{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,5]],"date-time":"2026-08-05T06:45:00Z","timestamp":1785912300817,"version":"3.56.0"},"reference-count":98,"publisher":"IEEE","license":[{"start":{"date-parts":[[2026,3,1]],"date-time":"2026-03-01T00:00:00Z","timestamp":1772323200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,3,1]],"date-time":"2026-03-01T00:00:00Z","timestamp":1772323200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,3,1]],"date-time":"2026-03-01T00:00:00Z","timestamp":1772323200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"NSF","award":["2323086,2223839"],"award-info":[{"award-number":["2323086,2223839"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,3]]},"DOI":"10.1109\/wacvw68408.2026.00142","type":"proceedings-article","created":{"date-parts":[[2026,8,4]],"date-time":"2026-08-04T19:17:40Z","timestamp":1785871060000},"page":"1307-1316","source":"Crossref","is-referenced-by-count":0,"title":["Stabilizing Intrinsic Explanations: A Geometric Perspective on Concept Leakage"],"prefix":"10.1109","author":[{"given":"Keun Hee","family":"Park","sequence":"first","affiliation":[{"name":"School of Computing and Augmented Intelligence, Arizona State University","place":["Tempe, AZ, USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jinyung","family":"Hong","sequence":"additional","affiliation":[{"name":"Barrow Neurological Institute","place":["Phoenix, AZ, USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Theodore P.","family":"Pavlic","sequence":"additional","affiliation":[{"name":"School of Life Sciences, Arizona State University","place":["Tempe, AZ, USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pavan","family":"Turaga","sequence":"additional","affiliation":[{"name":"The GAME School, Arizona State University","place":["Tempe, AZ, USA"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Sanity checks for saliency maps","volume":"31","author":"Adebayo","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"ref2","article-title":"Towards robust interpretability with self-explaining neural networks","volume":"31","author":"Melis","year":"2018","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref3","article-title":"On the robustness of interpretability methods","volume":"2","author":"Alvarez-Melis","year":"2018","journal-title":"arXiv preprint"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i6.20551"},{"key":"ref5","first-page":"17","article-title":"On manifold regularization.","volume-title":"International Workshop on Artificial Intelligence and Statistics","volume":"4","author":"Belkin","year":"2005"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1903070116"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-39593-2_1"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2013.50"},{"key":"ref9","first-page":"217","article-title":"When is \u201cnearest neighbor","volume-title":"meaningful? In International conference on database theory","volume":"2","author":"Beyer","year":"1999"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1995.7.1.108"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1175\/1520-0493(1950)078<0001:VOFEIT>2.0.CO;2"},{"key":"ref12","article-title":"This looks like that: deep learning for interpretable image recognition","volume":"32","author":"Chen","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1038\/s42256-020-00265-z"},{"key":"ref14","first-page":"1310","article-title":"Certified adversarial robustness via randomized smoothing","volume-title":"international conference on machine learning","author":"Cohen","year":"2019"},{"key":"ref15","first-page":"2206","article-title":"Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks","volume-title":"International conference on machine learning","author":"Croce","year":"2020"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01002"},{"key":"ref18","article-title":"An image is worth 16\u00d716 words: Transformers for image recognition at scale","author":"Dosovitskiy","year":"2020","journal-title":"arXiv preprint"},{"key":"ref19","article-title":"Adversarial robustness as a prior for learned representations","author":"Engstrom","year":"2019","journal-title":"arXiv preprint"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1007\/s10994-017-5663-3"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1090\/jams\/852"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.476"},{"key":"ref23","article-title":"Information-theoretic generalization analysis for expected calibration error","author":"Fujisawa","year":"2024","journal-title":"arXiv preprint"},{"key":"ref24","first-page":"7549","article-title":"Generalisation in humans and deep neural networks","author":"Geirhos","year":"2018","journal-title":"Advances in Neural Information Processing Systems (NeurIPS)"},{"key":"ref25","first-page":"23885","article-title":"Partial success in closing the gap between human and machine vision","volume":"34","author":"Geirhos","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.33013681"},{"key":"ref27","article-title":"Towards automatic concept-based explanations","author":"Ghorbani","year":"2019","journal-title":"Advances in Neural Information Processing Systems, 32"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.169"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2014.81"},{"key":"ref30","article-title":"Explaining and harnessing adversarial examples","author":"Goodfellow","year":"2014","journal-title":"arXiv preprint"},{"key":"ref31","article-title":"Explaining classifiers with causal concept effect (CaCE)","author":"Goyal","year":"2019","journal-title":"arXiv preprint"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0699"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.33013714"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1137\/1004050"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1699"},{"key":"ref36","article-title":"Formal guarantees on the robustness of a classifier against adversarial manipulation","volume":"30","author":"Hein","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.1.1"},{"key":"ref38","volume-title":"Putting the Mind and the Body Into AI Using Random Projections, Neuromodulation, and Cognitive Modules to Improve the Design of Modern AI Systems","author":"Hong","year":"2025"},{"key":"ref39","first-page":"315","article-title":"Randomly weighted neuromodulation in neural networks facilitates learning of manifolds common across tasks","volume-title":"Proceedings of UniReps: the First Workshop on Unifying Representations in Neural Models","author":"Hong","year":"2024"},{"key":"ref40","article-title":"Learning decomposable and debiased representations via attribute-centric information bottlenecks","author":"Hong","year":"2024","journal-title":"arXiv preprint"},{"key":"ref41","article-title":"Debiasing global workspace: A cognitive neural framework for learning debiased and interpretable representations","volume-title":"NeurIPS 2024 Workshop on Behavioral Machine Learning","author":"Hong","year":"2024"},{"key":"ref42","article-title":"A cognitive framework for learning debiased and interpretable representations via debiasing global workspace","volume-title":"UniReps: 2nd Edition of the Workshop on Unifying Representations in Neural Models","author":"Hong","year":"2024"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/WACV57701.2024.00481"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/WACV61042.2026.00374"},{"key":"ref45","first-page":"1173","article-title":"Partstacked CNN for fine-grained visual categorization","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Huang","year":"2016"},{"key":"ref46","article-title":"Spatial transformer networks","volume":"28","author":"Jaderberg","year":"2015","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref47","article-title":"Now you see me (cme): concept-based model extraction","author":"Kazhdan","year":"2020","journal-title":"arXiv preprint"},{"key":"ref48","first-page":"2668","article-title":"Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV)","volume-title":"International Conference on Machine Learning","author":"Kim","year":"2018"},{"key":"ref49","first-page":"5338","article-title":"Concept bottleneck models","volume-title":"International conference on machine learning","author":"Wei Koh","year":"2020"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v32i1.11771"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.170"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1145\/3236386.3241340"},{"key":"ref53","first-page":"11525","article-title":"Objectcentric learning with slot attention","volume":"33","author":"Locatello","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.3321\/j.issn:0529-6579.2007.z1.029"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.06083"},{"key":"ref56","article-title":"Measuring leakage in concept-based methods: An information theoretic approach","volume-title":"ICLR 2025 Workshop: XAI4Science: From Understanding Model Behavior to Discovering New Scientific Knowledge","volume":"2","author":"Makonnen","year":"2025"},{"key":"ref57","article-title":"Are sixteen heads really better than one?","volume":"32","author":"Michel","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1351"},{"key":"ref59","article-title":"Interpretable machine learning","volume":"2","author":"Molnar","year":"2020","journal-title":"Lulu. com"},{"key":"ref60","article-title":"Robustness of classifiers to universal perturbations: A geometric perspective","volume-title":"International Conference on Learning Representations","author":"Moosavi-Dezfooli","year":"2018"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.00269"},{"key":"ref62","article-title":"Synthesizing the preferred inputs for neurons in neural networks via deep generator networks","volume":"29","author":"Nguyen","year":"2016","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref63","article-title":"Leakage and interpretability in concept-based models","author":"Parisini","year":"2025","journal-title":"arXiv preprint"},{"key":"ref64","article-title":"Integrating adversarial training, noise injection, and mixup into xai: Pathways to enhancing data efficiency and generalizability","volume-title":"Master\u2019s thesis, Arizona State University","author":"Park","year":"2024"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-86380-7_12"},{"key":"ref66","article-title":"Adversarial training can hurt generalization","volume-title":"ICML 2019 Workshop on Identifying and Understanding Deep Learning Phenomena","volume":"2","author":"Raghunathan","year":"2019"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1145\/2939672.2939778"},{"key":"ref68","first-page":"833","article-title":"Contractive auto-encoders: Explicit invariance during feature extraction","volume-title":"Proceedings of the 28th international conference on international conference on machine learning","author":"Rifai","year":"2011"},{"key":"ref69","article-title":"Attention-based interpretability with concept transformers","volume-title":"International Conference on Learning Representations","author":"Rigotti","year":"2021"},{"key":"ref70","first-page":"410","article-title":"V-measure: A conditional entropy-based external cluster evaluation measure","volume-title":"Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL)","author":"Rosenberg","year":"2007"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19775-8_21"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/ICCVW69036.2025.00071"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.74"},{"key":"ref74","article-title":"On the adversarial robustness of vision transformers","author":"Shao","year":"2022","journal-title":"Transactions on Machine Learning Research"},{"key":"ref75","article-title":"Smoothgrad: removing noise by adding noise","author":"Smilkov","year":"2017","journal-title":"arXiv preprint"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2017.2708039"},{"issue":"1","key":"ref77","first-page":"1929","article-title":"Dropout: a simple way to prevent neural networks from overfitting","volume":"15","author":"Srivastava","year":"2014","journal-title":"The journal of machine learning research"},{"key":"ref78","first-page":"3319","article-title":"Axiomatic attribution for deep networks","volume-title":"International Conference on Machine Learning","author":"Sundararajan","year":"2017"},{"key":"ref79","article-title":"Intriguing properties of neural networks","volume":"2","author":"Szegedy","year":"2013","journal-title":"arXiv preprint"},{"key":"ref80","article-title":"Ensemble adversarial training: Attacks and defenses","volume-title":"6th International Conference on Learning Representations, ICLR 2018-Conference Track Proceedings","author":"Tram\u00e8r","year":"2018"},{"key":"ref81","first-page":"1747","article-title":"Pixel recurrent neural networks","volume-title":"International Conference on Machine Learning","author":"Van Den Oord","year":"2016"},{"key":"ref82","first-page":"2579","article-title":"Visualizing data using t-sne","volume":"9","author":"van der Maaten","year":"2008","journal-title":"Journal of Machine Learning Research"},{"key":"ref83","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref84","first-page":"6438","article-title":"Manifold mixup: Better representations by interpolating hidden states","volume-title":"International conference on machine learning","author":"Verma","year":"2019"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1580"},{"key":"ref86","article-title":"Caltech-UCSD Birds-200\u20132011 (CUB-200\u20132011) dataset","volume-title":"Technical Report CNS-TR-2011-001, California Institute of Technology","author":"Wah","year":"2011"},{"key":"ref87","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01055"},{"key":"ref88","first-page":"842","article-title":"The application of twolevel attention models in deep convolutional neural network for fine-grained image classification","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Xiao","year":"2015"},{"key":"ref89","article-title":"Protopformer: Concentrating on prototypical parts in vision transformers for interpretable image recognition","author":"Xue","year":"2024","journal-title":"IJCAI"},{"key":"ref90","first-page":"20554","article-title":"On completenessaware concept-based explanations in deep neural networks","volume":"33","author":"Yeh","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref91","article-title":"Understanding neural networks through deep visualization","author":"Yosinski","year":"2015","journal-title":"arXiv preprint"},{"key":"ref92","article-title":"Concept embedding models","volume-title":"NeurIPS 2022-36th Conference on Neural Information Processing Systems","author":"Zarlenga","year":"2022"},{"key":"ref93","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.129"},{"key":"ref94","article-title":"mixup: Beyond empirical risk minimization","volume-title":"International Conference on Learning Representations","author":"Zhang","year":"2018"},{"key":"ref95","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10590-1_54"},{"key":"ref96","first-page":"5209","article-title":"Learning multi-attention convolutional neural network for finegrained image recognition","volume-title":"Proceedings of the IEEE International Conference on Computer Vision","author":"Zheng","year":"2017"},{"key":"ref97","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.319"},{"key":"ref98","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01237-3_8"}],"event":{"name":"2026 IEEE\/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW)","location":"Tucson, AZ, USA","start":{"date-parts":[[2026,3,6]]},"end":{"date-parts":[[2026,3,10]]}},"container-title":["2026 IEEE\/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11617369\/11617396\/11617691.pdf?arnumber=11617691","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,5]],"date-time":"2026-08-05T05:51:37Z","timestamp":1785909097000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11617691\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3]]},"references-count":98,"URL":"https:\/\/doi.org\/10.1109\/wacvw68408.2026.00142","relation":{},"subject":[],"published":{"date-parts":[[2026,3]]}}}