Search

In the rapidly advancing domain of artificial intelligence applied to materials science, systems are frequently called upon to make critical decisions under conditions of substantial uncertainty, such as selecting which candidate material to synthesize next, which experiment to prioritize for evaluation, or which property to measure in a given campaign. A fundamental aspect that current materials AI approaches largely ignore is the phenomenon of regret—the realization, after the fact, that a different choice would have produced a superior outcome, often carrying emotional, cognitive, and practical costs for the decision-maker. Regret theory, originating in decision theory and economics, provides a powerful alternative lens for understanding choice under uncertainty by incorporating not only expected utilities but also the anticipation and experience of post-decision disappointment or rejoicing. This paper proposes a conceptual framework for regret-aware materials AI systems that explicitly integrates regret quantification, theoretical regret bounds, regret minimization objectives, regret-aware acquisition functions, and regret communication mechanisms into the decision-making pipeline. The framework further delineates four primary types of regret encountered in materials contexts—synthesis regret, measurement regret, discovery regret, and resource regret—each arising from the irreversible, sequential, and high-stakes nature of experimental materials research. By embedding these elements, the proposed framework shifts materials AI from a narrow focus on reward maximization toward systems that more closely mirror the nuanced realities of scientific decision-making, where the avoidance of avoidable regret becomes a central design goal. Ultimately, embracing regret awareness promises more robust exploration of vast material spaces, better alignment between AI recommendations and laboratory constraints, and enhanced trust between human researchers and autonomous systems, thereby accelerating genuine discovery while mitigating the hidden costs of overlooked alternatives.

Materials machine learning has achieved remarkable success in building accurate predictive models for properties such as formation energy, band gap, and mechanical strength. Yet the ultimate purpose of the field is not prediction but design: the discovery of new materials that deliver target properties under real-world constraints. This perspective argues that the community must shift from prediction-centric to optimization-centric workflows. Prediction-centric approaches focus on minimizing mean absolute error on randomly held-out test sets, while optimization-centric workflows aim to maximize (or minimize) a target property value within practical budgets, constraints, and sequential decision-making. These are fundamentally different objectives that demand different methods, evaluation metrics, and research priorities. A prediction-centric model may achieve low error on interpolation tasks yet fail catastrophically when asked to guide the search for extreme or out-of-distribution materials. In contrast, optimization-centric workflows treat the machine-learning model as a surrogate within an iterative loop that actively selects the next experiment or simulation. Four core principles underpin this shift: explicit goal specification that includes both the objective and all relevant constraints; robust constraint handling that distinguishes hard feasibility requirements from soft trade-offs; uncertainty awareness, where every prediction is accompanied by well-calibrated uncertainty estimates essential for balancing exploration and exploitation; and closed-loop integration that connects the model directly to automated or high-throughput experimentation and simulation. Adopting these principles will require new evaluation protocols that measure the best material discovered, sample efficiency, regret, constraint satisfaction, and extrapolation distance rather than isolated accuracy metrics. The implications extend to model development, benchmark design, journal standards, and funding priorities. Only by embracing optimization-centric workflows can materials machine learning fulfill its promise of accelerating discovery and delivering materials that solve pressing societal challenges.

Multi-fidelity learning has been positioned as a transformative tool for high-throughput materials screening, offering the ability to fuse inexpensive low-fidelity data—generated from classical force fields, tight-binding models, or CALPHAD thermodynamics—with costly high-fidelity calculations such as density functional theory (DFT). The central premise is straightforward: surrogate models can learn systematic discrepancies between fidelity levels, delivering near-DFT accuracy at a fraction of the computational expense and thereby enabling the rapid exploration of vast compositional and structural spaces that would otherwise remain inaccessible. A substantial body of work in journals including npj Computational Materials, Machine Learning: Science and Technology, Digital Discovery, Journal of Chemical Theory and Computation, Journal of Chemical Physics, and Nature Machine Intelligence has explored this paradigm through linear corrections, Gaussian process discrepancy modeling, neural network transfer learning, multi-fidelity Bayesian optimization, ensemble approaches, and hierarchical correction chains. Yet the promises remain largely unfulfilled. Despite hundreds of reported case studies, the literature reveals few, if any, experimentally validated novel materials discovered solely through multi-fidelity acceleration. Efficiency gains rarely exceed 2–5× in realistic workflows, uncertainty estimates are frequently miscalibrated outside training domains, and extrapolation to chemically distinct regions consistently collapses to the accuracy of the underlying low-fidelity baseline. Overclaims regarding order-of-magnitude cost reductions, guaranteed high-fidelity fidelity, and accelerated discovery have not materialized in practice. This review provides the first systematic synthesis of the 34 peer-reviewed publications that directly address multi-fidelity surrogate strategies in materials screening. It categorizes surrogate architectures, contrasts theoretical assumptions with empirical performance, inventories six recurring broken promises, and distills actionable lessons from the collective evidence. By foregrounding the persistent gaps between theory and application—non-stationary bias, limited transferability, and evaluation practices that favor interpolation over genuine discovery—this work offers a sober roadmap for future research in computational and data-driven materials engineering. The field must move beyond optimistic benchmarks toward honest reporting of real-world discovery metrics if multi-fidelity learning is to fulfill even a modest fraction of its original vision.

Closed-loop systems have become foundational to computational and data-driven materials engineering, integrating automated experimentation, machine learning inference, and orchestration software to compress the design-make-test-analyze cycle. These pipelines rely on continuous flows of data, models, and decisions, yet the mechanisms governing the transfer of decision authority between human experts and autonomous agents remain conceptually underdeveloped. Existing infrastructures emphasize optimization and execution but offer limited interpretive frameworks for how authority is dynamically delegated across epistemic states and pipeline stages. This manuscript presents the Decision Authority Delegation Cascade (DADC) Framework, an original systems-level architecture that formalizes delegated experimentation as a structured cascade of authority transfer. The framework delineates layered pipelines—from data representation through model inference and steering logics to execution and feedback—while emphasizing infrastructure trade-offs in representation fidelity, uncertainty quantification, and delegation thresholds. Synthesizing advances in Bayesian active learning, self-driving laboratories, and orchestration platforms, the DADC Framework interprets authority transfer not as a binary handover but as a continuous, computationally steered process that modulates discovery dynamics. The framework offers interpretive insights into scalable computational ecosystems, highlighting pathways to align human epistemic oversight with autonomous operation and to mitigate bottlenecks in closed-loop materials discovery. Its application reframes infrastructure design around explicit delegation logics, with implications for the next generation of autonomous materials platforms.