mass ratio to additive molecules per electrode area to optimize li-ion batteries.Digital Discovery, 3(7):1342–1349, 2024

    Bojing Zhang, Leon Merker, Monika V ogler, Fuzhan Rahmanian, and Helge S Stein. Apples to apples: shift from mass ratio to additive molecules per electrode area to optimize li-ion batteries.Digital Discovery, 3(7):1342–1349, 2024

    work page 2024

  45. [45]

    From powder to pouch cell: Setting up a sodium-ion battery reference system based on na3v2 (po4) 3/c and hard carbon.Batteries & Supercaps, 7(12):e202400406, 2024

    Pirmin Stüble, Cedric Müller, Nicole Bohn, Marcus Müller, Andreas Hofmann, Tolga Akçay, Julian Klemens, Arnd Koeppe, Satish Kolli, Deepalaxmi Rajagopal, et al. From powder to pouch cell: Setting up a sodium-ion battery reference system based on na3v2 (po4) 3/c and hard carbon.Batteries & Supercaps, 7(12):e202400406, 2024

    work page 2024

  46. [46]

    Bleken, Simon Stier, Eibar Flores, Casper Welzel Andersen, Marek Marcinek, Anna Szczesna-Chrzan, Miran Gaberscek, M

    Simon Clark, Francesca L. Bleken, Simon Stier, Eibar Flores, Casper Welzel Andersen, Marek Marcinek, Anna Szczesna-Chrzan, Miran Gaberscek, M. Rosa Palacin, Martin Uhrin, and Jesper Friis. Toward a unified description of battery data.Advanced Energy Materials, 12(17):2102702, 2022. doi:https://doi.org/10.1002/aenm.202102702. URLhttps://advanced.onlinelibr...

    work page doi:10.1002/aenm.202102702 2022

  47. [47]

    Bleken, Casper Welzel Andersen, Eibar Flores, Hendrik_Snijder, and Simon Stier

    Simon Clark, Jesper Friis, Francesca L. Bleken, Casper Welzel Andersen, Eibar Flores, Hendrik_Snijder, and Simon Stier. Big-map/battinfo: v0.6.0, August 2023. URLhttps://doi.org/10.5281/zenodo.8260800

    work page doi:10.5281/zenodo.8260800 2023

  48. [48]

    Cambridge University Press, 2023

    Roman Garnett.Bayesian optimization. Cambridge University Press, 2023

    work page 2023

  49. [49]

    MIT press Cambridge, MA, 2006

    Christopher KI Williams and Carl Edward Rasmussen.Gaussian processes for machine learning, volume 2-3. MIT press Cambridge, MA, 2006

    work page 2006

  50. [50]

    Differentiable expected hypervolume improvement for parallel multi-objective bayesian optimization.Advances in neural information processing systems, 33:9851–9864, 2020

    Samuel Daulton, Maximilian Balandat, and Eytan Bakshy. Differentiable expected hypervolume improvement for parallel multi-objective bayesian optimization.Advances in neural information processing systems, 33:9851–9864, 2020

    work page 2020

  51. [51]

    Human-lab-ai, March 2026

    Giovanna Tosato, Merker Leon, V ogler Monika, Selzer Michael, and Koeppe Arnd. Human-lab-ai, March 2026. URLhttps://doi.org/10.5281/zenodo.19360410. 21 arXivTemplateA PREPRINT Supporting Information 1.4 1e-3 1.2 0.6 Formation Cycling boundary Cell 139 EOL = 1173 Cell 137 EOL = 1758 Cell 140 EOL = 0 Cell 138 EOL = 567 Mean EOL = 874 Cycle 93Cycle 193Cycle ...

    work page doi:10.5281/zenodo.19360410 2026