\u5185\u5bb9\u6458\u8981\u53ca\u5609\u5bbe\u4ecb\u7ecd<\/b><\/strong><\/p>\n Dawn Song\u6559\u6388\uff0c\u52a0\u5dde\u5927\u5b66\u4f2f\u514b\u5229\u5206\u6821<\/p>\n \u9898\u76ee: AI and Security: lessons, challenges and future directions<\/strong><\/p>\n \u6458\u8981:<\/strong><\/p>\n In this talk, I will talk about challenges and exciting new opportunities at the intersection of AI and Security, how AI and deep learning can enable better security, and how Security can enable better AI. In particular, I will talk about secure deep learning and challenges and approaches to ensure the integrity of decisions made by deep learning. I will also give an overview on our work on using program analysis and transformation techniques to enable privacy-preserving data analytics and machine learning. I will also talk about how we can use deep learning for vulnerability detection. Finally, I will conclude with future directions at the intersection of AI and Security.<\/p>\n \u4e2a\u4eba\u4ecb\u7ecd:<\/strong><\/p>\n Dawn Song is a Professor in the Department of Electrical Engineering and Computer Science at UC Berkeley. Her research interest lies in deep learning and security. She has studied diverse security and privacy issues in computer systems and networks, including areas ranging from software security, networking security, database security, distributed systems security, applied cryptography, to the intersection of machine learning and security. She is the recipient of various awards including the MacArthur Fellowship, the Guggenheim Fellowship, the NSF CAREER Award, the Alfred P. Sloan Research Fellowship, the MIT Technology Review TR-35 Award, the George Tallman Ladd Research Award, the Okawa Foundation Research Award, the Li Ka Shing Foundation Women in Science Distinguished Lecture Series Award, the Faculty Research Award from IBM, Google and other major tech companies, and Best Paper Awards from top conferences. She obtained her Ph.D. degree from UC Berkeley. Prior to joining UC Berkeley as a faculty, she was a faculty at Carnegie Mellon University from 2002 to 2007.<\/p>\n Chengyu Song\u535a\u58eb\uff0c\u52a0\u5dde\u5927\u5b66\u6cb3\u6ee8\u5206\u6821<\/p>\n \u9898\u76ee:\u00a0Efficient Protection of Path-Sensitive Control Security<\/strong><\/p>\n \u6458\u8981:<\/strong><\/p>\n Control-Flow Integrity (CFI), as a means to prevent control-flow\u00a0hijacking attacks, enforces that each instruction transfers control to\u00a0an address in a set of valid targets.\u00a0 The security guarantee of CFI\u00a0thus depends on the definition of valid targets, which conventionally\u00a0are defined as the result of a static analysis.\u00a0 Unfortunately, previous\u00a0research has demonstrated that such a definition, and thus any\u00a0implementation that enforces it, still allows practical control-flow\u00a0attacks.<\/p>\n In this work, we present a path-sensitive variation of CFI that utilizes\u00a0runtime path-sensitive point-to analysis to compute the legitimate\u00a0control transfer targets.\u00a0 We have designed and implemented a runtime\u00a0environment, PittyPat, that enforces path-sensitive CFI efficiently by\u00a0combining commodity, low-overhead hardware monitoring and a novel\u00a0runtime points-to analysis.\u00a0 Our formal analysis and empirical\u00a0evaluation demonstrate that, compared to CFI based on static analysis,\u00a0PittyPat ensures that applications satisfy stronger security guarantees,\u00a0with acceptable overhead for security-critical contexts.<\/p>\n \u4e2a\u4eba\u4ecb\u7ecd:<\/strong><\/p>\n Chengyu Song is an Assistant Professor at the Computer Science and\u00a0Engineering department of UC Riverside. His primary research interests\u00a0are system and software security. He received his PhD from Georgia Tech\u00a0supervised by Prof. Wenke Lee and Prof. Taesoo Kim. He has published\u00a014 papers in top security and system conferences. One of his co-authored\u00a0paper won the 2015 Internet defense prize ($100k) and another one won\u00a0the CSAW’15 best applied security research paper. He received his M.Eng.\u00a0and B.S. from Peking University in 2010 and 2007.<\/p>\n \u8f6f\u4ef6\u667a\u80fd\u5206\u6790\u534f\u540c\u521b\u65b0\u56e2\u961f\u4ecb\u7ecd<\/strong><\/p>\n \u4e2d\u56fd\u79d1\u5b66\u9662\u8f6f\u4ef6\u7814\u7a76\u6240\u4e8e2016\u5e7412\u6708\uff0c\u7ecf\u5b66\u672f\u59d4\u5458\u4f1a\u8bba\u8bc1\u6210\u7acb\u4e86\u201c\u8f6f\u4ef6\u667a\u80fd\u5206\u6790\u201d\u534f\u540c\u521b\u65b0\u56e2\u961f\uff08Collaborative Research Group on Intelligent Software Analysis\uff0cISA Group\uff09\uff0c\u8be5\u56e2\u961f\u8054\u5408\u4e86\u6240\u5185\u53ef\u4fe1\u8ba1\u7b97\u4e0e\u4fe1\u606f\u4fdd\u969c\u5b9e\u9a8c\u5ba4\u3001\u8ba1\u7b97\u673a\u79d1\u5b66\u56fd\u5bb6\u91cd\u70b9\u5b9e\u9a8c\u5ba4\u7b49\u591a\u4e2a\u90e8\u95e8\u7684\u8f6f\u4ef6\u5206\u6790\u3001\u8f6f\u4ef6\u57fa\u7840\u7406\u8bba\u3001\u5927\u6570\u636e\u5206\u6790\u7b49\u65b9\u9762\u7684\u79d1\u7814\u4eba\u5458\uff0c\u662f\u4e00\u4e2a\u8de8\u90e8\u95e8\u7684\u534f\u540c\u521b\u65b0\u961f\u4f0d\u3002\u8be5\u56e2\u961f\u5c06\u9762\u5411\u8f6f\u4ef6\u6df1\u5ea6\u5206\u6790\u9700\u6c42\uff0c\u63a2\u7d22\u65b0\u7684\u65b9\u6cd5\u548c\u6280\u672f\uff0c\u63d0\u5347\u8f6f\u4ef6\u5206\u6790\u80fd\u529b\u3002<\/p>\n \u8be5\u56e2\u961f\u4ee5\u82cf\u749e\u777f\u7814\u7a76\u5458\u8bfe\u9898\u7ec4\u4e3a\u4e3b\u4f53\uff0c\u8054\u5408\u4e86\u8ba1\u7b97\u673a\u79d1\u5b66\u56fd\u5bb6\u91cd\u70b9\u5ba4\u8521\u5f66\u526f\u7814\u7a76\u5458\u8bfe\u9898\u7ec4\u3001\u65f6\u78ca\u526f\u7814\u7a76\u5458\u8bfe\u9898\u7ec4\uff0c\u5e76\u8058\u8bf7\u4e86\u6e05\u534e\u5927\u5b66\u5f20\u8d85\u526f\u6559\u6388\uff0c\u674e\u7426\u526f\u6559\u6388\u4f5c\u4e3a\u5ba2\u5ea7\u7814\u7a76\u5458\uff0c\u5171\u6709\u7814\u7a76\u4eba\u545830\u4f59\u4eba\u3002\u56e2\u961f\u5728\u8f6f\u4ef6\u52a8\u6001\u9006\u5411\u5206\u6790\u3001\u6076\u610f\u8f6f\u4ef6\u6df1\u5ea6\u5206\u6790\u3001\u8f6f\u4ef6\u6f0f\u6d1e\u5206\u6790\u4e0e\u5229\u7528\u3001\u8f6f\u4ef6\u5206\u6790\u4e0e\u6d4b\u8bd5\u3001\u53ef\u89c6\u5316\u5206\u6790\u4e0e\u7a0b\u5e8f\u5f02\u5e38\u68c0\u6d4b\u7b49\u5df2\u65b9\u9762\u5f62\u6210\u4e86\u4e30\u5bcc\u7684\u6280\u672f\u79ef\u7d2f\u3002<\/p>\n \u201c\u8f6f\u4ef6\u667a\u80fd\u5206\u6790\u201d\u5b66\u672f\u6c99\u9f99\u662f\u7531\u8f6f\u4ef6\u667a\u80fd\u5206\u6790\u534f\u540c\u521b\u65b0\u56e2\u961f\u7ec4\u7ec7\u7684\u5b66\u672f\u4ea4\u6d41\u5e73\u53f0\uff0c\u5c06\u9080\u8bf7\u56fd\u5185\u5916\u5b66\u8005\u5171\u540c\u7814\u8ba8\u8f6f\u4ef6\u5206\u6790\u9762\u4e34\u7684\u95ee\u9898\u3001\u6700\u65b0\u7684\u8fdb\u5c55\u3001\u7814\u7a76\u7684\u601d\u8def\u7b49\uff0c\u4fc3\u8fdb\u8f6f\u4ef6\u667a\u80fd\u5206\u6790\u5b66\u672f\u4ea4\u6d41\u3002<\/p>\n \u5173\u6ce8\u56e2\u961f\u79d1\u7814\u52a8\u6001\uff0c\u8bf7\u626b\u63cf\u9644\u540e\u4e8c\u7ef4\u7801\u3002<\/p>\n \u201c\u8f6f\u4ef6\u667a\u80fd\u5206\u6790\u201d\u5b66\u672f\u6c99\u9f99\u7b2c\u4e09\u6b21\u6d3b\u52a8\u5c06\u4e8e2017\u5e748\u670821\u65e5\uff08\u661f\u671f\u4e00\uff09\u4e0b\u5348\u4e24\u70b9\u534a\u5728\u4e2d\u56fd\u79d1\u5b66\u9662\u8f6f\u4ef6\u7814\u7a76\u6240… Continue reading ![\"\"](\"https:\/\/www.inforsec.org\/wp\/wp-content\/uploads\/2017\/08\/\u56fe\u72471.png\")
<\/p>\n![\"\"](\"https:\/\/www.inforsec.org\/wp\/wp-content\/uploads\/2017\/08\/\u56fe\u72472.png\")
<\/p>\n![\"\"](\"https:\/\/www.inforsec.org\/wp\/wp-content\/uploads\/2017\/08\/\u56fe\u72473.png\")
<\/p>\n![\"\"](\"https:\/\/www.inforsec.org\/wp\/wp-content\/uploads\/2017\/08\/\u56fe\u72474.png\")
<\/p>\n","protected":false},"excerpt":{"rendered":"