We propose Projection Pursuit CPCANet (PP-CPCANet), a covariance-free framework for domain generalization that avoids rank-deficient covariance estimation in mini-batch training. By jointly optimizing a global orthogonal basis on the Stiefel manifold via the Cayley transform and a robust PP dispersion objective, PP-CPCANet learns common principal components with stable optimization. Experiments on four DG benchmarks demonstrate SOTA performance.</p>\n","updatedAt":"2026-07-29T11:39:13.338Z","author":{"_id":"67e2063e1ee7f6db889849d6","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/67e2063e1ee7f6db889849d6/ihiwCCqbXlxQ2V_SSGnng.jpeg","fullname":"Yu-Hsi Chen","name":"wish44165","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7670679092407227},"editors":["wish44165"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/67e2063e1ee7f6db889849d6/ihiwCCqbXlxQ2V_SSGnng.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2607.22117","authors":[{"_id":"6a69e62eb88bbce873be3a19","user":{"_id":"67e2063e1ee7f6db889849d6","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/67e2063e1ee7f6db889849d6/ihiwCCqbXlxQ2V_SSGnng.jpeg","isPro":false,"fullname":"Yu-Hsi Chen","user":"wish44165","type":"user","name":"wish44165"},"name":"Yu-Hsi Chen","status":"claimed_verified","statusLastChangedAt":"2026-07-29T16:45:04.826Z","hidden":false},{"_id":"6a69e62eb88bbce873be3a1a","name":"Abd-Krim Seghouane","hidden":false}],"publishedAt":"2026-07-24T00:00:00.000Z","submittedOnDailyAt":"2026-07-29T00:00:00.000Z","title":"Projection Pursuit CPCANet for Domain Generalization","submittedOnDailyBy":{"_id":"67e2063e1ee7f6db889849d6","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/67e2063e1ee7f6db889849d6/ihiwCCqbXlxQ2V_SSGnng.jpeg","isPro":false,"fullname":"Yu-Hsi Chen","user":"wish44165","type":"user","name":"wish44165"},"summary":"Domain Generalization (DG) aims to learn representations robust to distribution shifts. Recent geometric alignment methods, such as CPCANet, extract domain-invariant structures through batch-wise Common Principal Component Analysis (CPCA). However, CPCANet suffers from rank-deficient covariance estimation due to the small-sample-size issue in mini-batch training. To address this limitation, we propose Projection Pursuit CPCANet (PP-CPCANet), a covariance-free framework that learns a global orthogonal basis on the Stiefel manifold and jointly optimizes it with network parameters via the Cayley transform. We further introduce a symmetry-breaking detached-median PP dispersion objective to extract common principal components (CPCs) with dense and robust optimization signals. Experiments on four DG benchmarks show that PP-CPCANet achieves SOTA performance while maintaining stable training.","upvotes":0,"discussionId":"6a69e62fb88bbce873be3a1b","githubRepo":"https://github.com/wish44165/PP-CPCANet","githubRepoAddedBy":"user","githubStars":0,"organization":{"_id":"6530ecba05ba6e63104a253a","name":"unimelb-nlp","fullname":"The University of Melbourne","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/6530ebbb32d27986ec470501/E49xVFH2XUolq-RM9JTaf.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[],"acceptLanguages":["en"],"organization":{"_id":"6530ecba05ba6e63104a253a","name":"unimelb-nlp","fullname":"The University of Melbourne","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/6530ebbb32d27986ec470501/E49xVFH2XUolq-RM9JTaf.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2607/2607.22117.md","query":{}}">
Projection Pursuit CPCANet for Domain Generalization
Abstract
Domain Generalization (DG) aims to learn representations robust to distribution shifts. Recent geometric alignment methods, such as CPCANet, extract domain-invariant structures through batch-wise Common Principal Component Analysis (CPCA). However, CPCANet suffers from rank-deficient covariance estimation due to the small-sample-size issue in mini-batch training. To address this limitation, we propose Projection Pursuit CPCANet (PP-CPCANet), a covariance-free framework that learns a global orthogonal basis on the Stiefel manifold and jointly optimizes it with network parameters via the Cayley transform. We further introduce a symmetry-breaking detached-median PP dispersion objective to extract common principal components (CPCs) with dense and robust optimization signals. Experiments on four DG benchmarks show that PP-CPCANet achieves SOTA performance while maintaining stable training.
Community
We propose Projection Pursuit CPCANet (PP-CPCANet), a covariance-free framework for domain generalization that avoids rank-deficient covariance estimation in mini-batch training. By jointly optimizing a global orthogonal basis on the Stiefel manifold via the Cayley transform and a robust PP dispersion objective, PP-CPCANet learns common principal components with stable optimization. Experiments on four DG benchmarks demonstrate SOTA performance.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2607.22117 in a model README.md to link it from this page.
Cite arxiv.org/abs/2607.22117 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2607.22117 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.