{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T06:04:35Z","timestamp":1790229875679,"version":"4.1.0"},"reference-count":43,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T00:00:00Z","timestamp":1790208000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>Art is increasingly recognized as an alternative asset class in wealth management. However, its non-standardized nature renders it difficult to treat on a par with conventional financial assets. Valuation risk arises from subjective pricing and expert bias, while liquidity risk is manifested in high auction failure rates. These challenges are difficult to address adequately with traditional linear econometric models. This study proposes a multimodal deep learning framework that integrates painting images, historical transaction data, artist reputation, and macroeconomic indicators. A ResNet-50 network, initialized on ImageNet and subsequently adapted to the artwork domain, serves as the visual encoder, while a multi-layer perceptron processes structured tabular variables; the two representations are combined in a fusion layer. The framework simultaneously performs two risk-assessment tasks: valuation deviation regression and unsold-lot binary classification. The model is estimated on 96,314 modern and contemporary auction lots offered between 2000 and 2023, of which 27.8% went unsold, using a strictly chronological training, validation, and test partition. The empirical results demonstrate that the proposed multimodal framework outperforms traditional benchmarks, including ordinary least squares (OLS), logistic regression, and random forests. The model achieves an Area Under the Curve (AUC) of 0.891 for liquidity risk prediction. It reduces the Root Mean Squared Error (RMSE) for valuation deviation by 32.9% compared to OLS. Ablation studies confirm that visual embeddings provide substantial incremental information beyond what can be attributed to increased model complexity alone, and permutation importance combined with linear probing shows that the visual representation encodes stylistic period, subject type, chromatic intensity, and compositional complexity\u2014the properties through which visual content is theorized to affect market outcomes. Robustness checks further delineate the model's performance boundaries under macroeconomic shocks. This study translates the aesthetic heterogeneity of non-standard assets into quantifiable risk signals. It offers financial institutions a data-calibrated basis for adjusting dynamic loan-to-value ratios and assists art funds in optimizing portfolio allocation.<\/jats:p>","DOI":"10.3389\/frai.2026.1946060","type":"journal-article","created":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T05:44:07Z","timestamp":1790228647000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Research on liquidity and valuation risk assessment of art financial assets based on deep learning"],"prefix":"10.3389","volume":"9","author":[{"given":"Ziwen","family":"Jiang","sequence":"first","affiliation":[{"name":"XuBeihong Art Academy, Shanghai Maritime University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jingzhou","family":"Zhao","sequence":"additional","affiliation":[{"name":"XuBeihong Art Academy, Shanghai Maritime University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiamin","family":"Yao","sequence":"additional","affiliation":[{"name":"College of Information Engineering, Shanghai Maritime University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zehao","family":"Bian","sequence":"additional","affiliation":[{"name":"XuBeihong Art Academy, Shanghai Maritime University","place":["Shanghai, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,9,24]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","DOI":"10.1145\/3292500.3330701","article-title":"\u201cOptuna: a next-generation hyperparameter optimization framework,\u201d","author":"Akiba","year":"2019","journal-title":"Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery and data mining"},{"key":"B2","doi-asserted-by":"publisher","first-page":"1714822","DOI":"10.3389\/frai.2026.1714822","article-title":"Modeling saudi stock index returns and volatility: a dual approach using garch and neural networks","volume":"9","author":"Al-Besher","year":"2026","journal-title":"Front. Artif. Intell"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1702414","DOI":"10.3389\/frai.2026.1702414","article-title":"Modeling and forecasting saudi banking stability using arima and exponential smoothing technique","volume":"9","author":"Alnajjar","year":"2026","journal-title":"Front. Artif. Intell"},{"key":"B4","doi-asserted-by":"publisher","first-page":"795","DOI":"10.1111\/jofi.13203","article-title":"Biased auctioneers","volume":"78","author":"Aubry","year":"2023","journal-title":"J. Finance"},{"key":"B5","doi-asserted-by":"publisher","first-page":"423","DOI":"10.1109\/TPAMI.2018.2798607","article-title":"Multimodal machine learning: a survey and taxonomy","volume":"41","author":"Baltru\u0161aitis","year":"2019","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell"},{"key":"B6","first-page":"1015","article-title":"Anchoring and buying-in at art auctions","volume":"26","author":"Beggs","year":"2008","journal-title":"Int. J. Indus. Organiz"},{"key":"B7","first-page":"142","article-title":"Does global fear predict the art market?","volume":"162","author":"Bouri","year":"2018","journal-title":"Econ. Lett"},{"key":"B8","doi-asserted-by":"publisher","first-page":"31","DOI":"10.3905\/jai.2008.705533","article-title":"Art as an alternative asset class","volume":"10","author":"Campbell","year":"2008","journal-title":"J. Altern. Invest"},{"key":"B9","doi-asserted-by":"publisher","first-page":"107","DOI":"10.1016\/j.eswa.2018.07.026","article-title":"Fine-tuning convolutional neural networks for fine art classification","volume":"114","author":"Cetinic","year":"2018","journal-title":"Expert Syst. Appl"},{"key":"B10","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/BF00148268","article-title":"The relevance of hedonic price indices: the case of paintings","volume":"20","author":"Chanel","year":"1996","journal-title":"J. Cult. Econ"},{"key":"B11","doi-asserted-by":"publisher","first-page":"1774013","DOI":"10.3389\/frai.2026.1774013","article-title":"Real-time dynamic graph learning with temporal attention for financial fraud detection","volume":"9","author":"Chen","year":"2026","journal-title":"Front. Artif. Intell"},{"key":"B12","first-page":"143","article-title":"Look at that painting! A hedonic approach to the art market","volume":"33","author":"Collins","year":"2009","journal-title":"J. Cult. Econ"},{"key":"B13","first-page":"187","article-title":"Econometric fine art valuation by combining hedonic and repeat-sales information","volume":"42","author":"Galbraith","year":"2018","journal-title":"J. Cult. Econ"},{"key":"B14","doi-asserted-by":"publisher","first-page":"222","DOI":"10.1257\/aer.101.3.222","article-title":"Art and money","volume":"101","author":"Goetzmann","year":"2011","journal-title":"Am. Econ. Rev"},{"key":"B15","volume-title":"Deep Learning","author":"Goodfellow","year":"2016"},{"key":"B16","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90","article-title":"\u201cDeep residual learning for image recognition,\u201d","author":"He","year":"2016","journal-title":"Proceedings of the IEEE conference on computer vision and pattern recognition"},{"key":"B17","doi-asserted-by":"publisher","first-page":"203","DOI":"10.1109\/ICORIS60118.2023.10352208","article-title":"\u201cMultimodal approach for painting price prediction,\u201d","author":"Indrawan","year":"2023","journal-title":"2023 5th International conference on cybernetics and intelligent system (ICORIS)"},{"key":"B18","doi-asserted-by":"publisher","first-page":"1752103","DOI":"10.3389\/frai.2026.1752103","article-title":"Fintextsim: a domain-specific sentence-transformer for extracting predictive latent topics from financial disclosures","volume":"9","author":"Jehnen","year":"2026","journal-title":"Front. Artif. Intell"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1007","DOI":"10.1093\/rfs\/hhv062","article-title":"Does it pay to invest in art? A selection-corrected returns perspective","volume":"29","author":"Korteweg","year":"2016","journal-title":"Rev. Fin. Stud"},{"key":"B20","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.ins.2014.02.137","article-title":"Evaluation of clustering algorithms for financial risk analysis","volume":"62","author":"Kou","year":"2014","journal-title":"Decis. Support Syst"},{"key":"B21","doi-asserted-by":"publisher","first-page":"38","DOI":"10.1016\/j.dss.2017.10.001","article-title":"Decision support from financial disclosures with deep neural networks and transfer learning","volume":"104","author":"Kraus","year":"2017","journal-title":"Decis. Support Syst"},{"key":"B22","doi-asserted-by":"publisher","first-page":"689","DOI":"10.1016\/j.ejor.2016.10.031","article-title":"Deep neural networks, gradient-boosted trees, random forests: statistical arbitrage on the s&p 500","volume":"259","author":"Krauss","year":"2017","journal-title":"Eur. J. Oper. Res"},{"key":"B23","doi-asserted-by":"publisher","first-page":"436","DOI":"10.1038\/nature14539","article-title":"Deep learning","volume":"521","author":"LeCun","year":"2015","journal-title":"Nature"},{"key":"B24","doi-asserted-by":"publisher","first-page":"124","DOI":"10.1016\/j.ejor.2015.05.030","article-title":"Benchmarking state-of-the-art classification algorithms for credit scoring: an update of research","volume":"247","author":"Lessmann","year":"2015","journal-title":"Eur. J. Oper. Res"},{"key":"B25","article-title":"\u201cDecoupled weight decay regularization,\u201d","volume-title":"Proceedings of the 7th international conference on learning representations (ICLR)","author":"Loshchilov","year":"2019"},{"key":"B26","doi-asserted-by":"publisher","first-page":"744","DOI":"10.1257\/aer.20160522","article-title":"A model of trading in the art market","volume":"108","author":"Lovo","year":"2018","journal-title":"Am. Econ. Rev"},{"key":"B27","first-page":"4765","article-title":"\u201cA unified approach to interpreting model predictions,\u201d","volume-title":"Advances in neural information processing systems 30 (NIPS)","author":"Lundberg","year":"2017"},{"key":"B28","doi-asserted-by":"publisher","first-page":"104004","DOI":"10.1016\/j.euroecorev.2021.104004","article-title":"Colors, emotions, and the auction value of paintings","volume":"142","author":"Ma","year":"2022","journal-title":"Eur. Econ. Rev"},{"key":"B29","doi-asserted-by":"publisher","first-page":"11006","DOI":"10.3390\/app142311006","article-title":"Tabular data models for predicting art auction results","volume":"14","author":"Mauer","year":"2024","journal-title":"Appl. Sci"},{"key":"B30","author":"McAndrew","year":"","journal-title":"The Art Basel and UBS Global Art Market Report 2025"},{"key":"B31","author":"McAndrew","year":"","journal-title":"The Art Basel and UBS Global Art Market Report 2025: A Survey of Global Collecting"},{"key":"B32","doi-asserted-by":"publisher","first-page":"1656","DOI":"10.1257\/000282802762024719","article-title":"Art as an investment and the underperformance of masterpieces","volume":"92","author":"Mei","year":"2002","journal-title":"Am. Econ. Rev"},{"key":"B33","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2512.23078","article-title":"Deep learning for art market valuation","author":"Mei","year":"2025","journal-title":"arXiv Preprint"},{"key":"B34","doi-asserted-by":"publisher","first-page":"2066","DOI":"10.3390\/ai5040101","article-title":"Deep learning in finance: a survey of applications and techniques","volume":"5","author":"Mienye","year":"2024","journal-title":"AI"},{"key":"B35","first-page":"506","article-title":"Art as an asset class: Implications for investors","volume":"41","author":"Oosterlinck","year":"2017","journal-title":"Res. Int. Business Fin"},{"key":"B36","doi-asserted-by":"publisher","first-page":"106384","DOI":"10.1016\/j.asoc.2020.106384","article-title":"Deep learning for financial applications: a survey","volume":"93","author":"Ozbayoglu","year":"2020","journal-title":"Appl. Soft Comput"},{"key":"B37","doi-asserted-by":"publisher","first-page":"432","DOI":"10.1016\/j.econlet.2014.01.008","article-title":"Sentiment and art prices","volume":"122","author":"Penasse","year":"2014","journal-title":"Econ. Lett"},{"key":"B38","doi-asserted-by":"publisher","first-page":"1819994","DOI":"10.3389\/frai.2026.1819994","article-title":"Transformer-based nlp approaches for credit risk prediction: a systematic review","volume":"9","author":"Raliphada","year":"2026","journal-title":"Front. Artif. Intell"},{"key":"B39","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1287\/mnsc.1120.1580","article-title":"Buying beauty: on prices and returns in the art market","volume":"59","author":"Renneboog","year":"2013","journal-title":"Manage. Sci"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.74","article-title":"\u201cGrad-cam: visual explanations from deep networks via gradient-based localization,\u201d","author":"Selvaraju","year":"2017","journal-title":"Proceedings of the IEEE international conference on computer vision (ICCV)"},{"key":"B41","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1409.1556","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2015","journal-title":"arXiv Preprint"},{"key":"B42","doi-asserted-by":"publisher","first-page":"1443372","DOI":"10.1080\/23322039.2018.1443372","article-title":"The comparison of the hedonic, repeat sales, and hybrid models: evidence from the chinese paintings market","volume":"6","author":"Wang","year":"2018","journal-title":"Cogent Econ. Fin"},{"key":"B43","doi-asserted-by":"publisher","first-page":"6978","DOI":"10.1109\/TASE.2023.3335145","article-title":"A spatial-temporal gated network for credit card fraud detection by learning transactional representations","volume":"21","author":"Xie","year":"2023","journal-title":"IEEE Trans. Autom. Sci. Eng."}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1946060\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,24]],"date-time":"2026-09-24T05:44:09Z","timestamp":1790228649000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1946060\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,24]]},"references-count":43,"alternative-id":["10.3389\/frai.2026.1946060"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1946060","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,24]]},"article-number":"1946060"}}