{"id":"7b576709-3ef3-49a3-8ce4-4def925856a0","arxiv_id":"2603.05238","paper_version":2,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Multi-fidelity ML interatomic potentials with global defect charge embeddings recover charged-defect structures and thermodynamics in Sb2Se3 in quantitative agreement with hybrid DFT at much lower cost.","lead":"Researchers built machine-learning force fields that handle charged defects in the semiconductor Sb2Se3 by adding charge-state embeddings and training on mixed cheap and expensive quantum data. This could make accurate defect thermodynamics far cheaper than pure hybrid DFT for materials design.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"Global charge embeddings plus multi-fidelity mixing may not capture charge-state-specific local bonding and long-range electrostatics without explicit electronic degrees of freedom.","rationale":"The reader’s weakest-assumption statement already isolates the precise load-bearing premise (global charge tag + multi-fidelity sufficiency without explicit electrons). No stronger internal contradiction is visible in the abstract; the methodological claim is coherent and the quantitative-agreement language is clear. Because energies, forces, ablations, electrostatic treatment and code remain unavailable, the concern cannot be resolved further, so the CONDITIONAL verdict with low confidence is unchanged. The proposed ablation directly tests whether the embeddings are essential to the claimed hybrid-level accuracy.","tokens_in":2040,"tokens_out":476,"duration_ms":20486,"concrete_test":"When full paper, data and code appear, retrain an otherwise identical multi-fidelity model that omits the global charge embeddings; recompute formation-energy diagrams and relaxed geometries for at least two charge states of a representative defect (e.g. V_Se) against the same hybrid reference. If the no-embedding errors stay within ~0.1 eV / 10 % of the embedded model, the embeddings are not load-bearing as claimed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that a global (system-level) defect-charge embedding is sufficient for an MLIP to learn distinct bonding and energy landscapes of different charge states in Sb2Se3, and that mixing semi-local reference data with hybrid energies/forces improves rather than pollutes hybrid-level defect thermodynamics. This premise is least secure because charged point defects typically involve local charge localization (or polarons) and long-range 1/r electrostatics; a short-range MLIP that receives only a global tag has no explicit mechanism for either, yet the abstract asserts the embeddings “distinguish the bonding characteristics of different charge states” and that the multi-fidelity mix “describe[s] well the subtleties of the defect energy landscape.” Without ablations, electrostatic corrections, or charge-density diagnostics, it remains unclear whether the reported quantitative agreement with hybrid DFT is robust or fortuitous for the chosen defects.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript addresses charged point defects in machine-learning interatomic potentials (MLIPs). It reports that current foundation MLIPs fail to describe the defect physics of the semiconductor Sb2Se3. The authors introduce global defect charge embeddings intended to distinguish bonding of different charge states, and a multi-fidelity strategy that mixes low-cost semi-local DFT data with hybrid-functional energies and forces. They claim the resulting potentials recover stable defect configurations and defect thermodynamics in quantitative agreement with direct hybrid DFT, at a fraction of the cost.","tokens_in":2243,"tokens_out":922,"duration_ms":14540,"significance":"Charged point defects control electronic and optoelectronic behavior in semiconductors; scalable MLIPs that reach hybrid-level defect thermodynamics would be practically valuable for screening and device-relevant defect engineering. Targeting a documented failure of foundation models, and combining multi-fidelity training with an explicit charge-state inductive bias, is a sensible methodological direction. If the claimed hybrid-level quantitative agreement is supported by rigorous, ablated validation (errors, baselines, electrostatic handling), the work would be a useful advance for defect-capable force fields. On the abstract alone those supporting results cannot be assessed.","major_comments":[{"comment":"Only the abstract is available for review. The central claim of “quantitative agreement with direct quantum mechanical calculations” for structures and defect thermodynamics is therefore not checkable: no error metrics, error bars, training-set sizes, validation protocols, baselines against foundation MLIPs, or comparison tables appear in the provided text. Load-bearing numerical support must be present and scrutinizable before acceptance can be considered.","section":"Abstract"},{"comment":"The abstract asserts that global defect charge embeddings “distinguish the bonding characteristics of different charge states.” Charged defects typically involve local charge localization (or polarons) and long-range 1/r electrostatics. A short-range MLIP receiving only a system-level charge tag has no explicit mechanism for either. Without ablations (embedding on/off), charge-density or localization diagnostics, and a clear statement of how finite-size/electrostatic corrections are treated (or learned), it is unclear whether reported hybrid-level agreement is robust or fortuitous for the chosen Sb2Se3 defects.","section":"Abstract (global defect charge embeddings)"},{"comment":"The multi-fidelity claim—that mixing semi-local reference data with hybrid energies/forces “describe[s] well the subtleties of the defect energy landscape”—is load-bearing. Semi-local and hybrid functionals can disagree qualitatively on defect localization and level positions. The manuscript must show (e.g., via fidelity-weight ablations or hold-out hybrid-only tests) that low-fidelity data improve rather than pollute hybrid-level defect thermodynamics; that evidence is not available from the abstract.","section":"Abstract (multi-fidelity approach)"}],"minor_comments":[{"comment":"The abstract states that foundation MLIPs “do not describe the defect physics of … Sb2Se3” without specifying which models, which defects/charge states, or which observables failed. Even in the abstract, naming the models and the failure mode (e.g., wrong ground-state geometry, wrong formation energy ordering) would orient the reader.","section":"Abstract"},{"comment":"“Quantitative agreement” and “fraction of the computational cost” should be backed by at least one indicative number or relative cost factor in the abstract once the full results are fixed.","section":"Abstract"}],"recommendation":"uncertain","confidential_remarks":"This report is based solely on the abstract (full text not available). I cannot responsibly recommend accept/minor/major/reject without the methods, data, and validation sections. If the full manuscript is supplied, the main risks to re-check are: (i) whether global charge embeddings plus short-range MLIPs actually capture localization and long-range electrostatics, (ii) multi-fidelity ablations showing hybrid-level defect energetics are not polluted, and (iii) explicit treatment of charged-defect finite-size corrections. Scope (cond-mat.mtrl-sci / MLIP methods) appears appropriate if the claims hold."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing to know is that this is an abstract-only methods paper. The claim is clear: foundation MLIPs fail on charged defects in Sb2Se3, and a multi-fidelity potential that mixes semi-local and hybrid DFT data plus a global defect-charge embedding recovers hybrid-level structures and defect thermodynamics at MLIP cost. That is a useful target if it holds.\n\nWhat looks new is the combination, not either piece alone. Multi-fidelity MLIPs and charge-aware potentials already exist; the distinctive move is global charge embeddings aimed specifically at distinguishing charge-state bonding for point defects, demonstrated where foundation models break. The workflow is standard supervised training against external DFT references, so circularity is low. They are honest that bulk-trained foundation models do not transfer to these defects, which is a real practical problem for photovoltaics and related semiconductors.\n\nThe soft spots are exactly what you expect with only an abstract. No error metrics, no training-set sizes, no baselines beyond “foundation models fail,” no ablations on the embedding or the multi-fidelity mix, no code or data. The stress-test concern is fair: a short-range MLIP that only gets a system-level charge tag has no explicit electrons and no built-in long-range 1/r electrostatics, so it is not obvious that it can capture polaronic localization or charge-state-specific bonding without fortuitous cancellation. That premise is load-bearing and currently untested in the text we have. I would not call it a fatal flaw—just the thing a referee must demand numbers and diagnostics for.\n\nThis is for people who build or use MLIPs for defects in semiconductors. If the full paper ships quantitative agreement tables, electrostatic handling, and released models, it is worth a serious look and possible citation. As it stands I would not cite it yet. It still deserves a proper referee rather than a desk reject: the problem is real, the approach is coherent, and the claim is falsifiable once the numbers appear. Send it out; require the metrics and the ablations.","headline":"Abstract-only methods claim: multi-fidelity MLIPs with global charge tags for charged defects in Sb2Se3; plausible but uncheckable without numbers.","tokens_in":2871,"tokens_out":527,"would_cite":false,"duration_ms":4750,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Global charge embeddings plus multi-fidelity training let ML interatomic potentials handle charged point defects in Sb2Se3","keywords":["multi-fidelity machine learning","interatomic potentials","charged point defects","Sb2Se3","defect thermodynamics","hybrid functionals","charge embeddings","semiconductors"],"falsifier":"An independent hybrid-functional calculation on a charged defect in Sb2Se3 whose formation energy or thermodynamic transition level differs from the multi-fidelity MLIP prediction by more than ~0.1 eV, or whose MLIP-relaxed geometry differs substantially from the hybrid-relaxed geometry.","tokens_in":2884,"feed_emoji":"⚡","tokens_out":780,"duration_ms":18825,"temperature":0.7,"pith_summary":"Machine-learning interatomic potentials now match first-principles energies and forces for perfect crystals, yet they still fail on charged point defects where local coordination and electron counts differ from the bulk. This paper argues that a simple global tag for the overall defect charge state, together with training that mixes cheap semi-local DFT data with a smaller set of accurate hybrid-functional energies and forces, is enough to capture the distinct bonding of each charge state. The resulting potentials locate stable defect geometries and recover defect thermodynamics for the semiconductor Sb2Se3 that agree quantitatively with direct hybrid calculations, at a small fraction of the cost. A reader who cares about semiconductors would care because charged defects set carrier concentrations, recombination rates and optical response; being able to screen them without a hybrid DFT calculation on every candidate structure would make materials design far more practical.","feed_headline":"Charge tags let ML force fields master charged defects","feed_subtitle":"Multi-fidelity training recovers hybrid-level defect thermodynamics in Sb2Se3 at a fraction of the cost","key_machinery":"Global defect-charge embeddings that inject the overall charge state of the supercell as a system-level tag, allowing the potential to distinguish the bonding characteristics of different charge states, combined with multi-fidelity training that blends abundant semi-local DFT data with high-quality hybrid-functional energies and forces.","core_discovery":"Defect-capable multi-fidelity machine-learning interatomic potentials that include global defect-charge embeddings can find stable structural configurations and predict defect thermodynamics for charged point defects in Sb2Se3 in quantitative agreement with direct hybrid-functional quantum-mechanical calculations, at a fraction of the computational cost.","pith_inferences":["The same global-tag idea may transfer to other low-symmetry semiconductors and to defect-migration barriers if the training set is expanded accordingly.","If a global charge label works, local charge-density features may not be strictly required for many defect energy landscapes once the charge state is known.","Multi-fidelity blending of semi-local and hybrid data could become a general strategy for bringing hybrid accuracy into large-scale defect sampling for photovoltaics and thermoelectrics.","Failure of the method on a second, chemically dissimilar material would indicate that the global embedding is material-specific rather than universal."],"forward_implications":["Stable charged-defect geometries in Sb2Se3 can be located by MLIP-driven relaxation rather than repeated hybrid DFT.","Defect formation energies and charge-transition levels become available at hybrid accuracy without a hybrid calculation on every configuration.","Foundation MLIPs that currently fail on defect physics can be upgraded by the same charge-embedding and multi-fidelity recipe.","The approach supplies a practical route to hybrid-level defect thermodynamics for other semiconductors once the corresponding multi-fidelity data are generated."],"fun_headline_variants":["Charge embeddings equip multi-fidelity MLIPs for charged defects","Multi-fidelity ML force fields match hybrid defect thermodynamics","Global charge tags yield hybrid-accurate Sb2Se3 defect ML potentials","Defect-charge embeddings enable cheap hybrid-level defect modeling","Multi-fidelity training recovers charged defect physics in Sb2Se3"],"cache_read_input_tokens":128,"weakest_assumption_plain":"That a single global charge tag, together with a mixture of semi-local and hybrid reference data, is sufficient to capture the distinct bonding and energy landscape of each charge state without needing explicit electronic-structure degrees of freedom.","fun_headline_variants_meta":{"raw":{"variants":["Charge embeddings equip multi-fidelity MLIPs for charged defects","Multi-fidelity ML force fields match hybrid defect thermodynamics","Global charge tags yield hybrid-accurate Sb2Se3 defect ML potentials","Defect-charge embeddings enable cheap hybrid-level defect modeling","Multi-fidelity training recovers charged defect physics in Sb2Se3"]},"model":"grok-4.5","effort":"low","cost_usd":0.005962,"raw_usage":{"total_tokens":1487,"prompt_tokens":692,"num_sources_used":0,"completion_tokens":76,"cost_in_usd_ticks":59620000,"prompt_tokens_details":{"text_tokens":692,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":719,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":692,"tokens_out":76,"duration_ms":5272,"temperature":1.0,"reasoning_tokens":719,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-15T14:42:37.893874+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"An independent hybrid-functional calculation on a charged defect in Sb2Se3 whose formation energy or thermodynamic transition level differs from the multi-fidelity MLIP prediction by more than ~0.1 eV, or whose MLIP-relaxed geometry differs substantially from the hybrid-relaxed geometry.","supporting_citations":[],"review_version":1}