ترغب بنشر مسار تعليمي؟ اضغط هنا

Game Theoretic Malware Detection

133   0   0.0 ( 0 )
 نشر من قبل Revan MacQueen
 تاريخ النشر 2020
  مجال البحث الهندسة المعلوماتية
والبحث باللغة English




اسأل ChatGPT حول البحث

Large software platforms (e.g., mobile app stores, social media, email service providers) must ensure that files on their platform do not contain malicious code. Platform hosts use security tools to analyze those files for potential malware. However, given the expensive runtimes of tools coupled with the large number of exchanged files, platforms are not able to run all tools on every incoming file. Moreover, malicious parties look to find gaps in the coverage of the analysis tools, and exchange files containing malware that exploits these vulnerabilities. To address this problem, we present a novel approach that models the relationship between malicious parties and the security analyst as a leader-follower Stackelberg security game. To estimate the parameters of our model, we have combined the information from the VirusTotal dataset with the more detailed reports from the National Vulnerability Database. Compared to a set of natural baselines, we show that our model computes an optimal randomization over sets of available security analysis tools.



قيم البحث

اقرأ أيضاً

Off-chain protocols constitute one of the most promising approaches to solve the inherent scalability issue of blockchain technologies. The core idea is to let parties transact on-chain only once to establish a channel between them, leveraging later on the resulting channel paths to perform arbitrarily many peer-to-peer transactions off-chain. While significant progress has been made in terms of proof techniques for off-chain protocols, existing approaches do not capture the game-theoretic incentives at the core of their design, which led to overlooking significant attack vectors like the Wormhole attack in the past. This work introduces the first game-theoretic model that is expressive enough to reason about the security of off-chain protocols. We advocate the use of Extensive Form Games - EFGs and introduce two instances of EFGs to capture security properties of the closing and the routing of the Lightning Network. Specifically, we model the closing protocol, which relies on punishment mechanisms to disincentivize the uploading on-chain of old channel states, as well as the routing protocol, thereby formally characterizing the Wormhole attack, a vulnerability that undermines the fee-based incentive mechanism underlying the Lightning Network.
We consider pricing and selection with fading channels in a Stackelberg game framework. A channel server decides the channel prices and a client chooses which channel to use based on the remote estimation quality. We prove the existence of an optimal deterministic and Markovian policy for the client, and show that the optimal policies of both the server and the client have threshold structures when the time horizon is finite. Value iteration algorithm is applied to obtain the optimal solutions for both the server and client, and numerical simulations and examples are given to demonstrate the developed result.
97 - Lin Chen , Lei Xu , Zhimin Gao 2020
Recent advances in the blockchain research have been made in two important directions. One is refined resilience analysis utilizing game theory to study the consequences of selfish behaviors of users (miners), and the other is the extension from a li near (chain) structure to a non-linear (graphical) structure for performance improvements, such as IOTA and Graphcoin. The first question that comes to peoples minds is what improvements that a blockchain system would see by leveraging these new advances. In this paper, we consider three major metrics for a blockchain system: full verification, scalability, and finality-duration. We { establish a formal framework and} prove that no blockchain system can achieve full verification, high scalability, and low finality-duration simultaneously. We observe that classical blockchain systems like Bitcoin achieves full verification and low finality-duration, Harmony and Ethereum 2.0 achieve low finality-duration and high scalability. As a complementary, we design a non-linear blockchain system that achieves full verification and scalability. We also establish, for the first time, the trade-off between scalability and finality-duration.
To achieve general intelligence, agents must learn how to interact with others in a shared environment: this is the challenge of multiagent reinforcement learning (MARL). The simplest form is independent reinforcement learning (InRL), where each agen t treats its experience as part of its (non-stationary) environment. In this paper, we first observe that policies learned using InRL can overfit to the other agents policies during training, failing to sufficiently generalize during execution. We introduce a new metric, joint-policy correlation, to quantify this effect. We describe an algorithm for general MARL, based on approximate best responses to mixtures of policies generated using deep reinforcement learning, and empirical game-theoretic analysis to compute meta-strategies for policy selection. The algorithm generalizes previous ones such as InRL, iterated best response, double oracle, and fictitious play. Then, we present a scalable implementation which reduces the memory requirement using decoupled meta-solvers. Finally, we demonstrate the generality of the resulting policies in two partially observable settings: gridworld coordination games and poker.
The spreading dynamics of an epidemic and the collective behavioral pattern of the population over which it spreads are deeply intertwined and the latter can critically shape the outcome of the former. Motivated by this, we design a parsimonious game -theoretic behavioral--epidemic model, in which an interplay of realistic factors shapes the co-evolution of individual decision-making and epidemics on a network. Although such a co-evolution is deeply intertwined in the real-world, existing models schematize population behavior as instantaneously reactive, thus being unable to capture human behavior in the long term. Our model offers a unified framework to model and predict complex emergent phenomena, including successful collective responses, periodic oscillations, and resurgent epidemic outbreaks. The framework also allows to assess the effectiveness of different policy interventions on ensuring a collective response that successfully eradicates the outbreak. Two case studies, inspired by real-world diseases, are presented to illustrate the potentialities of the proposed model.
التعليقات
جاري جلب التعليقات جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
mircosoft-partner

هل ترغب بارسال اشعارات عن اخر التحديثات في شمرا-اكاديميا