{"id":496,"date":"2020-02-02T11:16:00","date_gmt":"2020-02-02T03:16:00","guid":{"rendered":"http:\/\/www.91tfboys.com\/?p=496"},"modified":"2020-07-14T23:39:13","modified_gmt":"2020-07-14T15:39:13","slug":"%e5%9f%ba%e4%ba%8e%e6%b7%b1%e5%ba%a6%e5%ad%a6%e4%b9%a0%e7%9a%84%e7%bd%91%e7%ab%99%e6%8c%87%e7%ba%b9","status":"publish","type":"post","link":"https:\/\/www.91tfboys.com\/?p=496","title":{"rendered":"\u57fa\u4e8e\u6df1\u5ea6\u5b66\u4e60\u7684\u7f51\u7ad9\u6307\u7eb9"},"content":{"rendered":"\n

Tor\u901a\u8fc7\u5bf9\u5185\u5bb9\u7684\u52a0\u5bc6\u548c\u91cd\u91cd\u968f\u673a\u4ee3\u7406\uff0c\u4e3a\u4e92\u8054\u7f51\u7528\u6237\u63d0\u4f9b\u4e86\u533f\u540d\u670d\u52a1\uff0c\u80fd\u6709\u6548\u5bf9\u6297\u76d1\u5bdf\u8005\u7684\u5ba1\u67e5\u3002\u7136\u800c\uff0c\u7814\u7a76\u4eba\u5458\u53d1\u73b0\uff0c\u7531\u4e8eTor\u7684\u6d41\u91cf\u4e2d\u4ecd\u6709\u4e00\u4e9b\u7279\u5f81\u53ef\u4ee5\u8fdb\u884c\u8bc6\u522b\uff0c\u901a\u8fc7\u5bf9Tor\u7684WF\uff08Website Fingerprinting\uff0c\u7f51\u7ad9\u6307\u7eb9\uff09\u653b\u51fb\uff0c\u53ef\u4ee5\u5224\u65ad\u51fa\u7528\u6237\u6b63\u5728\u8bbf\u95ee\u7684\u7f51\u7ad9\u3002\u8fd1\u5e74\u6765\uff0c\u7814\u7a76\u4eba\u5458\u5148\u540e\u63d0\u51fa\u4f7f\u7528\u5305\u62eck<\/em>-NN[1]\u3001SVM[2]\u548c\u968f\u673a\u68ee\u6797\uff08Random Forest\uff0cRF\uff09[3]\u7b49\u4e0d\u540c\u673a\u5668\u5b66\u4e60\u7b97\u6cd5\u8fdb\u884c\u7f51\u7ad9\u6307\u7eb9\u7684\u65b9\u6cd5\uff0c\u5e76\u5728\u6d4b\u8bd5\u96c6\u4e0a\u53d6\u5f97\u4e8690%\u4ee5\u4e0a\u7684\u51c6\u786e\u7387\u3002<\/p>\n\n\n\n

\u4e0d\u8fc7\u8fd9\u4e9b\u65b9\u6cd5\u6ce8\u91cd\u4e8e\u7279\u5f81\u9009\u62e9\uff0c\u4ece\u800c\u5bf9\u6d41\u91cf\u7684\u7279\u5f81\u53d8\u5316\u8f83\u4e3a\u654f\u611f\uff0c\u800c\u4e14\u53ea\u5728\u5c01\u95ed\u73af\u5883\uff08Closed world\uff09\u4e0b\u8fdb\u884c\u6d4b\u8bd5\uff0c\u5728\u5b9e\u9645\u4f7f\u7528\u4e2d\u4ecd\u6709\u5c40\u9650\u6027\u3002\u4e3a\u6b64\uff0c\u4f5c\u8005\u63d0\u51fa\u4e86\u57fa\u4e8e\u6df1\u5ea6\u5b66\u4e60\u7684\u81ea\u52a8\u5316\u7f51\u7ad9\u6307\u7eb9\u65b9\u6cd5\uff0c\u5206\u522b\u91c7\u7528\u5806\u6808\u53bb\u566a\u81ea\u7f16\u7801\u5668\uff08Stacked Denoising Autoencoder\uff0cSDAE\uff09\u3001\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff08Convolutional Neural Network\uff0cCNN\uff09\u548c\u957f\u77ed\u671f\u8bb0\u5fc6\uff08Long-Short Term Memory\uff0cLSTM\uff09\u7f51\u7edc\uff0c\u901a\u8fc7\u5b66\u4e60\u539f\u59cb\u6d41\u91cf\u8fdb\u884c\u7279\u5f81\u63d0\u53d6\u6784\u5efa\u6307\u7eb9\uff0c\u8fdb\u800c\u5bf9\u7f51\u7ad9\u8fdb\u884c\u8bc6\u522b\u3002<\/p>\n\n\n\n

\"\"
\u56fe1 SDAE\u5206\u7c7b\u5668<\/figcaption><\/figure><\/div>\n\n\n\n
\"\"
\u56fe2 CNN\u5206\u7c7b\u5668<\/figcaption><\/figure><\/div>\n\n\n\n
\"\"<\/figure><\/div>\n\n\n\n
\"\"
\u56fe3 LSTM\u7f51\u7edc\u548cLSTM\u5355\u5143<\/figcaption><\/figure><\/div>\n\n\n\n

\u81ea\u7f16\u7801\u5668\uff08Autoencoder\uff0cAE\uff09\u662f\u4e00\u79cd\u524d\u9988\u7f51\u7edc\uff0c\u4e13\u95e8\u7528\u4e8e\u901a\u8fc7\u964d\u7ef4\u6765\u8fdb\u884c\u7279\u5f81\u5b66\u4e60\u3002\u5c06\u591a\u4e2aAE\u5806\u79ef\u5f62\u6210\u6df1\u5c42\u6a21\u578b\uff0c\u53ef\u4ee5\u5bf9\u8f93\u5165\u6570\u636e\u7684\u6700\u663e\u7740\u7279\u5f81\u8fdb\u884c\u5206\u5c42\u63d0\u53d6\uff0c\u5e76\u6839\u636e\u6d3e\u751f\u7279\u5f81\u8fdb\u884c\u5206\u7c7b\uff0c\u4f7f\u5f97SDAE\u6a21\u578b\u80fd\u591f\u7528\u4e8eWF\u3002CNN\u7531\u4e00\u7cfb\u5217\u5377\u79ef\u5c42\u6784\u6210\u3002\u5377\u79ef\u5c42\u4e5f\u7528\u4e8e\u7279\u5f81\u63d0\u53d6\uff0c\u4ece\u7b2c\u4e00\u5c42\u7684\u4f4e\u7ea7\u7279\u5f81\u5f00\u59cb\uff0c\u9010\u7ea7\u62bd\u8c61\u3002\u5377\u79ef\u5c42\u5b66\u4e60\u591a\u79cd\u8fc7\u6ee4\u5668\uff0c\u8fd9\u4e9b\u8fc7\u6ee4\u5668\u53ef\u4ee5\u663e\u793a\u51fa\u8f93\u5165\u6570\u636e\u4e2d\u5305\u542b\u7279\u5b9a\u7279\u5f81\u7684\u533a\u57df\uff0c\u7136\u540e\u5bf9\u8fd9\u4e9b\u8f93\u5165\u5b9e\u4f8b\u8fdb\u884c\u964d\u91c7\u6837\uff0c\u5e76\u4fdd\u7559\u7279\u6b8a\u533a\u57df\u3002\u8fd9\u6837\uff0cCNN\u53ef\u4ee5\u68c0\u7d22\u6700\u91cd\u8981\u7684\u7279\u5f81\u6765\u8fdb\u884c\u5206\u7c7b\u3002SDAE\u9700\u8981\u9010\u5757\u8fdb\u884c\u9884\u8bad\u7ec3\uff0c\u800cCNN\u4e0d\u9700\u8981\u5927\u91cf\u7684\u9884\u5904\u7406\u3002LSTM\u5206\u7c7b\u5668\u662f\u7279\u6b8a\u7c7b\u578b\u7684\u5faa\u73af\u795e\u7ecf\u7f51\u7edc\uff08Recurrent Neural Network\uff09\uff0c\u5177\u6709\u589e\u5f3a\u8bb0\u5fc6\u80fd\u529b\u3002LSTM\u7684\u8bbe\u8ba1\u5141\u8bb8\u5b66\u4e60\u6570\u636e\u4e2d\u7684\u957f\u671f\u4f9d\u8d56\u5173\u7cfb\uff0c\u4f7f\u5206\u7c7b\u5668\u80fd\u591f\u89e3\u91ca\u65f6\u95f4\u5e8f\u5217\u3002\u6b64\u5904\u8f93\u5165\u7684\u6d41\u91cf\u8f68\u8ff9\u57fa\u672c\u4e0a\u662fTor\u5355\u5143\u7684\u65f6\u95f4\u5e8f\u5217\uff0c\u8fd9\u4e9b\u52a8\u6001\u65f6\u95f4\u5e8f\u5217\u5f88\u53ef\u80fd\u5305\u542b\u7f51\u7ad9\u6307\u7eb9\u3002<\/p>\n\n\n\n

\u4f5c\u8005\u901a\u8fc7Tor\u548cTor\u6d4f\u89c8\u5668\u6536\u96c6\u4e86360\u4e07\u4e2a\u5bf9Alexa Top 1200\u7684\u8bbf\u95ee\u6570\u636e\uff0c\u5e76\u7ecf\u8fc7\u7b5b\u9009\uff0c\u5f62\u6210\u4e864\u4e2a\u4e3b\u8981\u7684\u6570\u636e\u96c6CW<\/em>100<\/sub>\u3001CW<\/em>200<\/sub>\u3001CW<\/em>500<\/sub>\u548cCW<\/em>900<\/sub>\uff0c\u5206\u522b\u5305\u542b\u4e86\u7b5b\u9009\u540e\u7684100\u4e2a\u3001200\u4e2a\u3001500\u4e2a\u548c900\u4e2a\u7f51\u7ad9\u7684\u54042500\u6b21\u8bbf\u95ee\u6570\u636e\uff0c\u4f5c\u4e3a\u5c01\u95ed\u73af\u5883\uff08Closed world\uff09\u7684\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6\u3002\u4f5c\u8005\u5bf9Alexa Top 400, 000\u7684\u7f51\u7ad9\u8fdb\u884c\u8bbf\u95ee\uff0c\u6536\u96c6400, 000\u6761\u6570\u636e\uff0c\u5bf9CW<\/em>200<\/sub>\u4e2d\u7684\u7f51\u7ad9\u5206\u522b\u8fdb\u884c\u4e862000\u6b21\u8bbf\u95ee\uff0c\u6536\u96c6400, 000\u6761\u6570\u636e\uff0c\u8fd9800, 000\u6761\u6570\u636e\u6784\u6210\u5f00\u653e\u73af\u5883\uff08Open world\uff09\u7684\u6570\u636e\u96c6\u3002<\/p>\n\n\n\n

\u4f5c\u8005\u7684\u6d4b\u8bd5\u7ed3\u679c\u5982\u56fe4\u548c\u56fe5\u3002\u7531\u4e8e\u968f\u7740\u65f6\u95f4\u7684\u53d8\u5316\uff0c\u7f51\u7ad9\u5185\u5bb9\u4f1a\u53d1\u751f\u53d8\u5316\uff0c\u4f7f\u7528\u5148\u524d\u6536\u96c6\u7684\u6570\u636e\u8bad\u7ec3\u51fa\u6765\u7684\u6a21\u578b\u4f1a\u9020\u6210\u51c6\u786e\u7387\u4e0b\u964d\uff0c\u56e0\u6b64\u4f5c\u8005\u53c8\u8fdb\u884c\u4e86\u65f6\u95f4\u4e0a\u7684\u5bf9\u6bd4\u5b9e\u9a8c\uff0c\u5982\u56fe6\u3002\u7efc\u5408\u4e0a\u8ff0\u5b9e\u9a8c\uff0c\u53ef\u4ee5\u770b\u51fa\uff0c\u6574\u4f53\u4e0aSDAE\u6a21\u578b\u8868\u73b0\u8f83\u597d\uff0c\u4e14\u6700\u4e3a\u7a33\u5b9a\u3002\u7531\u4e8eCNN\u9700\u8981\u66f4\u5c11\u7684\u53c2\u6570\uff0c\u5b83\u7684\u8bad\u7ec3\u662f\u6700\u5feb\u7684\uff0c\u4f46\u4e5f\u66f4\u5bb9\u6613\u8fc7\u62df\u5408\u3002LSTM\u6267\u884c\u901f\u5ea6\u6700\u6162\uff0c\u4f46\u7531\u4e8e\u5176\u7ed3\u6784\u7279\u6027\uff0c\u5b83\u5177\u5907\u6700\u4f73\u7684\u6cdb\u5316\u80fd\u529b\uff0c\u800c\u5176\u53cd\u5411\u4f20\u64ad\u7684\u7279\u6027\u4f1a\u9020\u6210\u5904\u7406\u957f\u65f6\u95f4\u7684\u6570\u636e\u65f6\u7684\u6027\u80fd\u4e0b\u964d\u3002<\/p>\n\n\n\n

\"\"
\u56fe4 \u5c01\u95ed\u73af\u5883\u4e0b\u7684\u6d4b\u8bd5\uff08\u7ecf\u8fc7\u8c03\u53c2\u540e\uff0c\u5bf9 CW<\/em>1<\/sub>00<\/sub> \u8fdb\u884c\u6d4b\u8bd5\uff0c\u83b7\u53d6\u7684\u968f\u5355\u4e2a\u7f51\u7ad9\u8bbf\u95ee\u6b21\u6570\uff08trace\uff09\u589e\u52a0\u800c\u53d8\u5316\u7684\u51c6\u786e\u7387\uff08Accuracy\uff09\u548c\u635f\u5931\uff08Loss\uff09\uff0c\u5176\u4e2d\u7531\u4e8eLSTM\u5728\u65f6\u95f4\u9650\u5236\u4e0b\u7684\u53cd\u5411\u4f20\u64ad\uff0cLSTM\u7684\u8bad\u7ec3\u8f93\u5165\u6570\u636e\u91cf\u662f\u6700\u5c0f\u7684\uff0c\u4e14\u4fdd\u6301\u4e0d\u53d8\uff09 <\/figcaption><\/figure><\/div>\n\n\n\n
\"\"
\u56fe5 \u5f00\u653e\u73af\u5883\u4e0b\u7684\u6d4b\u8bd5\uff08CW<\/em>200<\/sub>\u76842000\u6b21\u8bbf\u95ee\u6570\u636e\u548cAlexa Top 400 000\u7684\u8bbf\u95ee\u6570\u636e\uff09 <\/figcaption><\/figure><\/div>\n\n\n\n
\"\"
\u56fe6 \u65f6\u95f4\u7684\u589e\u52a0\u5bf9\u51c6\u786e\u7387\u7684\u5f71\u54cd\uff08\u6570\u636e\u96c6\u4e3aCW<\/em>200<\/sub>\uff0cCUMUL\u4e3a[2]\u4e2d\u63d0\u51fa\u7684\u65b9\u6cd5\uff0c\u4f5c\u4e3a\u5bf9\u6bd4\uff09 <\/figcaption><\/figure><\/div>\n\n\n\n

\u4e16\u754c\u4e0a\u6ca1\u6709\u7edd\u5bf9\u7684\u5b89\u5168\uff0c\u4f7f\u7528Tor\u4e5f\u4e0d\u80fd\u5e26\u6765\u5b8c\u5168\u7684\u533f\u540d\u3002\u968f\u7740\u65f6\u95f4\u548c\u6280\u672f\u7684\u53d1\u5c55\uff0c\u9488\u5bf9\u6df1\u5ea6\u5b66\u4e60\u7684\u5bf9\u6297\u4e5f\u7ec8\u5c06\u8fd0\u7528\u4e8eTor\u7f51\u7edc\u4e2d\u3002\u653b\u9632\u662f\u573a\u957f\u671f\u7684\u535a\u5f08\u6218\uff0c\u6b64\u8d77\u5f7c\u4f0f\uff0c\u6b64\u6d88\u5f7c\u957f\u3002\u800c\u5728\u653b\u9632\u7684\u5bf9\u6297\u4e2d\uff0c\u6280\u672f\u5c06\u59cb\u7ec8\u662f\u6838\u5fc3\uff0c\u5e76\u5f97\u4ee5\u4e0d\u65ad\u5347\u534e\u3002<\/p>\n\n\n\n

\u53c2\u8003\u6587\u732e\uff1a<\/p>\n\n\n\n

[1] T. Wang, X. Cai, R. Nithyanand, R. Johnson, and I. Goldberg, \u201cEffective Attacks and Provable Defenses for Website Fingerprinting,\u201d in USENIX Security Symposium. USENIX Association, 2014, pp. 143\u2013157.<\/p>\n\n\n\n

[2] A. Panchenko, F. Lanze, A. Zinnen, M. Henze, J. Pennekamp, K.Wehrle, T. Engel, \u201cWebsite fingerprinting at internet scale,\u201d in Network & Distributed System Security Symposium (NDSS). IEEE Computer Society, 2016, pp. 1\u201315.<\/p>\n\n\n\n

[3] J. Hayes and G. Danezis, \u201ck-fingerprinting: a Robust Scalable Website Fingerprinting Technique,\u201d in USENIX Security Symposium. USENIX Association, 2016, pp. 1\u201317.<\/p>\n\n\n\n

\uff08\u8bba\u6587\u539f\u6587\uff1aVera Rimmer\u3001Davy Preuveneers\u3001Marc Juarezx\u3001Tom Van Goethem\u3001Wouter Joosen. Automated Website Fingerprinting through Deep Learning. NDSS 2018. https:\/\/www.ndss-symposium.org\/ndss2018\/programme\/#03A\uff09<\/p>\n\n\n\n

\uff08\u672c\u6587\u5b8c\u6210\u4e8e2018\u5e744\u670826\u65e5\uff0c2018\u5e744\u670827\u65e5\u9996\u53d1\u4e8eArkTeam<\/a>\uff09 <\/p>\n","protected":false},"excerpt":{"rendered":"

\u7531\u4e8eTor\u7684\u6d41\u91cf\u4e2d\u4ecd\u6709\u4e00\u4e9b\u7279\u5f81\uff0c\u901a\u8fc7\u5bf9Tor\u7684\u7f51\u7ad9\u6307\u7eb9\u653b\u51fb\uff0c\u53ef\u4ee5\u5224\u65ad\u51fa\u7528\u6237\u6b63\u5728\u8bbf\u95ee\u7684\u7f51\u7ad9\u3002\u672c\u6587\u63d0\u51fa\u4e86\u57fa\u4e8e\u6df1\u5ea6\u5b66\u4e60\u7684\u81ea\u52a8\u5316\u7f51\u7ad9\u6307\u7eb9\u65b9\u6cd5\uff0c\u901a\u8fc7\u5b66\u4e60\u539f\u59cb\u6d41\u91cf\u8fdb\u884c\u7279\u5f81\u63d0\u53d6\u6784\u5efa\u6307\u7eb9\uff0c\u8fdb\u800c\u5bf9\u7f51\u7ad9\u8fdb\u884c\u8bc6\u522b\u3002<\/p>\n","protected":false},"author":1,"featured_media":498,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[14,35,34],"class_list":["post-496","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-running","tag-tor","tag-35","tag-34"],"_links":{"self":[{"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/posts\/496","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=496"}],"version-history":[{"count":4,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/posts\/496\/revisions"}],"predecessor-version":[{"id":570,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/posts\/496\/revisions\/570"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=\/wp\/v2\/media\/498"}],"wp:attachment":[{"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=496"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=496"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.91tfboys.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=496"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}