聯(lián)系人：李女士
座機(jī)：010-59433750
手機(jī)：15313878231
郵 箱：angelals@secspace.com
地 址：北京市海淀區(qū)金碼大廈B座2328
【InForSec@復(fù)旦大學(xué)】德克薩斯大學(xué)達(dá)拉斯分校助理教授楊威與您相約復(fù)旦大學(xué)！
While machine learning-based techniques have been widely applied in security domains, being able to explain the rationale behind their decision making process remains as a largely open problem. Recent techniques on interpreting decision making of neural networks either provide local explanation for each input instance or approximate the original model based on a set of input-output instances.
雖然基于機(jī)器學(xué)習(xí)的技術(shù)已廣泛應(yīng)用于安全領(lǐng)域，但能夠解釋其決策過程背后的基本原理仍然是一個(gè)基本上是開放的問題。解釋神經(jīng)網(wǎng)絡(luò)決策的最新技術(shù)或者為每個(gè)輸入實(shí)例提供局部解釋，或者基于一組輸入 - 輸出實(shí)例來近似原始模型
The quality of explanation provided by these techniques is limited by the scope of inputs used to generate approximated models or explanations. However, the inherent nature of security research requires us to understand the intrinsic characteristics of a neural network model instead of just parts of model behaviors.
這些技術(shù)提供的解釋質(zhì)量受到用于生成近似模型或解釋的輸入范圍的限制。然而，安全研究的固有性質(zhì)要求我們理解神經(jīng)網(wǎng)絡(luò)模型的內(nèi)在特征，而不僅僅是模型行為的一部分。
In this talk, I will first introduce REINAM as an example of applying machine learning technique for security research purpose. REINAM is a reinforcement-learning approach for synthesizing probabilistic context-free program input grammars without any seed inputs.
在本次演講中，我將首先介紹REINAM作為將機(jī)器學(xué)習(xí)技術(shù)應(yīng)用于安全研究目的的一個(gè)例子。REINAM是一種強(qiáng)化學(xué)習(xí)方法，用于在沒有任何種子輸入的情況下合成概率無上下文程序輸入語(yǔ)法。
Then, I will introduce DENAS, a novel input-independent neural-network explanation approach dedicated for security applications. DENAS is capable of efficiently generating decision rules which could interpret the decision making of a neural network without providing any input. Finally, I will briefly introduce iRuler, an IoT analysis framework that leverages Satisfiability Modulo Theories (SMT) solving and model checking to discover inter-rule vulnerabilities.
然后，我將介紹DENAS，一種專用于安全應(yīng)用的新穎的獨(dú)立于輸入的神經(jīng)網(wǎng)絡(luò)解釋方法。DENAS能夠有效地生成決策規(guī)則，該規(guī)則可以解釋神經(jīng)網(wǎng)絡(luò)的決策而無需提供任何輸入。最后，我將簡(jiǎn)要介紹iRuler，這是一個(gè)利用可滿足模數(shù)理論（SMT）求解和模型檢查來發(fā)現(xiàn)規(guī)則間漏洞的物聯(lián)網(wǎng)分析框架。
Wei Yang is an assistant professor in the Department of Computer Science at the University of Texas at Dallas.
Wei Yang是德克薩斯大學(xué)達(dá)拉斯分校計(jì)算機(jī)科學(xué)系的助理教授。
He received his Ph.D. in Computer Science from the University of Illinois at Urbana-Champaign in 2018, advised by Prof. Carl A. Gunter and Prof. Tao Xie, an M.S. in Computer Science from North Carolina State University in 2013, advised by Prof. Tao Xie, and a B.E. in Software Engineering from Shanghai Jiao Tong University in 2011, advised by Prof. Jianjun Zhao. He was a visiting researcher in University of California, Berkeley in 2017, invited by Prof. Dawn Song.
他獲得了博士學(xué)位。他于2018年在伊利諾伊大學(xué)厄巴納 - 香檳分校獲得計(jì)算機(jī)科學(xué)專業(yè)，由Carl A. Gunter教授和謝教授教授。2013年獲得北卡羅來納州立大學(xué)計(jì)算機(jī)科學(xué)學(xué)士學(xué)位，陶燮教授和B.E. 2011年，上海交通大學(xué)軟件工程專業(yè)，趙建軍教授。他是2017年加州大學(xué)伯克利分校的訪問研究員，受到Dawn Song教授的邀請(qǐng)。
【文章轉(zhuǎn)載自：網(wǎng)安國(guó)際】