diff options
41 files changed, 151 insertions, 196 deletions
diff --git a/site/content/pages/research/munich_security_conference/assets/embassy_counts_summary_dataset.csv b/site/content/pages/research/munich_security_conference/assets/embassy_counts_summary_dataset.csv index 89f3c226..3a439821 100755 --- a/site/content/pages/research/munich_security_conference/assets/embassy_counts_summary_dataset.csv +++ b/site/content/pages/research/munich_security_conference/assets/embassy_counts_summary_dataset.csv @@ -1,5 +1,5 @@ dataset,images -ibm_dif,389 -megaface,5679 -vgg_face,1 -who_goes_there,2372 +IBM Diversity in Faces,389 +MegaFace,5679 +VGG Face,1 +Who Goes There,2372
\ No newline at end of file diff --git a/site/content/pages/research/munich_security_conference/assets/megapixels_origins_top.csv b/site/content/pages/research/munich_security_conference/assets/megapixels_origins_top.csv index 081b4636..ae6e8f11 100755 --- a/site/content/pages/research/munich_security_conference/assets/megapixels_origins_top.csv +++ b/site/content/pages/research/munich_security_conference/assets/megapixels_origins_top.csv @@ -1,9 +1,8 @@ source,images -Search Engines,30127200 -Flickr.com,11783888 -IMDb.com,5251410 -CCTV,959312 -Wikimedia.org,183500 -Mugshots,113268 -YouTube.com,31888 -Other Sources Combined,37044 +Internet Search Engines,15063600 +Flickr.com,5891944 +Internet Movie Database (IMDB.com),2625705 +CCTV,479656 +Wikimedia.org,91750 +Mugshots,56634 +YouTube.com,15944
\ No newline at end of file diff --git a/site/content/pages/research/munich_security_conference/index.md b/site/content/pages/research/munich_security_conference/index.md index 2a97a7c0..0f8a5bda 100644 --- a/site/content/pages/research/munich_security_conference/index.md +++ b/site/content/pages/research/munich_security_conference/index.md @@ -3,7 +3,7 @@ status: published title: MSC slug: munich-security-conference -desc: Analyzing the Transnational Flow of Facial Recognition Training Data +desc: Analyzing Transnational Flows of Face Recognition Image Training Data subdesc: Where does face data originate and who's using it? cssclass: dataset image: assets/background.jpg @@ -13,7 +13,7 @@ authors: Adam Harvey ------------ -## Analysis for the Munich Security Conference Transnational Security Report +## Face Datasets and Information Supply Chains ### sidebar @@ -21,21 +21,30 @@ authors: Adam Harvey + Datasets Analyzed: 30 + Years: 2006 - 2018 + Status: Ongoing Investigation -+ Last Updated: June 27, 2019 ++ Last Updated: June 28, 2019 ### end sidebar +National AI strategies often rely on transnational data sources to capitalize on recent advancements in deep learning and neural networks. Researchers benefiting from these transnational data flows can yield quick and significant gains across diverse sectors from health care to biometrics. But new challenges emerge when national AI strategies collide with national interests. -Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum." +Our earlier research on the [MS Celeb](/datsets) and [Duke](/datsets/duke_mtmc) datasets published with the Financial Times revealed that several computer vision image datasets created by US companies and universities were unexpectedly also used for research by the National University of Defense Technology in China, along with top Chinese surveillance firms including SenseTime, SenseNets, CloudWalk, Hikvision, and Megvii/Face++ which have all been linked to the oppressive surveillance of Uighur Muslims in Xinjiang. + +In this new research for the Munich Security Conference's Transnational Security Report we provide summary statistics about the origins and endpoints of facial recognition information supply chains. To make it more personal, we gathered additional data on the number of public photos from Embassies that are currently being used in facial recognition datasets. + + +### 24 Million Non-Cooperative Faces + +In total, we analyzed 30 publicly available face recognition and face analysis datasets that collectively include over 24 million non-cooperative images. Of these 24 million images, over 15 million face images are from Internet search engines, over 5.8 million from Flickr.com, over 2.5 million from the Internet Movie Database (IMDb.com), and nearly 500,000 from CCTV footage. All 24 million images were collected without any explicit consent, a type of face image researchers call "in the wild". + +Next we manually verified 1,134 publicly available research papers that cite these datasets to determine who was using the data and where it was being used. Even though all of the images originated in the United States, the publicly available research citations show that only about 25% citations are from the country of the origin while the majority of citations are from China. -Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum." === columns 2 ``` single_pie_chart /site/research/munich_security_conference/assets/megapixels_origins_top.csv -Caption: Sources of Publicly Available Face Training Data 2006 - 2018 +Caption: Sources of Publicly Available Non-Cooperative Face Image Training Data 2006 - 2018 Top: 10 OtherLabel: Other ``` @@ -44,85 +53,32 @@ OtherLabel: Other ``` single_pie_chart /site/research/munich_security_conference/assets/summary_countries.csv -Caption: Locations Where Face Data Is Used +Caption: Locations Where Face Data Is Used Based on Public Research Citations Top: 14 OtherLabel: Other ``` === end columns +### 6,000 Embassy Photos Being Used To Train Facial Recognition -=== columns 2 - -#### Sources of Face Data - -Add text - -| Source | Images | -| --- | --- | -|Search Engines | 30,127,200 | -|Flickr.com | 11,783,888 | -|IMDb.com | 5,251,410 | -|CCTV | 959,312 | -|Wikimedia.org | 183,500 | -|Mugshots | 113,268 | -|Other Sources Combined | 37,044 | -|YouTube.com | 31,888 | - -=== - -#### Where Face Data Is Used - -Add text - -|country | citations| -| --- | --- | -|China | 327| -|United States | 302| -|United Kingdom | 187| -|Australia | 38| -|Germany | 35| -|Singapore | 27| -|Canada | 25| -|Netherlands | 25| -|Italy | 22| -|France | 17| -|India | 14| -|South Korea | 12| -|Spain | 10| -|Switzerland | 9| - -=== end columns - - - -## Over 6,000 Embassy Images on Flickr Found in Face Recognition Datasets - -Including over 2,000 more for racial analysis - - - - - - - - +Of the 5.8 million Flickr images we found over 6,000 public photos from Embassy Flickr accounts were used to train facial recognition technologies. These images were used in the MegaFace, IBM Diversity in Faces datasets. Over 2,000 more images were used in the Who Goes There datasets used for facial ethnicity analysis research. A few of the embassy images found in facial recognition datasets are shown below. === columns 2 ``` -single_pie_chart /site/research/munich_security_conference/assets/megapixels_origins_top.csv -Caption: Sources of Face Training Data -Top: 5 -OtherLabel: Other Countries +single_pie_chart /site/research/munich_security_conference/assets/country_counts.csv +Caption: Photos from these embassies are being used to train face recognition software +Top: 4 +OtherLabel: Other Colors: categoryRainbow ``` -=========== +===== ``` single_pie_chart /site/research/munich_security_conference/assets/embassy_counts_summary_dataset.csv -Caption: Dataset sources +Caption: Embassy images were found in these datasets Top: 4 OtherLabel: Other Colors: categoryRainbow @@ -130,6 +86,29 @@ Colors: categoryRainbow === end columns + + + + + + +This brief research aims to shed light on the emerging politics of data. A photo is no longer just a photo when it can also be surveillance training data, and datasets can no longer be separated from the development of software when software is now built with data. "Our relationship to computers has changed", says Geoffrey Hinton, one of the founders of modern day neural networks and deep learning. "Instead of programming them, we now show them and they figure it out."[^hinton]. + +National AI strategies might also want to include transnational dataset strategies. + +*This research post is going and will updated during July and August, 2019.* + +### Further Reading + +- [MS Celeb Dataset Analysis](/datasets/msceleb) +- [Brainwash Dataset Analysis](/datasets/brainwash) +- [Duke MTMC Dataset Analysis](/datasets/duke_mtmc) +- [Unconstrained College Students Dataset Analysis](/datasets/uccs) +- [Duke MTMC dataset author apologies to students](https://www.dukechronicle.com/article/2019/06/duke-university-facial-recognition-data-set-study-surveillance-video-students-china-uyghur) +- [BBC coverage of MS Celeb dataset takedown](https://www.bbc.com/news/technology-48555149) +- [Spiegel coverage of MS Celeb dataset takdown](https://www.spiegel.de/netzwelt/web/microsoft-gesichtserkennung-datenbank-mit-zehn-millionen-fotos-geloescht-a-1271221.html) + + {% include 'supplementary_header.html' %} ``` @@ -137,5 +116,10 @@ load_file /site/research/munich_security_conference/assets/embassy_counts_public Headings: Images, Dataset, Embassy, Flickr ID, URL, Guest, Host ``` +{% include 'cite_our_work.html' %} + +### Footnotes + +[^hinton]: "Heroes of Deep Learning: Andrew Ng interviews Geoffrey Hinton". Published on Aug 8, 2017. <https://www.youtube.com/watch?v=-eyhCTvrEtE> + -{% include 'cite_our_work.html' %}
\ No newline at end of file diff --git a/site/public/about/assets/LICENSE/index.html b/site/public/about/assets/LICENSE/index.html index f1e3a9fd..40929e4f 100644 --- a/site/public/about/assets/LICENSE/index.html +++ b/site/public/about/assets/LICENSE/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/about/attribution/index.html b/site/public/about/attribution/index.html index 15270150..4e7474b0 100644 --- a/site/public/about/attribution/index.html +++ b/site/public/about/attribution/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-about"> diff --git a/site/public/about/index.html b/site/public/about/index.html index 16a2e967..a46653c6 100644 --- a/site/public/about/index.html +++ b/site/public/about/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-about"> diff --git a/site/public/about/legal/index.html b/site/public/about/legal/index.html index 49ed926d..8beafeea 100644 --- a/site/public/about/legal/index.html +++ b/site/public/about/legal/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-about"> diff --git a/site/public/about/news/index.html b/site/public/about/news/index.html index fcba7877..de44468e 100644 --- a/site/public/about/news/index.html +++ b/site/public/about/news/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-about"> diff --git a/site/public/datasets/brainwash/index.html b/site/public/datasets/brainwash/index.html index 3dacd6e1..18600b6f 100644 --- a/site/public/datasets/brainwash/index.html +++ b/site/public/datasets/brainwash/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/duke_mtmc/index.html b/site/public/datasets/duke_mtmc/index.html index 9a70a3f6..fc141450 100644 --- a/site/public/datasets/duke_mtmc/index.html +++ b/site/public/datasets/duke_mtmc/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/helen/index.html b/site/public/datasets/helen/index.html index a7ada42a..44ef462e 100644 --- a/site/public/datasets/helen/index.html +++ b/site/public/datasets/helen/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/hrt_transgender/index.html b/site/public/datasets/hrt_transgender/index.html index 02324a2f..2e5e9c62 100644 --- a/site/public/datasets/hrt_transgender/index.html +++ b/site/public/datasets/hrt_transgender/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/ibm_dif/index.html b/site/public/datasets/ibm_dif/index.html index 1c465f93..be5dbfe4 100644 --- a/site/public/datasets/ibm_dif/index.html +++ b/site/public/datasets/ibm_dif/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/ijb_c/index.html b/site/public/datasets/ijb_c/index.html index a36fac14..abe7d5ed 100644 --- a/site/public/datasets/ijb_c/index.html +++ b/site/public/datasets/ijb_c/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/index.html b/site/public/datasets/index.html index 1fb83352..a634b877 100644 --- a/site/public/datasets/index.html +++ b/site/public/datasets/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/datasets/megaface/index.html b/site/public/datasets/megaface/index.html index 33abf6c1..712af28a 100644 --- a/site/public/datasets/megaface/index.html +++ b/site/public/datasets/megaface/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/msceleb/assets/notes/index.html b/site/public/datasets/msceleb/assets/notes/index.html index cac21eef..36c32429 100644 --- a/site/public/datasets/msceleb/assets/notes/index.html +++ b/site/public/datasets/msceleb/assets/notes/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/datasets/msceleb/index.html b/site/public/datasets/msceleb/index.html index 7109cc9b..42a44571 100644 --- a/site/public/datasets/msceleb/index.html +++ b/site/public/datasets/msceleb/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> @@ -212,8 +212,8 @@ <p>Despite the recent termination of the <a href="https://msceleb.org">msceleb.org</a> website, the dataset still exists in several repositories on GitHub, the hard drives of countless researchers, and will likely continue to be used in research projects around the world.</p> <p>For example, on October 28, 2019, the MS Celeb dataset will be used for a new competition called "<a href="https://ibug.doc.ic.ac.uk/resources/lightweight-face-recognition-challenge-workshop/">Lightweight Face Recognition Challenge & Workshop</a>" where the best face recognition entries will be awarded $5,000 from Huawei and $3,000 from DeepGlint. The competition is part of the <a href="http://iccv2019.thecvf.com/program/workshops">ICCV 2019 conference</a>. This time the challenge is no longer being organized by Microsoft, who created the dataset, but instead by Imperial College London (UK) and <a href="https://github.com/deepinsight/insightface">InsightFace</a> (CN). The organizers provide a <a href="https://ibug.doc.ic.ac.uk/resources/lightweight-face-recognition-challenge-workshop/">25GB download of cropped faces</a> from MS Celeb for anyone to download (in .rec format).</p> <p>And in June, shortly after <a href="https://twitter.com/adamhrv/status/1134511293526937600">posting</a> about the disappearance of the MS Celeb dataset, it reemerged on <a href="https://academictorrents.com/details/9e67eb7cc23c9417f39778a8e06cca5e26196a97/tech">Academic Torrents</a>. As of June 10, the MS Celeb dataset files have been redistributed in at least 9 countries and downloaded 44 times without any restrictions. The files were seeded and are mostly distributed by an AI company based in China called Hyper.ai, which states that it redistributes MS Celeb and other datasets for "teachers and students of service industry-related practitioners and research institutes."<a class="footnote_shim" name="[^hyperai_readme]_1"> </a><a href="#[^hyperai_readme]" class="footnote" title="Footnote 6">6</a></p> -<p>Earlier in 2019 images from the MS Celeb were also repackaged into another face dataset called <em>Racial Faces in the Wild (RFW)</em>. To create it, the RFW authors uploaded face images from the MS Celeb dataset to the Face++ API and used the inferred racial scores to segregate people into four subsets: Caucasian, Asian, Indian, and African each with 3,000 subjects. That dataset then appeared in a subsequent research project from researchers affiliated with IIIT-Delhi and IBM TJ Watson called <a href="https://arxiv.org/abs/1904.01219">Deep Learning for Face Recognition: Pride or Prejudiced?</a>, which aims to reduce bias but also inadvertently furthers racist language and ideologies that can not be repeated here.</p> -<p>The estimated racial scores for the MS Celeb face images used in the RFW dataset were computed using the Face++ API, which is owned by Megvii Inc, a company that has been repeatedly linked to the oppressive surveillance of Uighur Muslims in Xinjiang, China. According to posts from the <a href="https://chinai.substack.com/p/chinai-newsletter-11-companies-involved-in-expanding-chinas-public-security-apparatus-in-xinjiang">ChinAI Newsletter</a> and <a href="https://www.buzzfeednews.com/article/ryanmac/us-money-funding-facial-recognition-sensetime-megvii">BuzzFeedNews</a>, Megvii announced in 2017 at the China-Eurasia Security Expo in Ürümqi, Xinjiang, that it would be the official technical support unit of the "Public Security Video Laboratory" in Xinjiang, China. If they didn't already, it's highly likely that Megvii has a copy of everyone's biometric faceprint from the MS Celeb dataset, either from uploads to the Face++ API or through the research projects explicitly referencing MS Celeb dataset usage, such as a 2018 paper called <a href="https://arxiv.org/pdf/1808.06210.pdf">GridFace: Face Rectification via Learning Local Homography Transformations</a> jointly published by 3 authors, all of whom worked for Megvii.</p> +<p>Earlier in 2019 images from the MS Celeb were also repackaged into another face dataset called <em>Racial Faces in the Wild (RFW)</em>. To create it, the RFW authors uploaded face images from the MS Celeb dataset to the Face++ API and used the inferred racial scores to segregate people into four subsets: Caucasian, Asian, Indian, and African each with 3,000 subjects. That dataset then appeared in a subsequent research project from researchers affiliated with IIIT-Delhi and IBM TJ Watson called <a href="https://arxiv.org/abs/1904.01219">Deep Learning for Face Recognition: Pride or Prejudiced?</a>, which aims to reduce bias but also inadvertently furthers racist ideologies, using discredited racial terminology that cannot be repeated here.</p> +<p>The estimated racial scores for the MS Celeb face images used in the RFW dataset were computed using the Face++ API, which is owned by Megvii Inc, a company that has been repeatedly linked to the oppressive surveillance of Uighur Muslims in Xinjiang, China. According to posts from the <a href="https://chinai.substack.com/p/chinai-newsletter-11-companies-involved-in-expanding-chinas-public-security-apparatus-in-xinjiang">ChinAI Newsletter</a> and <a href="https://www.buzzfeednews.com/article/ryanmac/us-money-funding-facial-recognition-sensetime-megvii">BuzzFeedNews</a>, Megvii announced in 2017 at the China-Eurasia Security Expo in Ürümqi, Xinjiang, that it would be the official technical support unit of the "Public Security Video Laboratory" in Xinjiang, China. If they didn't already, it's highly likely that Megvii has a copy of everyone's biometric faceprint from the MS Celeb dataset, either from uploads to the Face++ API or through research projects explicitly referencing MS Celeb dataset usage, such as a 2018 paper called <a href="https://arxiv.org/pdf/1808.06210.pdf">GridFace: Face Rectification via Learning Local Homography Transformations</a> jointly published by 3 authors, all of whom worked for Megvii.</p> <h2>Commercial Usage</h2> <p>Microsoft's <a href="http://web.archive.org/web/20180218212120/http://www.msceleb.org/download/sampleset">MS Celeb website</a> says it was created for "non-commercial research purpose only." Publicly available research citations and competitions show otherwise.</p> <p>In 2017 Microsoft Research organized a face recognition competition at the International Conference on Computer Vision (ICCV), one of the top 2 computer vision conferences worldwide, where industry and academia used the MS Celeb dataset to compete for the highest performance scores. The 2017 winner was Beijing-based OrionStar Technology Co., Ltd.. In their <a href="https://www.prnewswire.com/news-releases/orionstar-wins-challenge-to-recognize-one-million-celebrity-faces-with-artificial-intelligence-300494265.html">press release</a>, OrionStar boasted a 13% increase on the difficult set over last year's winner. The prior year's competitors included Beijing-based Faceall Technology Co., Ltd., a company providing face recognition for "smart city" applications.</p> diff --git a/site/public/datasets/oxford_town_centre/index.html b/site/public/datasets/oxford_town_centre/index.html index 40f8bbc6..11fb436f 100644 --- a/site/public/datasets/oxford_town_centre/index.html +++ b/site/public/datasets/oxford_town_centre/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/uccs/assets/notes/index.html b/site/public/datasets/uccs/assets/notes/index.html index c8daf796..ce36f3d9 100644 --- a/site/public/datasets/uccs/assets/notes/index.html +++ b/site/public/datasets/uccs/assets/notes/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/datasets/uccs/index.html b/site/public/datasets/uccs/index.html index 96ab1e09..2dcf88a1 100644 --- a/site/public/datasets/uccs/index.html +++ b/site/public/datasets/uccs/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/datasets/who_goes_there/index.html b/site/public/datasets/who_goes_there/index.html index 3db77ff7..a00fd151 100644 --- a/site/public/datasets/who_goes_there/index.html +++ b/site/public/datasets/who_goes_there/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> diff --git a/site/public/index.html b/site/public/index.html index e5a6cd62..98b780b2 100644 --- a/site/public/index.html +++ b/site/public/index.html @@ -49,7 +49,7 @@ <div class='links'> <a href="/datasets/" class='aboutLink'>DATASETS</a> <a href="/about/" class='aboutLink'>ABOUT</a> - <a href="/about/news" class='updateLink'>News</a> + <a href="/research" class='updateLink'>Research</a> </div> </header> <div class="splash"> diff --git a/site/public/info/index.html b/site/public/info/index.html index f6280e58..51b4e5f8 100644 --- a/site/public/info/index.html +++ b/site/public/info/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/research/_from_1_to_100_pixels/index.html b/site/public/research/_from_1_to_100_pixels/index.html index 74f334cc..a978b264 100644 --- a/site/public/research/_from_1_to_100_pixels/index.html +++ b/site/public/research/_from_1_to_100_pixels/index.html @@ -50,27 +50,13 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> - <section> - <h1>From 1 to 100 Pixels</h1> - <div class='meta'> - <div> - <div class='gray'>Posted</div> - <div>2018-12-04</div> - </div> - <div> - <div class='gray'>By</div> - <div>Adam Harvey</div> - </div> - - </div> - </section> - - <section><h3>High resolution insights from low resolution data</h3> + <section><h1>From 1 to 100 Pixels</h1> +<h3>High resolution insights from low resolution data</h3> <p>This post will be about the meaning of "face". How do people define it? How to biometrics researchers define it? How has it changed during the last decade.</p> <p>What can you know from a very small amount of information?</p> <ul> diff --git a/site/public/research/_introduction/index.html b/site/public/research/_introduction/index.html index 66905247..8b17c016 100644 --- a/site/public/research/_introduction/index.html +++ b/site/public/research/_introduction/index.html @@ -50,27 +50,13 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> - <section> - <h1>Introducing MegaPixels</h1> - <div class='meta'> - <div> - <div class='gray'>Posted</div> - <div>2018-12-15</div> - </div> - <div> - <div class='gray'>By</div> - <div>Adam Harvey</div> - </div> - - </div> - </section> - - <section><p>Face recognition has become the focal point for ...</p> + <section><h1>Introduction</h1> +<p>Face recognition has become the focal point for ...</p> <p>Add 68pt landmarks animation</p> <p>But biometric currency is ...</p> <p>Add rotation 3D head</p> @@ -82,7 +68,7 @@ <li>Posted: Dec. 15</li> <li>Author: Adam Harvey</li> </ul> -</section><section class='applet_container'><div class='applet' data-payload='{"command": "load_file /site/research/00_introduction/assets/summary_countries_top.csv", "fields": ["country, Xcitations"]}'></div></section><section><p>Paragraph text to test css formatting. Paragraph text to test css formatting. Paragraph text to test css formatting. Paragraph text to test css formatting. Paragraph text to test css formatting.</p> +</section><section class='applet_container'><div class='applet' data-payload='{"command": "load_file /site/research/00_introduction/assets/summary_countries_top.csv", "fields": ["Headings: country, Xcitations"]}'></div></section><section><p>Paragraph text to test css formatting. Paragraph text to test css formatting. Paragraph text to test css formatting. Paragraph text to test css formatting. Paragraph text to test css formatting.</p> <p>[ page under development ]</p> </section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/_introduction/assets/test.png' alt=' This is the caption'><div class='caption'> This is the caption</div></div></section> diff --git a/site/public/research/_what_computers_can_see/index.html b/site/public/research/_what_computers_can_see/index.html index 003dd733..35f6d47d 100644 --- a/site/public/research/_what_computers_can_see/index.html +++ b/site/public/research/_what_computers_can_see/index.html @@ -50,27 +50,13 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> - <section> - <h1>What Computers Can See</h1> - <div class='meta'> - <div> - <div class='gray'>Posted</div> - <div>2018-12-15</div> - </div> - <div> - <div class='gray'>By</div> - <div>Adam Harvey</div> - </div> - - </div> - </section> - - <section><p>Rosalind Picard on Affective Computing Podcast with Lex Fridman</p> + <section><h1>What Computers Can See About Your Face</h1> +<p>Rosalind Picard on Affective Computing Podcast with Lex Fridman</p> <ul> <li>we can read with an ordinary camera on your phone, from a neutral face if</li> <li>your heart is racing</li> diff --git a/site/public/research/index.html b/site/public/research/index.html index 571b8230..f4f90531 100644 --- a/site/public/research/index.html +++ b/site/public/research/index.html @@ -50,13 +50,22 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> <section><h1>Research Blog</h1> -</section><div class='research_index'><a href='/research/_introduction/'><section class='wide'><img src='data:image/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==' alt='Research post' /><section><h1>Introducing MegaPixels</h1><h2></h2></section></section></a><a href='/research/munich_security_conference/'><section class='wide'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/background.jpg' alt='Research post' /><section><h1>Transnational Data Analysis of Publicly Available Face Recognition Training Datasets</h1><h2></h2></section></section></a></div> +</section><div class='research_index'> + <a href='/research/munich_security_conference/'><section class='wide' style='background-image: url(https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/background.jpg);' /> + <section> + <h4><span class='bgpad'>28 June 2019</span></h4> + <h2><span class='bgpad'>Analyzing Transnational Flows of Face Recognition Image Training Data</span></h2> + <h3><span class='bgpad'>Where does face data originate and who's using it?</span></h3> + <h4 class='readmore'><span class='bgpad'>Read more...</span></h4> + </section> + </section></a> + </div> </div> <footer> diff --git a/site/public/research/munich_security_conference/index.html b/site/public/research/munich_security_conference/index.html index 499d8e9f..b0503f84 100644 --- a/site/public/research/munich_security_conference/index.html +++ b/site/public/research/munich_security_conference/index.html @@ -4,7 +4,7 @@ <title>MegaPixels: MSC</title> <meta charset="utf-8" /> <meta name="author" content="Adam Harvey" /> - <meta name="description" content="Analyzing the Transnational Flow of Facial Recognition Data" /> + <meta name="description" content="Analyzing Transnational Flows of Face Recognition Image Training Data" /> <meta property="og:title" content="MegaPixels: MSC"/> <meta property="og:type" content="website"/> <meta property="og:summary" content="MegaPixels is an art and research project about face recognition datasets created \"in the wild\"/> @@ -50,31 +50,36 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-dataset"> - <section> - <h1>MSC</h1> - <div class='meta'> - <div> - <div class='gray'>Posted</div> - <div>2019-4-18</div> - </div> - <div> - <div class='gray'>By</div> - <div>Adam Harvey</div> - </div> - - </div> - </section> - - <section class='intro_section' style='background-image: url(https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/background.jpg)'><div class='inner'><div class='hero_desc'><span class='bgpad'>Analyzing the Transnational Flow of Facial Recognition Data</span></div><div class='hero_subdesc'><span class='bgpad'>Where does face data originate and who's using it? -</span></div></div></section><section><p>[page under devlopment]</p> -<p>Intro paragraph.</p> -<p>[ add montage of extracted faces here]</p> -</section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/montage_placeholder.jpg' alt=' Placeholder caption'><div class='caption'> Placeholder caption</div></div></section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/bar_placeholder.png' alt=' Placeholder caption'><div class='caption'> Placeholder caption</div></div></section><section><div class='columns columns-2'><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/pie_placeholder.png' alt=' Placeholder caption'><div class='caption'> Placeholder caption</div></div></section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/pie_placeholder.png' alt=' Placeholder caption'><div class='caption'> Placeholder caption</div></div></section></div></section><section> + <section class='intro_section' style='background-image: url(https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/background.jpg)'><div class='inner'><div class='hero_desc'><span class='bgpad'>Analyzing Transnational Flows of Face Recognition Image Training Data</span></div><div class='hero_subdesc'><span class='bgpad'>Where does face data originate and who's using it? +</span></div></div></section><section><h2>Face Datasets and Information Supply Chains</h2> +</section><section><div class='right-sidebar'><div class='meta'><div class='gray'>Images Analyzed</div><div>24,302,637</div></div><div class='meta'><div class='gray'>Datasets Analyzed</div><div>30</div></div><div class='meta'><div class='gray'>Years</div><div>2006 - 2018</div></div><div class='meta'><div class='gray'>Status</div><div>Ongoing Investigation</div></div><div class='meta'><div class='gray'>Last Updated</div><div>June 28, 2019</div></div></div><p>National AI strategies often rely on transnational data sources to capitalize on recent advancements in deep learning and neural networks. Researchers benefiting from these transnational data flows can yield quick and significant gains across diverse sectors from health care to biometrics. But new challenges emerge when national AI strategies collide with national interests.</p> +<p>Our earlier research on the <a href="/datsets">MS Celeb</a> and <a href="/datsets/duke_mtmc">Duke</a> datasets published with the Financial Times revealed that several computer vision image datasets created by US companies and universities were unexpectedly also used for research by the National University of Defense Technology in China, along with top Chinese surveillance firms including SenseTime, SenseNets, CloudWalk, Hikvision, and Megvii/Face++ which have all been linked to the oppressive surveillance of Uighur Muslims in Xinjiang.</p> +<p>In this new research for the Munich Security Conference's Transnational Security Report we provide summary statistics about the origins and endpoints of facial recognition information supply chains. To make it more personal, we gathered additional data on the number of public photos from Embassies that are currently being used in facial recognition datasets.</p> +<h3>24 Million Non-Cooperative Faces</h3> +<p>In total, we analyzed 30 publicly available face recognition and face analysis datasets that collectively include over 24 million non-cooperative images. Of these 24 million images, over 15 million face images are from Internet search engines, over 5.8 million from Flickr.com, over 2.5 million from the Internet Movie Database (IMDb.com), and nearly 500,000 from CCTV footage. All 24 million images were collected without any explicit consent, a type of face image researchers call "in the wild".</p> +<p>Next we manually verified 1,134 publicly available research papers that cite these datasets to determine who was using the data and where it was being used. Even though all of the images originated in the United States, the publicly available research citations show that only about 25% citations are from the country of the origin while the majority of citations are from China.</p> +</section><section><div class='columns columns-2'><section class='applet_container'><div class='applet' data-payload='{"command": "single_pie_chart /site/research/munich_security_conference/assets/megapixels_origins_top.csv", "fields": ["Caption: Sources of Publicly Available Non-Cooperative Face Image Training Data 2006 - 2018", "Top: 10", "OtherLabel: Other"]}'></div></section><section class='applet_container'><div class='applet' data-payload='{"command": "single_pie_chart /site/research/munich_security_conference/assets/summary_countries.csv", "fields": ["Caption: Locations Where Face Data Is Used Based on Public Research Citations", "Top: 14", "OtherLabel: Other"]}'></div></section></div></section><section><h3>6,000 Embassy Photos Being Used To Train Facial Recognition</h3> +<p>Of the 5.8 million Flickr images we found over 6,000 public photos from Embassy Flickr accounts were used to train facial recognition technologies. These images were used in the MegaFace, IBM Diversity in Faces datasets. Over 2,000 more images were used in the Who Goes There datasets used for facial ethnicity analysis research. A few of the embassy images found in facial recognition datasets are shown below.</p> +</section><section><div class='columns columns-2'><section class='applet_container'><div class='applet' data-payload='{"command": "single_pie_chart /site/research/munich_security_conference/assets/country_counts.csv", "fields": ["Caption: Photos from these embassies are being used to train face recognition software", "Top: 4", "OtherLabel: Other", "Colors: categoryRainbow"]}'></div></section><section class='applet_container'><div class='applet' data-payload='{"command": "single_pie_chart /site/research/munich_security_conference/assets/embassy_counts_summary_dataset.csv", "fields": ["Caption: Embassy images were found in these datasets", "Top: 4", "OtherLabel: Other", "Colors: categoryRainbow"]}'></div></section></div></section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/4606260362.jpg' alt=' An image in the MegaFace dataset obtained from United Kingdoms Embassy in Italy'><div class='caption'> An image in the MegaFace dataset obtained from United Kingdom's Embassy in Italy</div></div> +<div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/4749096858.jpg' alt=' An image in the MegaFace dataset obtained from the Flickr account of the United States Embassy in Kabul, Afghanistan'><div class='caption'> An image in the MegaFace dataset obtained from the Flickr account of the United States Embassy in Kabul, Afghanistan</div></div></section><section class='images'><div class='image'><img src='https://nyc3.digitaloceanspaces.com/megapixels/v1/site/research/munich_security_conference/assets/4730007024.jpg' alt=' An image in the MegaFace dataset obtained from U.S. Embassy Canberra'><div class='caption'> An image in the MegaFace dataset obtained from U.S. Embassy Canberra</div></div></section><section><p>This brief research aims to shed light on the emerging politics of data. A photo is no longer just a photo when it can also be surveillance training data, and datasets can no longer be separated from the development of software when software is now built with data. "Our relationship to computers has changed", says Geoffrey Hinton, one of the founders of modern day neural networks and deep learning. "Instead of programming them, we now show them and they figure it out."<a class="footnote_shim" name="[^hinton]_1"> </a><a href="#[^hinton]" class="footnote" title="Footnote 1">1</a>.</p> +<p>National AI strategies might also want to include transnational dataset strategies.</p> +<p><em>This research post is going and will updated during July and August, 2019.</em></p> +<h3>Further Reading</h3> +<ul> +<li><a href="/datasets/msceleb">MS Celeb Dataset Analysis</a></li> +<li><a href="/datasets/brainwash">Brainwash Dataset Analysis</a></li> +<li><a href="/datasets/duke_mtmc">Duke MTMC Dataset Analysis</a></li> +<li><a href="/datasets/uccs">Unconstrained College Students Dataset Analysis</a></li> +<li><a href="https://www.dukechronicle.com/article/2019/06/duke-university-facial-recognition-data-set-study-surveillance-video-students-china-uyghur">Duke MTMC dataset author apologies to students</a></li> +<li><a href="https://www.bbc.com/news/technology-48555149">BBC coverage of MS Celeb dataset takedown</a></li> +<li><a href="https://www.spiegel.de/netzwelt/web/microsoft-gesichtserkennung-datenbank-mit-zehn-millionen-fotos-geloescht-a-1271221.html">Spiegel coverage of MS Celeb dataset takdown</a></li> +</ul> +</section><section> <div class="hr-wave-holder"> <div class="hr-wave-line hr-wave-line1"></div> @@ -83,8 +88,7 @@ <h2>Supplementary Information</h2> -</section><section><p>[ add a download button for CSV data ]</p> -</section><section class='applet_container'><div class='applet' data-payload='{"command": "load_file /site/research/munich_security_conference/assets/embassy_counts_public.csv", "fields": ["Images, Dataset, Embassy, Flickr ID, URL, Guest, Host"]}'></div></section><section> +</section><section class='applet_container'><div class='applet' data-payload='{"command": "load_file /site/research/munich_security_conference/assets/embassy_counts_public.csv", "fields": ["Headings: Images, Dataset, Embassy, Flickr ID, URL, Guest, Host"]}'></div></section><section> <h4>Cite Our Work</h4> <p> @@ -101,7 +105,8 @@ }</pre> </p> -</section> +</section><section><h3>References</h3><section><ul class="footnotes"><li>1 <a name="[^hinton]" class="footnote_shim"></a><span class="backlinks"><a href="#[^hinton]_1">a</a></span>"Heroes of Deep Learning: Andrew Ng interviews Geoffrey Hinton". Published on Aug 8, 2017. <a href="https://www.youtube.com/watch?v=-eyhCTvrEtE">https://www.youtube.com/watch?v=-eyhCTvrEtE</a> +</li></ul></section></section> </div> <footer> diff --git a/site/public/test/chart/index.html b/site/public/test/chart/index.html index 33fafb48..e3134df9 100644 --- a/site/public/test/chart/index.html +++ b/site/public/test/chart/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/citations/index.html b/site/public/test/citations/index.html index a5fbcc76..3c630adc 100644 --- a/site/public/test/citations/index.html +++ b/site/public/test/citations/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/csv/index.html b/site/public/test/csv/index.html index d3ca0953..f1204c90 100644 --- a/site/public/test/csv/index.html +++ b/site/public/test/csv/index.html @@ -50,14 +50,14 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> <section><h1>CSV Test</h1> <h3><a href="/test/">← Back to test index</a></h3> -</section><section class='applet_container'><div class='applet' data-payload='{"command": "load_file /site/test/assets/test.csv", "fields": ["Name, Images, Year, Gender, Description, URL"]}'></div></section> +</section><section class='applet_container'><div class='applet' data-payload='{"command": "load_file /site/test/assets/test.csv", "fields": ["Headings: Name, Images, Year, Gender, Description, URL"]}'></div></section> </div> <footer> diff --git a/site/public/test/datasets/index.html b/site/public/test/datasets/index.html index 136fbd60..fccc5367 100644 --- a/site/public/test/datasets/index.html +++ b/site/public/test/datasets/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/face_search/index.html b/site/public/test/face_search/index.html index 3545bb00..52279498 100644 --- a/site/public/test/face_search/index.html +++ b/site/public/test/face_search/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/gallery/index.html b/site/public/test/gallery/index.html index b1061867..79adb78e 100644 --- a/site/public/test/gallery/index.html +++ b/site/public/test/gallery/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/index.html b/site/public/test/index.html index 626f1f0f..81e805ef 100644 --- a/site/public/test/index.html +++ b/site/public/test/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/map/index.html b/site/public/test/map/index.html index 6cba3b6f..4b4f2a4b 100644 --- a/site/public/test/map/index.html +++ b/site/public/test/map/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/name_search/index.html b/site/public/test/name_search/index.html index 1be553a5..50fc4a1b 100644 --- a/site/public/test/name_search/index.html +++ b/site/public/test/name_search/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/public/test/pie_chart/index.html b/site/public/test/pie_chart/index.html index 6b13aad6..4ba67b78 100644 --- a/site/public/test/pie_chart/index.html +++ b/site/public/test/pie_chart/index.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-"> diff --git a/site/templates/home.html b/site/templates/home.html index 19bdafe6..b74c0408 100644 --- a/site/templates/home.html +++ b/site/templates/home.html @@ -49,7 +49,7 @@ <div class='links'> <a href="/datasets/" class='aboutLink'>DATASETS</a> <a href="/about/" class='aboutLink'>ABOUT</a> - <a href="/about/news" class='updateLink'>News</a> + <a href="/research" class='updateLink'>Research</a> </div> </header> <div class="splash"> diff --git a/site/templates/layout.html b/site/templates/layout.html index b170b2b1..7c9052b2 100644 --- a/site/templates/layout.html +++ b/site/templates/layout.html @@ -50,7 +50,7 @@ <div class='links'> <a href="/datasets/">Datasets</a> <a href="/about/">About</a> - <a href="/about/news">News</a> + <a href="/research">Research</a> </div> </header> <div class="content content-{{ metadata.cssclass }}"> |