{"id":55,"date":"2025-05-09T23:01:08","date_gmt":"2025-05-09T23:01:08","guid":{"rendered":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/?page_id=55"},"modified":"2025-12-11T06:31:06","modified_gmt":"2025-12-11T06:31:06","slug":"results","status":"publish","type":"page","link":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/","title":{"rendered":"Results"},"content":{"rendered":"\n<p>The table below compares the accuracy on balanced test set for different methods<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"250\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM-1024x250.png\" alt=\"\" class=\"wp-image-114\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM-1024x250.png 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM-300x73.png 300w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM-768x187.png 768w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM.png 1264w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption class=\"wp-element-caption\">Table 1. Comparison of class-balanced and bias-conflicting accuracies on different baseline methods on standard benchmarks.<\/figcaption><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"232\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.19-AM-1024x232.png\" alt=\"\" class=\"wp-image-116\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.19-AM-1024x232.png 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.19-AM-300x68.png 300w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.19-AM-768x174.png 768w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.19-AM.png 1256w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption class=\"wp-element-caption\">Table 2. Comparison of class-balanced and bias-conflict accuracies on different baseline methods on challenging benchmarks.<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\">Visualizations<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"451\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/samples_page-0001-1024x451.jpg\" alt=\"\" class=\"wp-image-117\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/samples_page-0001-1024x451.jpg 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/samples_page-0001-300x132.jpg 300w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/samples_page-0001-768x338.jpg 768w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/samples_page-0001-1536x677.jpg 1536w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/samples_page-0001.jpg 1823w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption class=\"wp-element-caption\">Figure 1. Illustration of the NMS debiasing process. The top row shows the original images, while the bottom row shows the context-swapped images generated using the NMS strategy. NMS effectively swaps contextual attributes while preserving the target attribute, facilitating the removal of spurious correlations.<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\">Ablation Studies<\/h2>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"223\" src=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/ablation_page-0001-1024x223.jpg\" alt=\"\" class=\"wp-image-120\" srcset=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/ablation_page-0001-1024x223.jpg 1024w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/ablation_page-0001-300x65.jpg 300w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/ablation_page-0001-768x167.jpg 768w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/ablation_page-0001-1536x334.jpg 1536w, https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/ablation_page-0001-2048x445.jpg 2048w\" sizes=\"auto, (max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><figcaption class=\"wp-element-caption\">Figure 2. Effect of target subspace dimension k on debiasing performance across three datasets.<\/figcaption><\/figure>\n\n\n\n<p>The ablation study examines the effect of different choices of ( k ) (the target subspace dimension) on performance. The results are presented in Figure 2., where each column corresponds to a dataset: CelebA-G&amp;B-0%, CelebA-H&amp;E-0%, and Dogs&amp;Cats-Fur-0%. The blue line displays performance on class-balanced samples, while the red line shows performance on bias-conflicting samples. The yellow bar in each subplot represents the distribution of singular values in the Pairwise Difference Matrix.<\/p>\n\n\n\n<p>From the figure, it is evident that the choice of ( k ) significantly impacts debiasing performance. Specifically, for bias-conflicting samples, the success of debiasing, measured by bias-conflicting accuracy gradually declines as ( k ) increases. This occurs because a higher ( k ) retains more nuisance attributes, and NMS fails to marginalize these attributes. However, for class-balanced samples, performance does not always decrease as ( k ) increases. For instance, in CelebA-G&amp;B, accuracy steadily declines with increasing ( k ), whereas in CelebA-H&amp;E, it gradually improves. In Dogs&amp;Cats-Fur, accuracy initially rises before dropping. This suggests that while increasing ( k ) can introduce more bias, in some cases, this added bias may improve overall accuracy by enhancing performance on bias-aligned samples, depending on the dataset\u2019s characteristics.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>The table below compares the accuracy on balanced test set for different methods Visualizations Ablation Studies The ablation study examines the effect of different choices of ( k ) (the target subspace dimension) on performance. The results are presented in Figure 2., where each column corresponds to a dataset: CelebA-G&amp;B-0%, CelebA-H&amp;E-0%, and Dogs&amp;Cats-Fur-0%. The blue &hellip; <\/p>\n<p class=\"link-more\"><a href=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/\" class=\"more-link\">Continue reading<span class=\"screen-reader-text\"> &#8220;Results&#8221;<\/span><\/a><\/p>\n","protected":false},"author":243,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-55","page","type-page","status-publish","hentry"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.4 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Results - Debiasing vision datasets<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Results - Debiasing vision datasets\" \/>\n<meta property=\"og:description\" content=\"The table below compares the accuracy on balanced test set for different methods Visualizations Ablation Studies The ablation study examines the effect of different choices of ( k ) (the target subspace dimension) on performance. The results are presented in Figure 2., where each column corresponds to a dataset: CelebA-G&amp;B-0%, CelebA-H&amp;E-0%, and Dogs&amp;Cats-Fur-0%. The blue &hellip; Continue reading &quot;Results&quot;\" \/>\n<meta property=\"og:url\" content=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/\" \/>\n<meta property=\"og:site_name\" content=\"Debiasing vision datasets\" \/>\n<meta property=\"article:modified_time\" content=\"2025-12-11T06:31:06+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1264\" \/>\n\t<meta property=\"og:image:height\" content=\"308\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"3 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/\",\"name\":\"Results - Debiasing vision datasets\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/wp-content\\\/uploads\\\/sites\\\/128\\\/2025\\\/12\\\/Screenshot-2025-12-11-at-12.47.04-AM-1024x250.png\",\"datePublished\":\"2025-05-09T23:01:08+00:00\",\"dateModified\":\"2025-12-11T06:31:06+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/#primaryimage\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/wp-content\\\/uploads\\\/sites\\\/128\\\/2025\\\/12\\\/Screenshot-2025-12-11-at-12.47.04-AM.png\",\"contentUrl\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/wp-content\\\/uploads\\\/sites\\\/128\\\/2025\\\/12\\\/Screenshot-2025-12-11-at-12.47.04-AM.png\",\"width\":1264,\"height\":308},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/results\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Results\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/#website\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/\",\"name\":\"Debiasing vision datasets\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025team12-1\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Results - Debiasing vision datasets","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/","og_locale":"en_US","og_type":"article","og_title":"Results - Debiasing vision datasets","og_description":"The table below compares the accuracy on balanced test set for different methods Visualizations Ablation Studies The ablation study examines the effect of different choices of ( k ) (the target subspace dimension) on performance. The results are presented in Figure 2., where each column corresponds to a dataset: CelebA-G&amp;B-0%, CelebA-H&amp;E-0%, and Dogs&amp;Cats-Fur-0%. The blue &hellip; Continue reading \"Results\"","og_url":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/","og_site_name":"Debiasing vision datasets","article_modified_time":"2025-12-11T06:31:06+00:00","og_image":[{"width":1264,"height":308,"url":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM.png","type":"image\/png"}],"twitter_card":"summary_large_image","twitter_misc":{"Est. reading time":"3 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/","url":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/","name":"Results - Debiasing vision datasets","isPartOf":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/#website"},"primaryImageOfPage":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/#primaryimage"},"image":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/#primaryimage"},"thumbnailUrl":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM-1024x250.png","datePublished":"2025-05-09T23:01:08+00:00","dateModified":"2025-12-11T06:31:06+00:00","breadcrumb":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/#primaryimage","url":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM.png","contentUrl":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-content\/uploads\/sites\/128\/2025\/12\/Screenshot-2025-12-11-at-12.47.04-AM.png","width":1264,"height":308},{"@type":"BreadcrumbList","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/results\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/"},{"@type":"ListItem","position":2,"name":"Results"}]},{"@type":"WebSite","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/#website","url":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/","name":"Debiasing vision datasets","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"}]}},"_links":{"self":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/pages\/55","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/users\/243"}],"replies":[{"embeddable":true,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/comments?post=55"}],"version-history":[{"count":4,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/pages\/55\/revisions"}],"predecessor-version":[{"id":121,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/pages\/55\/revisions\/121"}],"wp:attachment":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025team12-1\/wp-json\/wp\/v2\/media?parent=55"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}