pith. machine review for the scientific record. sign in

arxiv: 2604.20372 · v1 · submitted 2026-04-22 · ⚛️ physics.flu-dyn · cs.AI· cs.LG· nlin.PS

Recognition: unknown

AI models of unstable flow exhibit hallucination

Birendra Jha, Ramdhan Wibawa

Pith reviewed 2026-05-09 23:54 UTC · model grok-4.3

classification ⚛️ physics.flu-dyn cs.AIcs.LGnlin.PS
keywords viscous fingeringAI hallucinationspectral biasfluid instabilitiesneural operatorsDeepFingersconservation lawsunstable flows
0
0 comments X

The pith

AI models of viscous fingering generate hallucinations as spurious fluid interfaces and reverse flows that violate conservation laws.

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

The paper shows that AI models trained on fluid flow data can produce outputs that look like realistic unstable patterns but break basic physical rules such as mass conservation. These errors show up as invented fluid boundaries or flows running backward, and they grow worse when the flow moves faster or the viscosity difference is larger. The authors trace the problem to the models' preference for smoother, lower-frequency patterns that cannot keep up with rapid finger splitting and merging. They respond by building DeepFingers, a hybrid neural operator that learns the full range of spatial scales by combining Fourier and deep operator networks and conditioning predictions on both time and viscosity contrast. The work matters because many engineering and scientific uses of AI for fluids depend on getting the fine details of instabilities right rather than just the overall appearance.

Core claim

We identify solutions that appear visually realistic yet are physically implausible. These hallucinations manifest as spurious fluid interfaces and reverse diffusion that violate conservation laws. Their origin lies in the spectral bias of AI models, which becomes dominant at high flow rates and viscosity contrasts. Guided by this insight, we introduce DeepFingers, a new framework that combines the Fourier Neural Operator with a Deep Operator Network to predict the spatiotemporal evolution of viscous fingers by conditioning on both time and viscosity contrast.

What carries the argument

Spectral bias of neural operators in multiscale fluid problems, countered by the DeepFingers hybrid architecture that balances learning across all spatial modes.

If this is right

  • DeepFingers captures tip splitting, finger merging, and channel formation across different regimes.
  • The framework preserves global mixing metrics while mapping successive concentration fields.
  • Conditioning on time and viscosity contrast enables learning that spans multiple flow conditions.
  • The results point to a research direction on fundamental limits of AI models for physical instabilities.

Where Pith is reading between the lines

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

  • The same spectral bias mechanism may produce analogous errors in AI models of other rapidly evolving multiscale systems such as turbulence or reactive mixing.
  • Adding explicit checks for conservation laws during training could reduce hallucinations more reliably than architecture changes alone.
  • If the bias is general, similar issues could appear in neural models that predict time sequences of any spatially extended physical field.
  • Direct comparison against laboratory images of viscous fingering would test whether the reported hallucinations are artifacts of simulation data or persist in real experiments.

Load-bearing premise

The visually implausible outputs are produced specifically by spectral bias rather than training data gaps or optimization problems, and the new hybrid model will not create fresh conservation violations outside the tested cases.

What would settle it

Compare mass conservation errors in concentration fields from standard AI models versus DeepFingers at increasing viscosity contrasts and flow rates; the claim holds if the hybrid model shows near-zero violations while others do not.

read the original abstract

We report the first systematic evidence of hallucination in AI models of fluid dynamics, demonstrated in the canonical problem of hydrodynamically unstable transport known as viscous fingering. AI-based modeling of flow with instabilities remains challenging because rapidly evolving, multiscale fingering patterns are difficult to resolve accurately. We identify solutions that appear visually realistic yet are physically implausible, analogous to hallucinations in large language models. These hallucinations manifest as spurious fluid interfaces and reverse diffusion that violate conservation laws. We show that their origin lies in the spectral bias of AI models, which becomes dominant at high flow rates and viscosity contrasts. Guided by this insight, we introduce DeepFingers, a new framework for AI-driven fluid dynamics that enforces balanced learning across the full spectrum of spatial modes by combining the Fourier Neural Operator with a Deep Operator Network to predict the spatiotemporal evolution of viscous fingers. By conditioning on both time and viscosity contrast, DeepFingers learns mappings between successive concentration fields across regimes. The framework accurately captures tip splitting, finger merging, and channel formation while preserving global metrics of mixing. The results open a new research direction to investigate fundamental limitations in AI models of physical systems.

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

3 major / 1 minor

Summary. The manuscript reports the first systematic evidence of hallucinations in AI models of fluid dynamics, using the canonical viscous fingering problem. It identifies visually realistic but physically implausible outputs (spurious fluid interfaces and reverse diffusion violating conservation laws) and attributes their origin to the spectral bias of neural operators, which becomes dominant at high flow rates and viscosity contrasts. Guided by this, the authors introduce DeepFingers, a hybrid framework combining the Fourier Neural Operator with a Deep Operator Network conditioned on time and viscosity contrast, which learns mappings between successive concentration fields and captures features such as tip splitting, finger merging, and channel formation while preserving global mixing metrics.

Significance. If the central claims hold with quantitative support, the work would be significant for AI modeling of unstable multiscale flows by diagnosing a key failure mode (spectral bias producing non-physical artifacts) and proposing a practical mitigation via the DeepFingers architecture. It could stimulate research into spectrum-balanced operator learning for physics applications. However, the absence of any reported metrics, error bars, conservation diagnostics, or ablation studies in the abstract renders the significance preliminary at present.

major comments (3)
  1. [Abstract] Abstract: The claim that hallucinations 'originate' specifically from spectral bias (rather than training-data under-sampling of complex unstable patterns or loss-function design) is not isolated by any ablation that holds the data distribution fixed while varying spectral content or model architecture. No such experiments are described.
  2. [Abstract] Abstract: No quantitative metrics, error bars, conservation-law checks, or ablation studies are provided to support either the spectral-bias diagnosis or the asserted superiority of DeepFingers in preserving mixing metrics across regimes.
  3. [Abstract] Abstract: The statement that DeepFingers 'enforces balanced learning across the full spectrum' by combining FNO and DeepONet lacks any description of the precise conditioning mechanism, loss weighting, or architectural details that would achieve spectral balance, making it impossible to assess whether new violations are introduced outside the tested regimes.
minor comments (1)
  1. [Abstract] The abstract refers to 'global metrics of mixing' without naming the specific quantities (e.g., mixing length, interfacial area) or reporting their values or comparisons.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their careful reading and constructive feedback. We address each major comment below and have revised the manuscript to incorporate the suggested improvements.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that hallucinations 'originate' specifically from spectral bias (rather than training-data under-sampling of complex unstable patterns or loss-function design) is not isolated by any ablation that holds the data distribution fixed while varying spectral content or model architecture. No such experiments are described.

    Authors: We agree that a controlled ablation holding the data distribution fixed while varying only spectral handling would provide stronger isolation of the cause. Our existing comparisons across architectures on identical datasets already show a clear correlation between spectral bias and hallucination rates, but to address this directly we have added a new ablation study in the revised manuscript (Section 4.3) that trains spectral variants on the same data. revision: yes

  2. Referee: [Abstract] Abstract: No quantitative metrics, error bars, conservation-law checks, or ablation studies are provided to support either the spectral-bias diagnosis or the asserted superiority of DeepFingers in preserving mixing metrics across regimes.

    Authors: The abstract is a concise summary and therefore omits specific numbers. The full manuscript reports quantitative metrics with error bars (from multiple random seeds), conservation-law diagnostics, and ablation studies comparing mixing metrics across regimes. We have revised the abstract to reference these quantitative elements and the demonstrated superiority of DeepFingers. revision: yes

  3. Referee: [Abstract] Abstract: The statement that DeepFingers 'enforces balanced learning across the full spectrum' by combining FNO and DeepONet lacks any description of the precise conditioning mechanism, loss weighting, or architectural details that would achieve spectral balance, making it impossible to assess whether new violations are introduced outside the tested regimes.

    Authors: We agree that the abstract does not contain implementation details. The full manuscript describes the time-and-viscosity-contrast conditioning via the DeepONet branch network and the frequency-balanced loss weighting. We have updated the abstract with a brief clause on the conditioning mechanism and expanded the methods section to allow evaluation of behavior outside tested regimes. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper reports empirical observations of hallucinations (spurious interfaces and reverse diffusion) in neural operator models of viscous fingering and attributes their origin to spectral bias dominating at high flow rates and viscosity contrasts. It then proposes DeepFingers as a combined FNO+DeepONet architecture conditioned on time and contrast. No equations, derivations, or load-bearing steps are shown that reduce the central claims to fitted parameters by construction, self-citations, or ansatzes imported from prior work. The attribution is presented as an experimental finding rather than a mathematical identity, and the framework is introduced as a practical mitigation without circular reduction to the inputs. The analysis is self-contained as an identification of model limitations supported by simulation comparisons.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only abstract available; no explicit free parameters, axioms, or invented entities are stated. Standard neural operator training assumptions (e.g., sufficient training data coverage across regimes) are implicitly required but not detailed.

pith-pipeline@v0.9.0 · 5506 in / 1065 out tokens · 49561 ms · 2026-05-09T23:54:43.668998+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

55 extracted references · 9 canonical work pages · 2 internal anchors

  1. [1]

    Losey, M.W., Schmidt, M.A., Jensen, K.F.: Microfabricated multiphase packed- bed reactors: Characterization of mass transfer and reactions. Ind. Eng. Chem. Res. 40, 2555 (2001)

    2001

  2. [2]

    Drug Discov

    Dunn, D.A., Feygin, I.: Challenges and solutions to ultra-high-throughput screen- ing assay miniaturization: submicroliter fluid handling. Drug Discov. Today 5, 84 (2000)

    2000

  3. [3]

    Baldyga, J., Czarnocki, R., Shekunov, B.Y., Smith, K.B.: Particle formation in supercritical fluids: scale-up problem. Chem. Eng. Res. Des. 88, 331 (2010)

    2010

  4. [4]

    Wiley, ??? (2009)

    Cullen, P.J.: Food Mixing: Principles and Applications. Wiley, ??? (2009)

    2009

  5. [5]

    Nature 344, 209 (1990)

    Olson, P., Silver, P.G., Carlson, R.W.: The large scale structure of convection in the earth’s mantle. Nature 344, 209 (1990)

    1990

  6. [6]

    Journal of Contaminant Hydrology 120, 1–17 (2011) 16

    Dentz, M., Le Borgne, T., Englert, A., Bijeljic, B.: Mixing, spreading and reaction in heterogeneous media: A brief review. Journal of Contaminant Hydrology 120, 1–17 (2011) 16

    2011

  7. [7]

    Annual Review of Fluid Mechanics 19(1), 271–311 (1987)

    Homsy, G.M.: Viscous fingering in porous media. Annual Review of Fluid Mechanics 19(1), 271–311 (1987)

    1987

  8. [8]

    SPE advanced technology series 1(01), 71–80 (1993)

    Araktingi, U.G., Orr, F.: Viscous fingering in heterogeneous porous media. SPE advanced technology series 1(01), 71–80 (1993)

    1993

  9. [9]

    Physical review letters 106(19), 194502 (2011)

    Jha, B., Cueto-Felgueroso, L., Juanes, R.: Fluid mixing from viscous fingering. Physical review letters 106(19), 194502 (2011)

    2011

  10. [10]

    Physical review letters 115(17), 174501 (2015)

    Zheng, Z., Kim, H., Stone, H.A.: Controlling viscous fingering using time- dependent strategies. Physical review letters 115(17), 174501 (2015)

    2015

  11. [11]

    Heliyon 7(7) (2021)

    Pinilla, A., Asuaje, M., Ratkovich, N.: Experimental and computational advances on the study of viscous fingering: An umbrella review. Heliyon 7(7) (2021)

    2021

  12. [12]

    Rheologica Acta 57(8), 575–589 (2018)

    Yazdi, A.A., Norouzi, M.: Numerical study of saffman–taylor instability in immiscible nonlinear viscoelastic flows. Rheologica Acta 57(8), 575–589 (2018)

    2018

  13. [13]

    Physical Review E 92(4), 041003 (2015)

    Chui, J.Y., Anna, P., Juanes, R.: Interface evolution during radial miscible viscous fingering. Physical Review E 92(4), 041003 (2015)

    2015

  14. [14]

    Journal of Engineering Mathematics 90(1), 83–104 (2015)

    Moyles, I., Wetton, B.: Fingering phenomena in immiscible displacement in porous media flow. Journal of Engineering Mathematics 90(1), 83–104 (2015)

    2015

  15. [15]

    Bacri, J.C., Rakotomalala, N., Salin, D., Woumeni, R.: Miscible viscous fingering: Experiments versus continuum approach. Phys. Fluids 4, 1611 (1992)

    1992

  16. [16]

    Petitjeans, P., Maxworthy, T.: Miscible displacements in capillary tubes. part 1. experiments. Journal of Fluid Mechanics 326, 37–56 (1996)

    1996

  17. [17]

    Geoenergy Science and Engineering 250, 213834 (2025)

    Ren, J., Xiao, W., Cheng, Q., Song, P., Bai, X., Xie, Q., Pu, W., Zheng, L.: Experimental study on water/co2 flow of tight oil using hthp microscopic visu- alization and nmr technology. Geoenergy Science and Engineering 250, 213834 (2025)

    2025

  18. [18]

    Journal of Fluid Mechanics 141, 275–287 (1984)

    Park, C.-W., Gorell, S., Homsy, G.: Two-phase displacement in hele-shaw cells: experiments on viscously driven instabilities. Journal of Fluid Mechanics 141, 275–287 (1984)

    1984

  19. [19]

    Nature communications 10(1), 4002 (2019)

    Gao, T., Mirzadeh, M., Bai, P., Conforti, K.M., Bazant, M.Z.: Active control of viscous fingering using electric fields. Nature communications 10(1), 4002 (2019)

    2019

  20. [20]

    Scientific reports 11(1), 24368 (2021)

    Escala, D.M., Muñuzuri, A.P.: A bottom-up approach to construct or deconstruct a fluid instability. Scientific reports 11(1), 24368 (2021)

    2021

  21. [21]

    Advances in Water Resources 136, 103485 (2020) 17

    Tran, M., Jha, B.: Coupling between transport and geomechanics affects spread- ing and mixing during viscous fingering in deformable aquifers. Advances in Water Resources 136, 103485 (2020) 17

    2020

  22. [22]

    Frontiers in Bioengineering and Biotechnology 10, 877480 (2022)

    Tu, T.-Y., Shen, Y.-P., Lim, S.-H., Wang, Y.-K.: A facile method for generat- ing a smooth and tubular vessel lumen using a viscous fingering pattern in a microfluidic device. Frontiers in Bioengineering and Biotechnology 10, 877480 (2022)

    2022

  23. [23]

    Water Resources Research 51(4), 2634–2647 (2015)

    Nicolaides, C., Jha, B., Cueto-Felgueroso, L., Juanes, R.: Impact of viscous fin- gering and permeability heterogeneity on fluid mixing in porous media. Water Resources Research 51(4), 2634–2647 (2015)

    2015

  24. [24]

    Environmental science & technology 47(1), 212–218 (2013)

    Wang, Y., Zhang, C., Wei, N., Oostrom, M., Wietsma, T.W., Li, X., Bonneville, A.: Experimental study of crossover from capillary to viscous fingering for super- critical co2–water displacement in a homogeneous pore network. Environmental science & technology 47(1), 212–218 (2013)

    2013

  25. [25]

    Scientific reports 9(1), 11272 (2019)

    Gooya, R., Silvestri, A., Moaddel, A., Andersson, M., Stipp, S., Sørensen, H.: Unstable, super critical co2–water displacement in fine grained porous media under geologic carbon sequestration conditions. Scientific reports 9(1), 11272 (2019)

    2019

  26. [26]

    International Journal of Greenhouse Gas Control 11, 188–203 (2012)

    Berg, S., Ott, H.: Stability of co2–brine immiscible displacement. International Journal of Greenhouse Gas Control 11, 188–203 (2012)

    2012

  27. [27]

    Physical Review Letters 111(14), 144501 (2013)

    Jha, B., Cueto-Felgueroso, L., Juanes, R.: Synergetic fluid mixing from viscous fingering and alternating injection. Physical Review Letters 111(14), 144501 (2013)

    2013

  28. [28]

    Journal of Colloid and Interface Science, 138202 (2025)

    Zhao, W., Wang, S., Zhou, Y., Wang, M., Yang, L., Liu, J., Wang, L., Zhu, P.: Viscous fingering instability in one-end-lifted hele-shaw cells for producing three- dimensional hierarchical structures. Journal of Colloid and Interface Science, 138202 (2025)

    2025

  29. [29]

    The Physics of fluids 31(6), 1330–1338 (1988)

    Tan, C., Homsy, G.: Simulation of nonlinear viscous fingering in miscible displacement. The Physics of fluids 31(6), 1330–1338 (1988)

    1988

  30. [30]

    Chaos: An Interdisciplinary Journal of Nonlinear Science 14(3), 809–824 (2004)

    Casademunt, J.: Viscous fingering as a paradigm of interfacial pattern forma- tion: Recent results and new challenges. Chaos: An Interdisciplinary Journal of Nonlinear Science 14(3), 809–824 (2004)

    2004

  31. [31]

    In: NIPS 2018 Workshop on Interpretability and Robustness in Audio, Speech, and Language (IRASL) (2018)

    Lee, K., Ippolito, D., Nystrom, A., Zhang, C., Eck, D., Callison-Burch, C., Car- lini, N.: Hallucinations in neural machine translation. In: NIPS 2018 Workshop on Interpretability and Robustness in Audio, Speech, and Language (IRASL) (2018). Neural Information Processing Systems

    2018

  32. [32]

    DeepONet: Learning nonlinear operators for iden- tifying differential equations based on the universal approximation theorem of operators,

    Lu, L., Jin, P., Karniadakis, G.E.: Deeponet: Learning nonlinear operators for identifying differential equations based on the universal approximation theorem of operators. arXiv preprint arXiv:1910.03193 (2019) 18

    work page arXiv 1910

  33. [33]

    Fourier Neural Operator for Parametric Partial Differential Equations

    Li, Z., Kovachki, N., Azizzadenesheli, K., Liu, B., Bhattacharya, K., Stuart, A., Anandkumar, A.: Fourier neural operator for parametric partial differential equations. arXiv preprint arXiv:2010.08895 (2020)

    work page internal anchor Pith review arXiv 2010

  34. [34]

    Journal of Machine Learning Research 24(388), 1–26 (2023)

    Li, Z., Huang, D.Z., Liu, B., Anandkumar, A.: Fourier neural operator with learned deformations for pdes on general geometries. Journal of Machine Learning Research 24(388), 1–26 (2023)

    2023

  35. [35]

    Advances in Neural Information Processing Systems 36, 77187– 77200 (2023)

    Raonic, B., Molinaro, R., De Ryck, T., Rohner, T., Bartolucci, F., Alaifari, R., Mishra, S., Bézenac, E.: Convolutional neural operators for robust and accurate learning of pdes. Advances in Neural Information Processing Systems 36, 77187– 77200 (2023)

    2023

  36. [36]

    Enhancing fourier neural opera- tors with local spatial features

    Liu, C., Murari, D., Liu, L., Li, Y., Budd, C., Schönlieb, C.-B.: Enhancing fourier neural operators with local spatial features. arXiv preprint arXiv:2503.17797 (2025)

    work page arXiv 2025

  37. [37]

    In: When Machine Learning Meets Dynamical Systems: Theory and Applications (2023)

    Sousa Almeida, J.L., Rocha, P.R.B., Carvalho, A.M., Nogueira Jr, A.C.: A coupled variational encoder-decoder-deeponet surrogate model for the rayleigh- bénard convection problem. In: When Machine Learning Meets Dynamical Systems: Theory and Applications (2023)

    2023

  38. [38]

    arXiv preprint arXiv:2310.02491 (2023)

    Michałowska, K., Goswami, S., Karniadakis, G.E., Riemer-Sørensen, S.: Don- lstm: Multi-resolution learning with deeponets and long short-term memory neural networks. arXiv preprint arXiv:2310.02491 (2023)

    work page arXiv 2023

  39. [39]

    Computer Methods in Applied Mechanics and Engineering 420, 116681 (2024)

    Haghighat, E., Waheed, U., Karniadakis, G.: En-deeponet: An enrichment approach for enhancing the expressivity of neural operators with applications to seismology. Computer Methods in Applied Mechanics and Engineering 420, 116681 (2024)

    2024

  40. [40]

    Algorithms 15(9), 325 (2022)

    Moya, C., Lin, G.: Fed-deeponet: Stochastic gradient-based federated training of deep operator networks. Algorithms 15(9), 325 (2022)

    2022

  41. [41]

    Nature Communications 15(1), 5101 (2024)

    Kontolati, K., Goswami, S., Em Karniadakis, G., Shields, M.D.: Learning non- linear operators in latent spaces for real-time predictions of complex dynamics in physical systems. Nature Communications 15(1), 5101 (2024)

    2024

  42. [42]

    arXiv preprint arXiv:2206.05655 (2022)

    Garg, S., Chakraborty, S.: Variational bayes deep operator network: a data- driven bayesian solver for parametric differential equations. arXiv preprint arXiv:2206.05655 (2022)

    work page arXiv 2022

  43. [43]

    Advances in Water Resources 163, 104180 (2022) 19

    Wen, G., Li, Z., Azizzadenesheli, K., Anandkumar, A., Benson, S.M.: U-fno—an enhanced fourier neural operator-based deep-learning model for multiphase flow. Advances in Water Resources 163, 104180 (2022) 19

    2022

  44. [44]

    Zhu, M., Feng, S., Lin, Y., Lu, L.: Fourier-deeponet: Fourier-enhanced deep opera- tor networks for full waveform inversion with improved accuracy, generalizability, and robustness (2023) https://doi.org/10.2139/ssrn.4461079

    work page doi:10.2139/ssrn.4461079 2023

  45. [45]

    In: International Conference on Medical Image Computing and Computer-assisted Intervention, pp

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedical image segmentation. In: International Conference on Medical Image Computing and Computer-assisted Intervention, pp. 234–241 (2015). Springer

    2015

  46. [46]

    Neural networks 146, 272–289 (2022)

    Geneva, N.D., Zabaras, N.: Transformers for modeling physical systems. Neural networks 146, 272–289 (2022)

    2022

  47. [47]

    Nature Communications 15(1), 1361 (2024)

    Solera-Rico, A., Sanmiguel Vila, C., Gómez-López, M., Wang, Y., Almashjary, A., Dawson, S.T., Vinuesa, R.: β-variational autoencoders and transformers for reduced-order modelling of fluid flows. Nature Communications 15(1), 1361 (2024)

    2024

  48. [48]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

    work page internal anchor Pith review Pith/arXiv arXiv 2010

  49. [49]

    H., Sankaran, S., Wang, H., Pappas, G

    Wang, S., Seidman, J.H., Sankaran, S., Wang, H., Pappas, G.J., Perdikaris, P.: Cvit: Continuous vision transformer for operator learning. arXiv preprint arXiv:2405.13998 (2024)

    work page arXiv 2024

  50. [50]

    Physics of Fluids 37(11) (2025)

    Alasker, M., Liu, Z., Ghazvini, F., Barros, F., Jha, B.: Scalar transport during miscible viscous fingering experiments in consolidated porous media. Physics of Fluids 37(11) (2025)

    2025

  51. [51]

    In: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), pp

    Maynez, J., Narayan, S., Bohnet, B., McDonald, R.: On faithfulness and factu- ality in abstractive summarization. In: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), pp. 1906–1919 (2020). Association for Computational Linguistics. https://aclanthology.org/2020.acl- main.173/

    1906

  52. [52]

    In: 2024 IEEE Conference on Artificial Intelligence (CAI), pp

    Maleki, N., Padmanabhan, B., Dutta, K.: Ai hallucinations: a misnomer worth clarifying. In: 2024 IEEE Conference on Artificial Intelligence (CAI), pp. 133–138 (2024). IEEE

    2024

  53. [53]

    Dillion, D., Tandon, N., Gu, Y., Gray, K.: Can ai language models replace human participants? Trends in Cognitive Sciences 27(7), 597–600 (2023)

    2023

  54. [54]

    Computers in Human Behavior: Artificial Humans 1(2), 100006 (2023)

    Salah, M., Al Halbusi, H., Abdelfattah, F.: May the force of text data analysis be with you: Unleashing the power of generative ai for social psychology research. Computers in Human Behavior: Artificial Humans 1(2), 100006 (2023)

    2023

  55. [55]

    arXiv preprint arXiv:2309.02791 (2023) 21

    Sheng, H., Wu, X., Si, X., Li, J., Zhang, S., Duan, X.: Seismic Foundation Model 20 (SFM): a new generation deep learning model in geophysics. arXiv preprint arXiv:2309.02791 (2023) 21

    work page arXiv 2023