Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 13 min ago
PreprintCC BYChemRxiv

Accessible-State Descriptors Transfer Across Families but Fail at Larger Scales

  • Rithvik Kothuri1,2
  1. 1Department of Physics
  2. 2Brown University Providence
Open at sourcePDFDOICite

Abstract

Can descriptions of the states visited during quantum dynamics predict product-formula simulation requirements beyond physical and Hamiltonian controls, and do those relationships survive a change of scale? In a frozen nested leave-one-family-out evaluation of finite matter–light models, normalized accessibility and a combined state-space representation reduced equal-family mean absolute step-count error from 7.4512 to 5.5334 and 4.6771, respectively, although effects were heterogeneous and the primary accessibility increment on a separately registered external family was marginal. On the subsequent strict large-simulation-dimension endpoint, frozen R6 and the later R7 representation each performed worse than the already inaccurate Phase IV physical baseline, with the resolved-fraction support gate remaining inconclusive. These results support a bounded predictive association within the tested finite regime, not a state-space complexity law, a causal explanation, or an extrapolatable hardware-resource estimate.

Related

  1. Charge-Transfer-State Descriptors from Donor–Acceptor Interfaces Enable Improved Prediction of Indoor Organic Photovoltaic Efficiency
    Anagh Mukherjee
    ChemRxivPreprintCC BYFull textGeneral chemistryChemRxiv
  2. Optimal sparging and mixing for enhanced mass transfer in U-loop bioreactors across scales
    Malik Hassanaly, Pranav Jain, Mohammad J Rahimi +3
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  3. Evaluating transfer across modalities in autonomous materials systems
    Kinston Ackölf, Taylor D Sparks
    ChemRxivPreprintCC BYFull textMaterials & polymersChemRxiv
  4. Limits of geometric packing descriptors in molecular crystal energy ranking
    Pavlo Bondarenko
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  5. A Machine Learning Approach to logBB Prediction Combining Molecular Docking and Quantum-Chemical Descriptors
    Lilian Rouille
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv

Cite

BibTeXRIS
BibTeX
@misc{kothuri2026accessiblestate,
  title = {{Accessible-State Descriptors Transfer Across Families but Fail at Larger Scales}},
  author = {Rithvik Kothuri},
  howpublished = {ChemRxiv},
  year = {2026},
  month = oct,
  doi = {10.26434/chemrxiv.15011065/v1},
  url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15011065/v1},
  note = {CC BY},
}
RIS
TY  - UNPB
TI  - Accessible-State Descriptors Transfer Across Families but Fail at Larger Scales
AU  - Rithvik Kothuri
PY  - 2026
DA  - 2026/10/26
PB  - ChemRxiv
DO  - 10.26434/chemrxiv.15011065/v1
UR  - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15011065/v1
AB  - Can descriptions of the states visited during quantum dynamics predict product-formula simulation requirements beyond physical and Hamiltonian controls, and do those relationships survive a change of scale? In a frozen nested leave-one-family-out evaluation of finite matter–light models, normalized accessibility and a combined state-space representation reduced equal-family mean absolute step-count error from 7.4512 to 5.5334 and 4.6771, respectively, although effects were heterogeneous and the primary accessibility increment on a separately registered external family was marginal. On the subsequent strict large-simulation-dimension endpoint, frozen R6 and the later R7 representation each performed worse than the already inaccurate Phase IV physical baseline, with the resolved-fraction support gate remaining inconclusive. These results support a bounded predictive association within the tested finite regime, not a state-space complexity law, a causal explanation, or an extrapolatable hardware-resource estimate.
ER  -