pith. sign in

arxiv: 2606.30170 · v1 · pith:MJIIHU6Wnew · submitted 2026-06-29 · 💻 cs.LG · cond-mat.mes-hall· cond-mat.mtrl-sci· cs.AI· cs.CE

Beyond Drug Discovery: The Nanotechnology Molecular Optimization (NMO) Benchmark

Matthias Blaschke , Daniel Kienzle , Zsuzsanna Koczor-Benda , Julian Lorenz , Rainer Lienhart , Fabian Pauly This is my paper

Pith reviewed 2026-06-30 06:45 UTC · model grok-4.3

classification 💻 cs.LG cond-mat.mes-hallcond-mat.mtrl-scics.AIcs.CE
keywords nanotechnologymolecular optimizationgenerative modelsquantum simulationsbenchmarkstructural constraintsstructural motifspretraining
0
0 comments X

The pith

The NMO benchmark shows advanced generative models underperform simpler approaches on physics-based nanotechnology tasks.

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

The paper introduces the Nanotechnology Molecular Optimization (NMO) Benchmark to move molecular generative design beyond drug-like proxy tests toward real nanotechnology targets. It replaces those proxies with quantum simulations as oracles and adds strict protocols that prioritize scientific utility over leaderboard scores. The resulting tasks feature hard structural constraints and rugged fitness landscapes. On these tasks advanced optimization methods lag behind simpler baselines, which prompts a new baseline method using a novel representation for constraints plus domain-agnostic pretraining that removes pharmaceutical bias. This baseline improves physical properties and identifies previously unknown structural motifs.

Core claim

The NMO Benchmark acts as both an ML testbed and a nanotechnology discovery engine by replacing proxy oracles with quantum simulations and enforcing strict protocols that prioritize scientific utility. The physics-based tasks impose hard structural constraints and rugged fitness landscapes that pose fundamentally new requirements on generative models. Advanced molecular optimization methods underperform much simpler approaches, yet a new baseline method that introduces a novel representation for modeling structural constraints and a domain-agnostic pretraining strategy surpasses prior state-of-the-art physical properties and reveals previously unknown structural motifs.

What carries the argument

The NMO Benchmark suite that uses quantum simulations as oracles under strict protocols, together with the new baseline method's novel representation for modeling structural constraints and domain-agnostic pretraining strategy.

If this is right

  • Generative models must acquire new capabilities to navigate rugged fitness landscapes and hard structural constraints in physics-based domains.
  • Eliminating pharmaceutical dataset bias through domain-agnostic pretraining improves performance on non-drug molecular targets.
  • Machine learning optimization can directly contribute to uncovering new structural motifs relevant to nanotechnology.
  • Strict evaluation protocols can reduce overfitting to benchmarks and steer research toward genuine scientific utility.

Where Pith is reading between the lines

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

  • Benchmark designs that embed domain-specific oracles and constraints could be adapted to other scientific areas such as catalysis or materials discovery.
  • The newly identified structural motifs offer concrete candidates for laboratory synthesis and device testing in nanotechnology.
  • Domain-agnostic pretraining approaches may improve transferability of molecular models across multiple application fields beyond the original training distribution.

Load-bearing premise

Quantum simulations serve as faithful oracles for real nanotechnology targets and the strict protocols successfully prioritize scientific utility without introducing selection biases or overfitting risks.

What would settle it

Synthesizing and physically testing the top molecules found by the NMO baseline and comparing their measured properties against the quantum simulation scores; large mismatches would show the benchmark does not translate to real targets.

Figures

Figures reproduced from arXiv: 2606.30170 by Daniel Kienzle, Fabian Pauly, Julian Lorenz, Matthias Blaschke, Rainer Lienhart, Zsuzsanna Koczor-Benda.

Figure 1
Figure 1. Figure 1: Molecular systems: (a) A molecule contacted by gold surfaces on both sides forms a single [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (a) Fragment library. (b) Molecule with GS string representation (see Section A.3.2). (c) [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Top performing molecules surpassing previous literature results for (a) TE ( [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Validation of the phonon transport implementation. Left: Phonon transmission for a set [PITH_FULL_IMAGE:figures/full_fig_p019_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: (a) Supercell containing two principal layers (red boxes) for periodic calculation. (b) [PITH_FULL_IMAGE:figures/full_fig_p024_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of PTB and DFT calculated P values for the Gold database of Molecular Vibration Explorer [77] (about 2800 molecules), showing the mean absolute error (MAE) and coefficient of determination (R2 ). The 1:1 line is shown in black while the linear fit to data points is shown in red. To arrive at a molecular-level property, Koczor-Benda et al. [25] introduced the dimensionless quantity P that collect… view at source ↗
Figure 7
Figure 7. Figure 7: Illustration of mutation operations on the GGS graph structure. Each node represents a [PITH_FULL_IMAGE:figures/full_fig_p029_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Single-Point Crossover. Two parent DAGs (left) are partitioned at randomly selected edges [PITH_FULL_IMAGE:figures/full_fig_p030_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Building blocks, sorted by atom count, used for the construction of molecular candidates. [PITH_FULL_IMAGE:figures/full_fig_p031_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Overview of the framework. First, the agent is initialized with the pretraining weights. During optimization, the model first generates a batch of molecules, which are filtered and evaluated by the oracle. High-reward molecules are stored in a replay buffer, from which a GA proposes refined candidates, which are added to the buffer if they are sufficiently good. Finally, a training step is performed on th… view at source ↗
Figure 11
Figure 11. Figure 11: (a) Valid rate of the models and employed encoding according to the legend versus [PITH_FULL_IMAGE:figures/full_fig_p037_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Sampling statistics for the thermoelectric oracle. (a) Fitness plotted during training versus [PITH_FULL_IMAGE:figures/full_fig_p038_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Same as in Figure 12 but for the phonon oracle. [PITH_FULL_IMAGE:figures/full_fig_p038_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Same as in Figure 12 but for the optomechanical oracle. [PITH_FULL_IMAGE:figures/full_fig_p039_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Training stability analysis for the thermoelectric oracle using the transformer model with [PITH_FULL_IMAGE:figures/full_fig_p040_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Quantum transport properties of the best-performing candidate for the thermoelectric [PITH_FULL_IMAGE:figures/full_fig_p042_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Structural formulas for the molecule shown in Figure 3(a) (highlighted by a black box) and [PITH_FULL_IMAGE:figures/full_fig_p042_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: (a) Phononic transmission for the candidate shown in Figure 3(b), along with the calculated [PITH_FULL_IMAGE:figures/full_fig_p043_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Selected top-performing candidates for THz upconversion and their relevant properties, [PITH_FULL_IMAGE:figures/full_fig_p045_19.png] view at source ↗
read the original abstract

Generative molecular design is shaped by simple proxy benchmarks for drug-like properties and models pretrained on large pharmaceutical datasets. This combination yields strong benchmark metrics but limits transferability to domains structurally distinct from drug discovery. To overcome this limitation and drive discovery toward real, scientifically grounded targets, we introduce the Nanotechnology Molecular Optimization (NMO) Benchmark, which bridges machine learning (ML) and quantum materials science. NMO acts simultaneously as a rigorous testbed for the ML community and a discovery engine for nanotechnology research. The suite replaces proxy oracles with quantum simulations and introduces strict protocols that prioritize scientific utility over leaderboard-oriented overfitting. The physics-based NMO tasks impose hard structural constraints and rugged fitness landscapes, posing fundamentally new requirements on generative models. Notably, advanced molecular optimization methods underperform much simpler approaches on the NMO tasks. We develop a new baseline method identifying the critical components to solve the NMO tasks, including a novel representation for modeling structural constraints and a domain-agnostic pretraining strategy to eliminate pharmaceutical dataset bias. Our results surpass state-of-the-art physical properties and reveal previously unknown structural motifs, offering new insights for the nanotechnology community and demonstrating that ML can drive genuine scientific discovery.

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 / 1 minor

Summary. The paper introduces the Nanotechnology Molecular Optimization (NMO) Benchmark as a replacement for drug-like proxy tasks, using quantum simulations as oracles together with strict protocols to enforce scientific utility. It reports that advanced generative molecular optimization methods underperform simpler baselines on these tasks, presents a new baseline incorporating a novel structural-constraint representation and domain-agnostic pretraining, and claims that the resulting models surpass prior state-of-the-art physical properties while revealing previously unknown structural motifs.

Significance. If the quantum oracles and protocols are shown to be faithful, the benchmark could usefully redirect generative modeling research toward rugged, constraint-heavy landscapes outside pharmaceutical chemical space and provide a concrete demonstration of ML-driven discovery in nanotechnology. The reported performance reversal and motif discovery would then constitute a substantive contribution.

major comments (2)
  1. [Abstract] Abstract: the central claim that 'advanced molecular optimization methods underperform much simpler approaches' and that the new baseline 'surpass[es] state-of-the-art physical properties' is presented without any quantitative metrics, tables, or section references, preventing assessment of effect sizes or controls for the claimed reversal.
  2. [Abstract] Abstract: the assertion that quantum simulations replace proxy oracles and that the introduced protocols 'prioritize scientific utility' is not accompanied by any validation (correlation with experiment, sensitivity analysis of simulation parameters, or ablation of protocol rules), which is load-bearing for both the performance claims and the 'previously unknown structural motifs' discovery claim.
minor comments (1)
  1. [Abstract] The abstract uses the phrase 'domain-agnostic pretraining strategy to eliminate pharmaceutical dataset bias' without defining the pretraining corpus or the bias metric employed.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address each major comment below and will revise the manuscript to improve clarity and support for the claims made in the abstract.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that 'advanced molecular optimization methods underperform much simpler approaches' and that the new baseline 'surpass[es] state-of-the-art physical properties' is presented without any quantitative metrics, tables, or section references, preventing assessment of effect sizes or controls for the claimed reversal.

    Authors: We agree that the abstract would be strengthened by including quantitative metrics. In the revised version, we will incorporate specific performance numbers (e.g., the magnitude of underperformance by advanced methods relative to baselines and the improvements in physical properties achieved by the new baseline) along with explicit references to the relevant tables and result sections. This will enable direct assessment of effect sizes and controls. revision: yes

  2. Referee: [Abstract] Abstract: the assertion that quantum simulations replace proxy oracles and that the introduced protocols 'prioritize scientific utility' is not accompanied by any validation (correlation with experiment, sensitivity analysis of simulation parameters, or ablation of protocol rules), which is load-bearing for both the performance claims and the 'previously unknown structural motifs' discovery claim.

    Authors: We will add a sensitivity analysis of simulation parameters and an ablation study of the protocol rules to the methods and results sections to provide direct support for the protocols' utility. The quantum oracles are based on established quantum chemistry methods standard in the field; however, direct correlation with wet-lab experiments lies outside the computational scope of this benchmark paper and would require a separate experimental study. The structural motifs were identified via the optimization trajectories and cross-validated with additional quantum calculations. revision: partial

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper introduces the NMO benchmark by replacing proxy oracles with quantum simulations and presents a new baseline method with a novel representation and domain-agnostic pretraining. No equations, fitted parameters renamed as predictions, or self-citation chains are described that would reduce any claimed result or motif discovery to the inputs by construction. All central claims (performance reversal, new structural motifs) are framed as empirical outcomes on the introduced tasks rather than tautological redefinitions or self-referential fits. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only access prevents enumeration of free parameters, axioms, or invented entities; no equations or methods sections available to audit.

pith-pipeline@v0.9.1-grok · 5775 in / 1070 out tokens · 26179 ms · 2026-06-30T06:45:20.326603+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

106 extracted references · 1 canonical work pages

  1. [1]

    TARTARUS: A benchmarking platform for realistic and practical inverse molecular design,

    A. Nigam, R. Pollice, G. Tom, K. Jorner, J. Willes, L. Thiede, A. Kundaje, and A. Aspuru- Guzik, “TARTARUS: A benchmarking platform for realistic and practical inverse molecular design,” inAdvances in Neural Information Processing Systems (NeurIPS), vol. 36, 2023

    2023

  2. [2]

    GenMol: A drug discovery generalist with discrete diffusion,

    S. Lee, K. Kreis, S. P. Veccham, M. Liu, D. Reidenbach, Y . Peng, S. G. Paliwal, W. Nie, and A. Vahdat, “GenMol: A drug discovery generalist with discrete diffusion,” in International Conference on Machine Learning (ICML), 2025

    2025

  3. [3]

    Refine drugs, don’t complete them: Uniform-source discrete flows for fragment-based drug discovery,

    B. Kaech, L. Wyss, K. Borgwardt, and G. Grasso, “Refine drugs, don’t complete them: Uniform-source discrete flows for fragment-based drug discovery,” in The Fourteenth International Conference on Learning Representations (ICLR), 2026

    2026

  4. [4]

    Molecule generation with fragment retrieval augmentation,

    S. Lee, K. Kreis, S. P. Veccham, M. Liu, D. Reidenbach, S. Paliwal, A. Vahdat, and W. Nie, “Molecule generation with fragment retrieval augmentation,” inAdvances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  5. [5]

    Sample efficiency matters: a benchmark for practical molecular optimization,

    W. Gao, T. Fu, J. Sun, and C. Coley, “Sample efficiency matters: a benchmark for practical molecular optimization,” Advances in Neural Information Processing Systems (NeurIPS), vol. 35, 2022

    2022

  6. [6]

    Evaluation gaps in machine learning practice,

    B. Hutchinson, N. Rostamzadeh, C. Greer, K. Heller, and V . Prabhakaran, “Evaluation gaps in machine learning practice,” in Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, 2022, p. 1859–1876

    2022

  7. [7]

    Self-assembled monolayers of thiolates on metals as a form of nanotechnology,

    J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, and G. M. Whitesides, “Self-assembled monolayers of thiolates on metals as a form of nanotechnology,”Chemical Reviews, vol. 105, no. 4, pp. 1103–1170, 2005

    2005

  8. [8]

    Design and synthesis of an exceptionally stable and highly porous metal-organic framework,

    H. Li, M. Eddaoudi, M. O’Keeffe, and O. M. Yaghi, “Design and synthesis of an exceptionally stable and highly porous metal-organic framework,”Nature, vol. 402, no. 6759, pp. 276–279, 1999

    1999

  9. [9]

    Phonon interference in single-molecule junctions,

    S. C. Yelishala, Y . Zhu, P. Martinez, H. Chen, M. Habibi, G. Prampolini, J. C. Cuevas, W. Zhang, J. Vilhena, and L. Cui, “Phonon interference in single-molecule junctions,”Nature Materials, vol. 24, pp. 1258–1264, 2025

    2025

  10. [10]

    Tuning phonon transmission via single-atom substituents,

    Y . Luan, M. Blaschke, Y . Isshiki, J. Guan, F. Pauly, E. Meyhofer, and P. Reddy, “Tuning phonon transmission via single-atom substituents,”Nature Materials, 2026

    2026

  11. [11]

    Full thermoelectric characterization of a single molecule,

    A. Gemma, F. Tabatabaei, U. Drechsler, A. Zulji, H. Dekkiche, N. Mosso, T. Niehaus, M. R. Bryce, S. Merabia, and B. Gotsmann, “Full thermoelectric characterization of a single molecule,”Nature Communications, vol. 14, no. 1, p. 3868, 2023

    2023

  12. [12]

    Detecting mid-infrared light by molecular frequency upconversion in dual-wavelength nanoantennas,

    A. Xomalis, X. Zheng, R. Chikkaraddy, Z. Koczor-Benda, E. Miele, E. Rosta, G. A. E. Vandenbosch, A. Mart´ınez, and J. J. Baumberg, “Detecting mid-infrared light by molecular frequency upconversion in dual-wavelength nanoantennas,”Science, vol. 374, no. 6572, pp. 1268–1271, 2021

    2021

  13. [13]

    Genetic algorithms are strong baselines for molecule generation.arXiv preprint arXiv:2310.09267, 2023

    A. Tripp and J. M. Hern´andez-Lobato, “Genetic algorithms are strong baselines for molecule generation,”arXiv preprint arXiv:2310.09267, 2023

    work page arXiv 2023

  14. [14]

    A graph-based genetic algorithm and generative model/monte carlo tree search for the exploration of chemical space,

    J. H. Jensen, “A graph-based genetic algorithm and generative model/monte carlo tree search for the exploration of chemical space,” Chemical Science, vol. 10, no. 12, pp. 3567–3572, 2019

    2019

  15. [15]

    Genetic-guided GFlowNets for sample efficient molecular optimization,

    H. Kim, M. Kim, S. Choi, and J. Park, “Genetic-guided GFlowNets for sample efficient molecular optimization,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 37, 2024

    2024

  16. [16]

    Gotta be safe: a new framework for molecular design,

    E. Noutahi, C. Gabellini, M. Craig, J. S. C. Lim, and P. Tossou, “Gotta be safe: a new framework for molecular design,”Digital Discovery, vol. 3, pp. 796–804, 2024. 11

    2024

  17. [17]

    Molecular de-novo design through deep reinforcement learning,

    M. Olivecrona, T. Blaschke, O. Engkvist, and H. Chen, “Molecular de-novo design through deep reinforcement learning,”Journal of Cheminformatics, vol. 9, no. 1, 2017

    2017

  18. [18]

    Flow network based generative models for non-iterative diverse candidate generation,

    E. Bengio, M. Jain, M. Korablyov, D. Precup, and Y . Bengio, “Flow network based generative models for non-iterative diverse candidate generation,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 34, 2021

    2021

  19. [19]

    Biological sequence design with GFlowNets,

    M. Jain, E. Bengio, A. Hernandez-Garcia, J. Rector-Brooks, B. F. Dossou, C. A. Ekbote, J. Fu, T. Zhang, M. Kilgour, D. Zhang et al., “Biological sequence design with GFlowNets,” in International Conference on Machine Learning (ICML), 2022, pp. 9786–9801

    2022

  20. [20]

    Scaling deep learning for materials discovery,

    A. Merchant, S. Batzner, S. S. Schoenholz, M. Aykol, G. Cheon, and E. D. Cubuk, “Scaling deep learning for materials discovery,”Nature, vol. 624, no. 7990, pp. 80–85, 2023

    2023

  21. [21]

    Structural constraint integration in a generative model for the discovery of quantum materials,

    R. Okabe, M. Cheng, A. Chotrattanapituk, M. Mandal, K. Mak, D. C´ordova Carrizales, N. T. Hung, X. Fu, B. Han, Y . Wang, W. Xie, R. J. Cava, T. S. Jaakkola, Y . Cheng, and M. Li, “Structural constraint integration in a generative model for the discovery of quantum materials,” Nature Materials, 2025

    2025

  22. [22]

    Inverse design of metal-organic frameworks using deep dreaming approaches,

    C. Cleeton and L. Sarkisov, “Inverse design of metal-organic frameworks using deep dreaming approaches,”Nature Communications, vol. 16, no. 1, p. 4806, 2025

    2025

  23. [23]

    Revealing molecule-internal mechanisms that control phonon heat transport through single-molecule junctions by a genetic algorithm,

    M. Blaschke and F. Pauly, “Revealing molecule-internal mechanisms that control phonon heat transport through single-molecule junctions by a genetic algorithm,”ACS Nano, vol. 19, no. 36, pp. 32 093–32 107, 2025

    2025

  24. [24]

    Designing mechanosensitive molecules from molecular building blocks: A genetic algorithm-based approach,

    M. Blaschke and F. Pauly, “Designing mechanosensitive molecules from molecular building blocks: A genetic algorithm-based approach,” The Journal of Chemical Physics, vol. 159, no. 2, p. 024126, 2023

    2023

  25. [25]

    Molecular screening for terahertz detection with machine-learning-based methods,

    Z. Koczor-Benda, A. L. Boehmke, A. Xomalis, R. Arul, C. Readman, J. J. Baumberg, and E. Rosta, “Molecular screening for terahertz detection with machine-learning-based methods,” Physical Review X, vol. 11, p. 041035, 2021

    2021

  26. [26]

    Generative design of functional organic molecules for terahertz radiation detection,

    Z. Koczor-Benda, S. Chaudhuri, J. Gilkes, F. Bartucca, L. Li, and R. J. Maurer, “Generative design of functional organic molecules for terahertz radiation detection,”Digital Discovery, vol. 4, pp. 2852–2863, 2025

    2025

  27. [27]

    Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules,

    N. Gebauer, M. Gastegger, and K. Sch ¨utt, “Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules,”Advances in Neural Information Processing Systems (NeurIPS), vol. 32, 2019

    2019

  28. [28]

    SMILES, a chemical language and information system. 1. introduction to methodology and encoding rules,

    D. Weininger, “SMILES, a chemical language and information system. 1. introduction to methodology and encoding rules,”Journal of Chemical Information and Computer Sciences, vol. 28, no. 1, pp. 31–36, 1988

    1988

  29. [29]

    Self-referencing embed- ded strings (SELFIES): A 100% robust molecular string representation,

    M. Krenn, F. H¨ase, A. Nigam, P. Friederich, and A. Aspuru-Guzik, “Self-referencing embed- ded strings (SELFIES): A 100% robust molecular string representation,”Machine Learning: Science and Technology, vol. 1, no. 4, p. 045024, 2020

    2020

  30. [30]

    Group SELFIES: a robust fragment-based molecular string representation,

    A. H. Cheng, A. Cai, S. Miret, G. Malkomes, M. Phielipp, and A. Aspuru-Guzik, “Group SELFIES: a robust fragment-based molecular string representation,” Digital Discovery, vol. 2, no. 3, pp. 748–758, 2023

    2023

  31. [31]

    Improving the reliability of molecular string representations for generative chemistry,

    E. Reboul, Z. Wefers, H. Prabakaran, J. Waldispuhl, and A. Taly, “Improving the reliability of molecular string representations for generative chemistry,”Journal of Chemical Information and Modeling, vol. 65, no. 19, pp. 10 221–10 238, 2025

    2025

  32. [32]

    On the art of compiling and using ’drug-like’ chemical fragment spaces,

    J. Degen, C. Wegscheid-Gerlach, A. Zaliani, and M. Rarey, “On the art of compiling and using ’drug-like’ chemical fragment spaces,”ChemMedChem, vol. 3, no. 10, pp. 1503–1507, 2008

    2008

  33. [33]

    Extended tight-binding quantum chemistry methods,

    C. Bannwarth, E. Caldeweyher, S. Ehlert, A. Hansen, P. Pracht, J. Seibert, S. Spicher, and S. Grimme, “Extended tight-binding quantum chemistry methods,” WIREs Computational Molecular Science, vol. 11, no. 2, p. e1493, 2021. 12

    2021

  34. [34]

    Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions,

    P. Ertl and A. Schuffenhauer, “Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions,” Journal of Cheminformatics, vol. 1, 2009

    2009

  35. [35]

    Nanoscale thermal transport. II. 2003–2012,

    D. G. Cahill, P. V . Braun, G. Chen, D. R. Clarke, S. Fan, K. E. Goodson, P. Keblinski, W. P. King, G. D. Mahan, A. Majumdaret al., “Nanoscale thermal transport. II. 2003–2012,”Applied Physics Reviews, vol. 1, no. 1, p. 011305, 2014

    2003

  36. [36]

    A brief history of molecular electronics,

    M. Ratner, “A brief history of molecular electronics,”Nature Nanotechnology, vol. 8, no. 6, pp. 378–381, 2013

    2013

  37. [37]

    Thermoelectricity in molecular junctions,

    P. Reddy, S.-Y . Jang, R. A. Segalman, and A. Majumdar, “Thermoelectricity in molecular junctions,”Science, vol. 315, no. 5818, pp. 1568–1571, 2007

    2007

  38. [38]

    Terahertz technology and its applications,

    A. Y . Pawar, D. D. Sonawane, K. B. Erande, and D. V . Derle, “Terahertz technology and its applications,”Drug Invention Today, vol. 5, no. 2, pp. 157–163, 2013

    2013

  39. [39]

    Molecular platform for frequency upconversion at the single-photon level,

    P. Roelli, D. Martin-Cano, T. J. Kippenberg, and C. Galland, “Molecular platform for frequency upconversion at the single-photon level,”Physical Review X, vol. 10, p. 031057, 2020

    2020

  40. [40]

    Continuous-wave frequency up- conversion with a molecular optomechanical nanocavity,

    W. Chen, P. Roelli, H. Hu, S. Verlekar, S. P. Amirtharaj, A. I. Barreda, T. J. Kippenberg, M. Kovylina, E. Verhagen, A. Mart´ınez, and C. Galland, “Continuous-wave frequency up- conversion with a molecular optomechanical nanocavity,” Science, vol. 374, no. 6572, pp. 1264–1267, 2021

    2021

  41. [41]

    Syba: Bayesian estimation of synthetic accessibility of organic compounds,

    M. V orˇsil´ak, M. Kol ´aˇr, I. ˇCmelo, and D. Svozil, “Syba: Bayesian estimation of synthetic accessibility of organic compounds,”Journal of Cheminformatics, vol. 12, no. 1, p. 35, 2020

    2020

  42. [42]

    Learning to scale logits for temperature-conditional GFlownets,

    M. Kim, J. Ko, T. Yun, D. Zhang, L. Pan, W. C. Kim, J. Park, E. Bengio, and Y . Bengio, “Learning to scale logits for temperature-conditional GFlownets,” inForty-first International Conference on Machine Learning (ICML), 2024

    2024

  43. [43]

    QGFN: Controllable greediness with action values,

    E. Lau, S. Lu, L. Pan, D. Precup, and E. Bengio, “QGFN: Controllable greediness with action values,”Advances in Neural Information Processing Systems (NeurIPS), vol. 37, 2024

    2024

  44. [44]

    Relative trajectory balance is equivalent to Trust-PCL,

    T. Deleu, P. Nouri, Y . Bengio, and D. Precup, “Relative trajectory balance is equivalent to Trust-PCL,” 2nd edition of Frontiers in Probabilistic Inference: Learning meets Sampling, 2025

    2025

  45. [45]

    Adaptive divergence regularized policy optimization for fine-tuning generative models,

    J. Fan, T. Wei, C. Cheng, Y . Chen, and G. Liu, “Adaptive divergence regularized policy optimization for fine-tuning generative models,” inThe Thirty-ninth Annual Conference on Neural Information Processing Systems (NeurIPS), 2025

    2025

  46. [46]

    A roadmap for molecular thermoelectricity,

    A. Gemma and B. Gotsmann, “A roadmap for molecular thermoelectricity,” Nature Nanotechnology, vol. 16, no. 12, pp. 1299–1301, 2021

    2021

  47. [47]

    G-xTB: a general-purpose extended tight-binding electronic structure method for the elements H to Lr (Z= 1–103),

    T. Froitzheim, M. M¨uller, A. Hansen, and S. Grimme, “G-xTB: a general-purpose extended tight-binding electronic structure method for the elements H to Lr (Z= 1–103),” 2025

    2025

  48. [48]

    Thermal conductance of single-molecule junctions,

    L. Cui, S. Hur, Z. A. Akbar, J. C. Kl ¨ockner, W. Jeong, F. Pauly, S.-Y . Jang, P. Reddy, and E. Meyhofer, “Thermal conductance of single-molecule junctions,”Nature, vol. 572, no. 7771, pp. 628–633, 2019

    2019

  49. [49]

    Thermal transport through single-molecule junctions,

    N. Mosso, H. Sadeghi, A. Gemma, S. Sangtarash, U. Drechsler, C. Lambert, and B. Gotsmann, “Thermal transport through single-molecule junctions,” Nano Letters, vol. 19, no. 11, pp. 7614–7622, 2019

    2019

  50. [50]

    Tuning the thermal conductance of molecular junctions with interference effects,

    J.-C. Kloeckner, J. C. Cuevas, and F. Pauly, “Tuning the thermal conductance of molecular junctions with interference effects,”Physical Review B, vol. 96, no. 24, p. 245419, 2017

    2017

  51. [51]

    Rules of connectivity-dependent phonon interference in molecular junctions,

    L. Zheng, E. Norouzi Farahani, A. H. S. Daaoub, S. Sangtarash, and H. Sadeghi, “Rules of connectivity-dependent phonon interference in molecular junctions,” Nano Letters, vol. 25, no. 16, pp. 6524–6529, 2025. 13

    2025

  52. [52]

    Phonon interference effects in molecular junctions,

    T. Markussen, “Phonon interference effects in molecular junctions,”The Journal of Chemical Physics, vol. 139, no. 24, p. 244101, 2013

    2013

  53. [53]

    Absence of backscattering in the quantum Hall effect in multiprobe conductors,

    M. B¨uttiker, “Absence of backscattering in the quantum Hall effect in multiprobe conductors,” Physical Review B, vol. 38, no. 14, p. 9375, 1988

    1988

  54. [54]

    J. C. Cuevas and E. Scheer, Molecular Electronics, 2nd ed. World Scientific, 2017

    2017

  55. [55]

    Nonequilibrium Green’s function method for quantum thermal transport,

    J.-S. Wang, B. K. Agarwalla, H. Li, and J. Thingna, “Nonequilibrium Green’s function method for quantum thermal transport,”Frontiers of Physics, vol. 9, pp. 673–697, 2014

    2014

  56. [56]

    First-principles calculation of the thermo- electric figure of merit for [2,2]paracyclophane-based single-molecule junctions,

    M. B¨urkle, T. J. Hellmuth, F. Pauly, and Y . Asai, “First-principles calculation of the thermo- electric figure of merit for [2,2]paracyclophane-based single-molecule junctions,” Physical Review B, vol. 91, no. 16, p. 165419, 2015

    2015

  57. [57]

    Anharmonic phonon flow through molecular-sized junctions,

    N. Mingo, “Anharmonic phonon flow through molecular-sized junctions,”Physical Review B, vol. 74, no. 12, p. 125402, 2006

    2006

  58. [58]

    Mechanical properties and formation mechanisms of a wire of single gold atoms,

    G. Rubio-Bollinger, S. R. Bahn, N. Agra ¨ıt, K. W. Jacobsen, and S. Vieira, “Mechanical properties and formation mechanisms of a wire of single gold atoms,”Physical Review Letters, vol. 87, no. 2, p. 026101, 2001

    2001

  59. [59]

    Electrical properties and mechanical stability of anchoring groups for single-molecule electronics,

    R. Frisenda, S. Tarkuc ¸, E. Gal´an, M. L. Perrin, R. Eelkema, F. C. Grozema, and H. S. van der Zant, “Electrical properties and mechanical stability of anchoring groups for single-molecule electronics,”Beilstein Journal of Nanotechnology, vol. 6, no. 1, pp. 1558–1567, 2015

    2015

  60. [60]

    Mechanoelectric sensitivity reveals destructive quantum interference in single-molecule junctions,

    S. van der Poel, J. Hurtado-Gallego, M. Blaschke, R. L´opez-Nebreda, A. Gallego, M. Mayor, F. Pauly, H. S. van der Zant, and N. Agra¨ıt, “Mechanoelectric sensitivity reveals destructive quantum interference in single-molecule junctions,” Nature Communications, vol. 15, no. 1, p. 10097, 2024

    2024

  61. [61]

    Lower limit of thermal conductivity: diffusion versus localization,

    A. Majumdar, “Lower limit of thermal conductivity: diffusion versus localization,”Microscale Thermophysical Engineering, vol. 2, no. 1, pp. 5–9, 1998

    1998

  62. [62]

    dxtb—an efficient and fully differentiable framework for extended tight-binding,

    M. Friede, C. H¨olzer, S. Ehlert, and S. Grimme, “dxtb—an efficient and fully differentiable framework for extended tight-binding,” The Journal of Chemical Physics, vol. 161, no. 6, 2024

    2024

  63. [63]

    Cluster-based density-functional approach to quantum transport through molecular and atomic contacts,

    F. Pauly, J. K. Viljas, U. Huniar, M. H¨afner, S. Wohlthat, M. B¨urkle, J. C. Cuevas, and G. Sch¨on, “Cluster-based density-functional approach to quantum transport through molecular and atomic contacts,”New Journal of Physics, vol. 10, no. 12, p. 125019, 2008

    2008

  64. [64]

    Applicability of the wide-band limit in DFT-based molecular transport calculations,

    C. Verzijl, J. Seldenthuis, and J. Thijssen, “Applicability of the wide-band limit in DFT-based molecular transport calculations,”The Journal of Chemical Physics, vol. 138, no. 9, p. 094102, 2013

    2013

  65. [65]

    Highly convergent schemes for the calculation of bulk and surface Green functions,

    M. P. L. Sancho, J. M. L. Sancho, J. M. L. Sancho, and J. Rubio, “Highly convergent schemes for the calculation of bulk and surface Green functions,”Journal of Physics F: Metal Physics, vol. 15, no. 4, p. 851, Apr. 1985

    1985

  66. [66]

    A three-tiered hierarchical computational framework bridging molecular systems and junction-level charge transport,

    X. Ji, Q. Qi, Y . Chen, C. Zhou, and X. Yu, “A three-tiered hierarchical computational framework bridging molecular systems and junction-level charge transport,”Journal of Chemical Theory and Computation, vol. 21, no. 6, pp. 2961–2976, 2025

    2025

  67. [67]

    DFTB+, a sparse matrix-based implementation of the DFTB method,

    B. Aradi, B. Hourahine, and T. Frauenheim, “DFTB+, a sparse matrix-based implementation of the DFTB method,” The Journal of Physical Chemistry A, vol. 111, no. 26, pp. 5678–5684, 2007

    2007

  68. [68]

    How substituents tune quantum interference in meta-OPE3 molecular junctions to control thermoelectric transport,

    S. Yan, Y . Luan, H. Xu, H. Fan, L. Martin, A. K. Gupta, H. Linke, E. Meyhofer, P. Reddy, F. Paulyet al., “How substituents tune quantum interference in meta-OPE3 molecular junctions to control thermoelectric transport,”Nanoscale, vol. 16, no. 29, pp. 13 905–13 914, 2024

    2024

  69. [69]

    Quantum interference enhances the performance of single- molecule transistors,

    Z. Chen, I. M. Grace, S. L. Woltering, L. Chen, A. Gee, J. Baugh, G. A. D. Briggs, L. Bogani, J. A. Mol, C. J. Lambert et al., “Quantum interference enhances the performance of single- molecule transistors,”Nature Nanotechnology, vol. 19, no. 7, pp. 986–992, 2024. 14

    2024

  70. [70]

    Density functional theory and the band gap problem,

    J. P. Perdew, “Density functional theory and the band gap problem,”International Journal of Quantum Chemistry, vol. 28, no. S19, pp. 497–523, 1985

    1985

  71. [71]

    Generalized gradient approximation made simple,

    J. P. Perdew, K. Burke, and M. Ernzerhof, “Generalized gradient approximation made simple,” Physical Review Letters, vol. 77, no. 18, pp. 3865–3868, 1996

    1996

  72. [72]

    Oligoyne molecular junctions for efficient room temperature thermoelectric power generation,

    H. Sadeghi, S. Sangtarash, and C. J. Lambert, “Oligoyne molecular junctions for efficient room temperature thermoelectric power generation,”Nano Letters, vol. 15, no. 11, pp. 7467–7472, 2015

    2015

  73. [73]

    Quantum and phonon interference-enhanced molecular-scale thermoelectricity,

    H. Sadeghi, “Quantum and phonon interference-enhanced molecular-scale thermoelectricity,” The Journal of Physical Chemistry C, vol. 123, no. 20, pp. 12 556–12 562, 2019

    2019

  74. [74]

    33 unresolved questions in nanoscience and nanotechnology,

    C. A. Mirkin, S. H. Petrosko, N. Artzi, K. Aydin, A. Biaggne et al., “33 unresolved questions in nanoscience and nanotechnology,”ACS Nano, vol. 19, no. 36, pp. 31 933–31 968, 2025

    2025

  75. [75]

    Extreme nanophotonics from ultrathin metallic gaps,

    J. J. Baumberg, J. Aizpurua, M. H. Mikkelsen, and D. R. Smith, “Extreme nanophotonics from ultrathin metallic gaps,”Nature Materials, vol. 18, no. 7, pp. 668–678, 2019

    2019

  76. [76]

    Le Ru and P

    E. Le Ru and P. Etchegoin,Principles of Surface-Enhanced Raman Spectroscopy: and related plasmonic effects. Elsevier, 2008

    2008

  77. [77]

    Molecular vibration explorer: an online database and toolbox for surface-enhanced frequency conversion and infrared and Raman spectroscopy,

    Z. Koczor-Benda, P. Roelli, C. Galland, and E. Rosta, “Molecular vibration explorer: an online database and toolbox for surface-enhanced frequency conversion and infrared and Raman spectroscopy,”The Journal of Physical Chemistry A, vol. 126, no. 28, pp. 4657–4663, 2022

    2022

  78. [78]

    A non-self-consistent tight-binding electronic structure potential in a polarized double-ζ basis set for all spd-block elements up to Z= 86,

    S. Grimme, M. M¨uller, and A. Hansen, “A non-self-consistent tight-binding electronic structure potential in a polarized double-ζ basis set for all spd-block elements up to Z= 86,”The Journal of Chemical Physics, vol. 158, no. 12, p. 124111, 2023

    2023

  79. [79]

    Density-functional exchange-energy approximation with correct asymptotic behavior,

    A. D. Becke, “Density-functional exchange-energy approximation with correct asymptotic behavior,”Physical Review A, vol. 38, no. 6, p. 3098, 1988

    1988

  80. [80]

    Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density,

    C. Lee, W. Yang, and R. G. Parr, “Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density,”Physical Review B, vol. 37, no. 2, p. 785, 1988

    1988

Showing first 80 references.