This work established Monte Carlo (MC) Dropout–based multilayer perceptron (MLP) models to predict key electrochemical metrics such as initial capacity and long-term retention from high-dimensional descriptors encompassing composition, structure, and processing features, and demonstrates an iterative human-AI loop.
Abstract
The rapid advancement of autonomous materials research requires data-efficient frameworks that can integrate artificial intelligence (AI) with experimental knowledge and human intuition. In the context of sodium-ion batteries (SIBs), where compositional and processing complexity spans an immense design space, conventional high-throughput experiments still face bottlenecks in data utilization and decision-making efficiency. To address this challenge, we developed an AI-guided, uncertainty-aware workflow that couples high-throughput synthesis and characterization [1-3] with machine learning (ML) surrogate modeling and probabilistic deep learning.
Specifically, we established Monte Carlo (MC) Dropout–based multilayer perceptron (MLP) models to predict key electrochemical metrics such as initial capacity and long-term retention from high-dimensional descriptors encompassing composition, structure, and processing features. The MC-MLP provides not only accurate property predictions but also calibrated uncertainty estimates that quantify model confidence and guide active learning. These models are integrated with gradient boosting regression (GBR) surrogate mappings that link intermediate feature spaces to improve physical interpretability and reduce data sparsity. Tail-focused loss functions and temperature-balanced sampling further enhance sensitivity toward high-performance outliers, accelerating the identification of promising chemistries within limited datasets. By combining these ML strategies with automated synthesis and characterization workflows, we demonstrate an iterative human-AI loop: experimental data continuously refines the surrogate landscape, while uncertainty-driven candidate selection prioritizes new experiments. This approach achieved >3× improvement in data efficiency and enabled Pareto-optimized discovery of Na–Ni–Mn–based layered cathodes with enhanced air stability and electrochemical performance.
Moreover, the framework is transferable across chemical systems, synthesis conditions, and multidimensional battery characterization data, enabling mechanistic understanding and accelerated discovery of novel electrode materials. Overall, this work highlights how integrating MC-Dropout deep learning, surrogate modeling, and domain knowledge from real lab-generated data can transform high-throughput experimentation into a self-improving, autonomous discovery framework—paving the way toward truly intelligent, data-efficient materials research.
References
[1] Jia, Shipeng, et al. "High-throughput design of Na–Fe–Mn–O cathodes for Na-ion batteries." Journal of Materials Chemistry A 10.1 (2022): 251-265.
[2] Jia, Shipeng, et al. "Chemical speed dating: the impact of 52 dopants in Na–Mn–O cathodes." Chemistry of Materials 34.24 (2022): 11047-11061.
[3] Jia, Shipeng, et al. "Stabilization of Na‐Ion Cathode Surfaces: Combinatorial Experiments with Insights from Machine Learning Models." Advanced Energy and Sustainability Research (2024): 2400051
The discovery of advanced alloys capable of withstanding extreme environmental conditions such as high temperatures, intense radiation, corrosive atmospheres, and mechanical stress is critical for applications in aerospace, nuclear energy, deep-sea exploration, and space missions. Traditional experimental and computational approaches to alloy design are often time-consuming and resource-intensive. Recent advances in artificial intelligence (AI) and machine learning (ML) offer powerful tools to accelerate the discovery process by enabling high-throughput screening, property prediction, and design optimization. This paper presents a comprehensive review and methodology for AI-assisted alloy discovery, focusing on the integration of data-driven models with physical principles, high-fidelity simulations, and experimental validation. We highlight successful case studies, discuss the challenges of data scarcity and model interpretability, and propose a framework for closed-loop design that incorporates generative models and active learning. This AI-driven approach represents a paradigm shift toward faster, more cost-effective discovery of next-generation materials for extreme environments.
Venkatesh Iyer, Nandhini Ravi· International Journal of Mod...· 0 citations
Sodium-ion batteries are promising for large-scale energy storage due to resource abundance and low cost. Layered NaxTMO2 cathodes offer high specific capacity and tunable compositions but face challenges in structure-property correlations, sluggish kinetics, and phase instability. First-principles calculations elucidate sodium-storage mechanisms and thermodynamic stability yet struggle with high-dimensional design, finite-temperature effects, and long-timescale dynamics. Artificial intelligence (AI) now enables efficient property prediction, composition screening, and mechanism analysis. Integrating physics-based models with AI establishes an intelligent computing framework-combining mechanistic interpretability and data-driven efficiency-that shifts research from isolated computations to high-throughput, multiscale, closed-loop optimization. This review summarizes advances in theoretical calculations, AI applications, and their convergence toward this paradigm. It highlights how the "physics-constrained modeling-data-driven prediction-experimental feedback" loop addresses high-voltage stability, phase-transition control, and kinetic limitations, enabling targeted cathode design. Future directions include active learning, generative inverse design, and automated experimentation. Overall, intelligent computing is steering layered cathode development from empirical trial-and-error toward demand-oriented rational design.
Qingfeng Liu, Zitong Fei, Jinhua Shi et al.· Small· 0 citations
Redox flow batteries (RFBs) have garnered increasing interest for grid-scale energy storage applications due to their modularity, scalability, and ability to partially or fully decouple energy and power. Iron-based RFBs are of particular interest due to the low cost and earth abundance of iron. However, many iron-based RFB systems suffer from solubility-related limitations, resulting in limited volumetric power density or hindered performance due to precipitation of crossover species. Despite active research, a comprehensive, structured dataset of iron-based electrolyte formulations and properties remains limited. To establish a systematic understanding of electrolyte additive effects, we leverage the capabilities of the Material Innovation through Robotics & AI Laboratory (MIRAL) to develop a standardized protocol for RFB electrolyte formulation and evaluation. This work employs AI-guided high-throughput experimentation to systematically assess the impact of a library of organic and inorganic electrolyte additives. By employing a closed-loop workflow that couples researcher domain expertise with AI-driven predictions, the protocol accelerates identification and down-selection of ferrocyanide electrolyte formulations with enhanced solubility and ferrous chloride electrolyte formulations with improved pH stability in the presence of crossover species.
H. Petersen, Heather M. Job, Aaron Hollas et al.· ECS Meeting Abstracts· 0 citations
Lithium-ion batteries are among the most complex energy storage technologies, with performance driven by intricate interactions across diverse materials and operating conditions. Despite advances in modeling, predicting battery behavior from individual components remains a challenge, making physical experimentation essential. To accelerate discovery, Wildcat has developed a high-throughput research platform that generates high-quality, reproducible, and cross-comparable data; more importantly, including both successful and unsuccessful results. This comprehensive dataset provides a strong foundation for predictive modeling. Building on this, we are integrating artificial intelligence to uncover patterns, optimize designs, and enhance research efficiency. In this presentation, we will share Wildcat’s data-driven approach to battery innovation, highlight the capabilities of our high-throughput system, and discuss key insights from deploying machine learning models to transform Lithium-ion battery research.
Manganese-based layered oxides have emerged as promising cathode materials for potassium-ion batteries owing to their low cost, environmental benignity, structural diversity, and high theoretical capacity, yet suffer from poor rate performance. Compositional regulation offers an effective strategy to address this limitation, but traditional trial-and-error approaches are inefficient and often yield suboptimal results. Herein, interpretable machine learning combining Random Forest, high-throughput virtual screening, and Pareto-front optimization of capacities at 50 and 500 mA g-1 is employed to identify promising Mn-based layered cathodes. Two complementary parameters, Peukert capacity and Peukert-derived rate-decay descriptor, offer insight into capacity and rate-dependent behavior. Screening 14,558 candidate compositions derived from 22 dopant elements reveals compositional regions that balance practical capacity and rate performance. The model-recommended composition delivers approximately 67 mAh g-1 at 500 mA g-1 and retains over 70% of its initial capacity after 500 cycles. This work demonstrates an interpretable and applicability-aware methodological framework for Pareto-guided screening of high-rate cathodes, providing generalizable insights for accelerating the discovery of advanced energy-storage materials.
Qinggang Yue, Yingjiao Zhang, Juanjuan Cheng et al.· Small Methods· 0 citations
The solid-state electrolytes (SSE) are gaining tremendous attention in designing rechargeable batteries with remarkable energy density and safety features for next-generation energy storage device applications. The rational design of SSE with promising ionic conductivity, higher electrochemical stability windows, and stable electrode-electrolyte interfaces remain a formidable challenge, traditionally hindered by trial-and-error experimentation and computationally expensive theoretical simulations. Here, we systematically review the recent breakthroughs in the artificial intelligence (AI)-driven design of SSE, spanning electrochemical stability and ionic conductivity domains, with a particular focus on how machine learning (ML) and deep learning (DL) are fundamentally transforming the discovery and optimization landscape. We critically discuss the synergy between first-principles density functional theory (DFT), molecular dynamics (MD) simulations, and advanced AI algorithms including supervised and unsupervised learning (SL, UL), graph neural networks (GNNs), and Machine Learning Interatomic Potentials (MLIP) that collectively enable accurate prediction of ionic conductivity, elucidation of ion transport mechanisms, and high-throughput screening (HTS) of vast chemical spaces. Emphasis is placed on descriptor engineering that bridges atomic-level structural features (e.g., lattice parameters, activation energies, defect chemistry) with macroscopic electrochemical performance, as well as the emerging paradigm of closed-loop, self-driving laboratories for autonomous materials discovery. Furthermore, AI-guided strategies have demonstrated remarkable interfacial ionic transport mechanism. Despite these transformative advances, persistent challenges including data scarcity, limited descriptor transferability, discrepancies between theoretical predictions and experimental realization, remain significant challenges. Looking forward, the convergence of AI with high-throughput experimentation and multiscale modeling promises to redefine SSE discovery, accelerating the deployment of high-performance all solid-state batteries (ASSBs) for sustainable energy storage.
Jiaying He, Zama Jan, Heqin Guo et al.· AI for Materials· 0 citations