[{"status":"public","abstract":[{"lang":"eng","text":"While shallow decision trees may be interpretable, larger ensemble models like gradient-boosted trees, which often set the state of the art in machine learning problems involving tabular data, still remain black box models. As a remedy, the Shapley value (SV) is a well-known concept in explainable artificial intelligence (XAI) research for quantifying additive feature attributions of predictions. The model-specific TreeSHAP methodology solves the exponential complexity for retrieving exact SVs from tree-based models. Expanding beyond individual feature attribution, Shapley interactions reveal the impact of intricate feature interactions of any order. In this work, we present TreeSHAP-IQ, an efficient method to compute any-order additive Shapley interactions for predictions of tree-based models. TreeSHAP-IQ is supported by a mathematical framework that exploits polynomial arithmetic to compute the interaction scores in a single recursive traversal of the tree, akin to Linear TreeSHAP. We apply TreeSHAP-IQ on state-of-the-art tree ensembles and explore interactions on well-established benchmark datasets."}],"type":"conference","publication":"Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)","language":[{"iso":"eng"}],"keyword":["Explainable Artificial Intelligence"],"user_id":"93420","department":[{"_id":"660"}],"project":[{"name":"TRR 318 - C3: TRR 318 - Subproject C3","_id":"126"},{"_id":"109","name":"TRR 318: TRR 318 - Erklärbarkeit konstruieren"},{"name":"TRR 318 - C: TRR 318 - Project Area C","_id":"117"}],"_id":"53073","citation":{"mla":"Muschalik, Maximilian, et al. “Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles.” <i>Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)</i>, vol. 38, no. 13, 2024, pp. 14388–96, doi:<a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>.","short":"M. Muschalik, F. Fumagalli, B. Hammer, E. Huellermeier, in: Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2024, pp. 14388–14396.","bibtex":"@inproceedings{Muschalik_Fumagalli_Hammer_Huellermeier_2024, title={Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles}, volume={38}, DOI={<a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>}, number={13}, booktitle={Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)}, author={Muschalik, Maximilian and Fumagalli, Fabian and Hammer, Barbara and Huellermeier, Eyke}, year={2024}, pages={14388–14396} }","apa":"Muschalik, M., Fumagalli, F., Hammer, B., &#38; Huellermeier, E. (2024). Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles. <i>Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)</i>, <i>38</i>(13), 14388–14396. <a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">https://doi.org/10.1609/aaai.v38i13.29352</a>","ama":"Muschalik M, Fumagalli F, Hammer B, Huellermeier E. Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles. In: <i>Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)</i>. Vol 38. ; 2024:14388-14396. doi:<a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>","ieee":"M. Muschalik, F. Fumagalli, B. Hammer, and E. Huellermeier, “Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles,” in <i>Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)</i>, 2024, vol. 38, no. 13, pp. 14388–14396, doi: <a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">10.1609/aaai.v38i13.29352</a>.","chicago":"Muschalik, Maximilian, Fabian Fumagalli, Barbara Hammer, and Eyke Huellermeier. “Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles.” In <i>Proceedings of the AAAI Conference on Artificial Intelligence (AAAI)</i>, 38:14388–96, 2024. <a href=\"https://doi.org/10.1609/aaai.v38i13.29352\">https://doi.org/10.1609/aaai.v38i13.29352</a>."},"intvolume":"        38","page":"14388-14396","year":"2024","issue":"13","publication_status":"published","publication_identifier":{"issn":["2374-3468","2159-5399"]},"doi":"10.1609/aaai.v38i13.29352","title":"Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensembles","date_created":"2024-03-27T14:50:04Z","author":[{"last_name":"Muschalik","full_name":"Muschalik, Maximilian","first_name":"Maximilian"},{"last_name":"Fumagalli","id":"93420","full_name":"Fumagalli, Fabian","first_name":"Fabian"},{"first_name":"Barbara","last_name":"Hammer","full_name":"Hammer, Barbara"},{"full_name":"Huellermeier, Eyke","id":"48129","last_name":"Huellermeier","first_name":"Eyke"}],"volume":38,"date_updated":"2025-09-11T16:20:11Z"},{"status":"public","abstract":[{"text":"Explainability for machine learning gets more and more important in high-stakes decisions like real estate appraisal. While traditional hedonic house pricing models are fed with hard information based on housing attributes, recently also soft information has been incorporated to increase the predictive performance. This soft information can be extracted from image data by complex models like Convolutional Neural Networks (CNNs). However, these are intransparent which excludes their use for high-stakes financial decisions. To overcome this limitation, we examine if a two-stage modeling approach can provide explainability. We combine visual interpretability by Regression Activation Maps (RAM) for the CNN and a linear regression for the overall prediction. Our experiments are based on 62.000 family homes in Philadelphia and the results indicate that the CNN learns aspects related to vegetation and quality aspects of the house from exterior images, improving the predictive accuracy of real estate appraisal by up to 5.4%.","lang":"eng"}],"type":"conference","publication":"55th Annual Hawaii International Conference on System Sciences (HICSS-55)","language":[{"iso":"eng"}],"keyword":["Explainable Artificial Intelligence (XAI)","Regression Activation Maps","Real Estate Appraisal","Convolutional Block Attention Module","Computer Vision"],"user_id":"77066","department":[{"_id":"195"},{"_id":"196"}],"_id":"27506","citation":{"bibtex":"@inproceedings{Kucklick_2022, title={Visual Interpretability of Image-based Real Estate Appraisal}, booktitle={55th Annual Hawaii International Conference on System Sciences (HICSS-55)}, author={Kucklick, Jan-Peter}, year={2022} }","mla":"Kucklick, Jan-Peter. “Visual Interpretability of Image-Based Real Estate Appraisal.” <i>55th Annual Hawaii International Conference on System Sciences (HICSS-55)</i>, 2022.","short":"J.-P. Kucklick, in: 55th Annual Hawaii International Conference on System Sciences (HICSS-55), 2022.","apa":"Kucklick, J.-P. (2022). Visual Interpretability of Image-based Real Estate Appraisal. <i>55th Annual Hawaii International Conference on System Sciences (HICSS-55)</i>. Hawaii International Conference on System Science (HICSS), Virtual.","ama":"Kucklick J-P. Visual Interpretability of Image-based Real Estate Appraisal. In: <i>55th Annual Hawaii International Conference on System Sciences (HICSS-55)</i>. ; 2022.","ieee":"J.-P. Kucklick, “Visual Interpretability of Image-based Real Estate Appraisal,” presented at the Hawaii International Conference on System Science (HICSS), Virtual, 2022.","chicago":"Kucklick, Jan-Peter. “Visual Interpretability of Image-Based Real Estate Appraisal.” In <i>55th Annual Hawaii International Conference on System Sciences (HICSS-55)</i>, 2022."},"year":"2022","main_file_link":[{"open_access":"1","url":"https://scholarspace.manoa.hawaii.edu/bitstream/10125/79519/0149.pdf"}],"conference":{"location":"Virtual","end_date":"2022-01-07","start_date":"2022-01-03","name":"Hawaii International Conference on System Science (HICSS)"},"title":"Visual Interpretability of Image-based Real Estate Appraisal","date_created":"2021-11-17T07:08:15Z","author":[{"first_name":"Jan-Peter","last_name":"Kucklick","full_name":"Kucklick, Jan-Peter","id":"77066"}],"date_updated":"2022-01-06T06:57:40Z","oa":"1"},{"date_updated":"2022-01-26T08:24:30Z","oa":"1","date_created":"2022-01-26T08:22:03Z","author":[{"id":"77066","full_name":"Kucklick, Jan-Peter","last_name":"Kucklick","first_name":"Jan-Peter"}],"title":"Towards a model- and data-focused taxonomy of XAI systems","main_file_link":[{"open_access":"1","url":"https://aisel.aisnet.org/cgi/viewcontent.cgi?article=1056&context=wi2022"}],"conference":{"start_date":"2022-02-21","name":"Wirtschaftsinformatik 2022 (WI22)","location":"Nürnberg (online)","end_date":"2022-02-23"},"year":"2022","citation":{"apa":"Kucklick, J.-P. (2022). Towards a model- and data-focused taxonomy of XAI systems. <i>Wirtschaftsinformatik 2022 Proceedings</i>. Wirtschaftsinformatik 2022 (WI22), Nürnberg (online).","bibtex":"@inproceedings{Kucklick_2022, title={Towards a model- and data-focused taxonomy of XAI systems}, booktitle={Wirtschaftsinformatik 2022 Proceedings}, author={Kucklick, Jan-Peter}, year={2022} }","mla":"Kucklick, Jan-Peter. “Towards a Model- and Data-Focused Taxonomy of XAI Systems.” <i>Wirtschaftsinformatik 2022 Proceedings</i>, 2022.","short":"J.-P. Kucklick, in: Wirtschaftsinformatik 2022 Proceedings, 2022.","ama":"Kucklick J-P. Towards a model- and data-focused taxonomy of XAI systems. In: <i>Wirtschaftsinformatik 2022 Proceedings</i>. ; 2022.","chicago":"Kucklick, Jan-Peter. “Towards a Model- and Data-Focused Taxonomy of XAI Systems.” In <i>Wirtschaftsinformatik 2022 Proceedings</i>, 2022.","ieee":"J.-P. Kucklick, “Towards a model- and data-focused taxonomy of XAI systems,” presented at the Wirtschaftsinformatik 2022 (WI22), Nürnberg (online), 2022."},"_id":"29539","user_id":"77066","department":[{"_id":"195"},{"_id":"196"}],"keyword":["Explainable Artificial Intelligence","XAI","Interpretability","Decision Support Systems","Taxonomy"],"language":[{"iso":"eng"}],"type":"conference","publication":"Wirtschaftsinformatik 2022 Proceedings","abstract":[{"text":"Explainable Artificial Intelligence (XAI) is currently an important topic for the application of Machine Learning (ML) in high-stakes decision scenarios. Related research focuses on evaluating ML algorithms in terms of interpretability. However, providing a human understandable explanation of an intelligent system does not only relate to the used ML algorithm. The data and features used also have a considerable impact on interpretability. In this paper, we develop a taxonomy for describing XAI systems based on aspects about the algorithm and data. The proposed taxonomy gives researchers and practitioners opportunities to describe and evaluate current XAI systems with respect to interpretability and guides the future development of this class of systems.","lang":"eng"}],"status":"public"}]