Jarvis: Moving Towards a Smarter Internet of Things

Mudgerikar, Anand; Bertino, Elisa

doi:10.1109/icdcs47774.2020.00020

Cited by 7 publications

(8 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors of [4] use Q-Learning to build a task-recommendation service for daily activities in Smart Homes. In [3], the authors use Deep Reinforcement Learning to optimize temperature, used energy, and cost. The main difference between the aforementioned works and Q-SMASH is that the human values are respected via the value-reasoning process of Q-SMASH.…”

Section: Related Work and Discussionmentioning

confidence: 99%

Q-SMASH: Q-Learning-based Self-Adaptation of Human-Centered Internet of Things

Rahimi¹,

Trentin²,

Ramparany³

et al. 2021

Preprint

View full text Add to dashboard Cite

Section: Related Work and Discussionmentioning

confidence: 99%

Q-SMASH: Q-Learning-based Self-Adaptation of Human-Centered Internet of Things

Rahimi¹,

Trentin²,

Ramparany³

et al. 2021

Preprint

View full text Add to dashboard Cite

“…Jarvis is a Reinforcement Learning technique for predicting the optimal and safe IoT systems decisions [66]. Jarvis models the simulated IoT environment based on the states and actions.…”

Section: Hybrid Analysismentioning

confidence: 99%

“…ForeSee correctness properties specified as Linear Temporal Logic (LTL) formulas [44]. While Jarvis represents safe states and transitions in defined formulas [66]. On the other hand, FriendlyFire uses calculus in representing the properties [72].…”

Section: Formal and Mathematical Representationsmentioning

confidence: 99%

“…On the other hand, Jarvis relies on safe states to be able to run Reinforcement Learning (RL) to compare the the normal behavior vs the safe behavior [66]. IoTPraetor uses the Device Usage Rules in order to compare the device behavior with these predefined rules [58].…”

Section: Other Representationsmentioning

confidence: 99%

“…Some Literature tools consider applying different Artificial Intelligence (AI) and Machine Learning (ML) techniques to improve the analysis of IoT systems. For example, Jarvis [66] uses Reinforcement Learning, Deep Q learning network, along with deep neural network to train the agent in order to improve the detection and achieve the best quality of the results. IoTMon [42], iRuler [48], and IoT-Praetor [58], uses Natural Language Processing (NLP) in their IoT apps analysis.…”

Section: Optimization Techniquesmentioning

confidence: 99%

See 2 more Smart Citations

A Survey of Analysis Methods for Security and Safety verification in IoT Systems

Abuserrieh¹,

Alalfi²

2022

Preprint

View full text Add to dashboard Cite

Internet of Things (IoT) has been rapidly growing in the past few years in all life disciplines. IoT provides automation and smart control to its users in different domains such as home automation, healthcare systems, automotive, and many more. Given the tremendous number of connected IoT devices, this growth leads to enormous automatic interactions among sizeable IoT apps in their environment, making IoT apps more intelligent and more enjoyable to their users. But some unforeseen interactions of IoT apps and any potential malicious behaviour can seriously cause insecure and unsafe consequences to its users, primarily non-experts, who lack the required knowledge regarding the potential impact of their IoT automation processes. In this paper, we study the problem of security and safety verification of IoT systems. We survey techniques that utilize program analysis to verify IoT applications' security and safety properties. The study proposes a set of categorization and classification attributes to enhance our understanding of the research landscape in this domain. Moreover, we discuss the main challenges considered in the surveyed work and potential solutions that could be adopted to ensure the security and safety of IoT systems.

show abstract

A Security-Constrained Reinforcement Learning Framework for Software Defined Networks

Mudgerikar

Bertino

Lobo

et al. 2021

ICC 2021 - IEEE International Conference on Communications

Self Cite

View full text Add to dashboard Cite

Reinforcement Learning (RL) is an effective technique for building 'smart' SDN controllers because of its modelfree nature and ability to learn policies online without requiring extensive training data. However, as RL agents are geared to maximize functionality and explore the environment without constraints, security can be breached. In this paper, we propose Jarvis-SDN, a RL framework that constrains explorations by taking security into account. In Jarvis-SDN, the RL agent learns 'intelligent policies' which maximize functionality but not at the cost of security. Standard network flow based attack signatures obtained from intrusion detection system (IDS) datasets cannot be used as policies because they do not conform to the state model of the RL framework and thus have poor accuracy and high false positives. To address such issue, the security policies for constraining explorations in Jarvis-SDN are learnt in a semi-supervised manner in the form of 'partial attack signatures' from packet captures of IDS datasets that are then encoded in the objective function of the RL based optimization framework. These signatures are learnt using Deep Q-Networks (DQN). Our analysis shows that DQN based attack signatures perform better than classical machine learning techniques, like decision trees, random forests and deep neural networks (DNN), for common network attacks. We instantiate our framework for a SDN controller with the goal of intelligent rate control to further analyze the effectiveness of the attack signatures.

show abstract

Jarvis: Moving Towards a Smarter Internet of Things

Cited by 7 publications

References 17 publications

Q-SMASH: Q-Learning-based Self-Adaptation of Human-Centered Internet of Things

Q-SMASH: Q-Learning-based Self-Adaptation of Human-Centered Internet of Things

A Survey of Analysis Methods for Security and Safety verification in IoT Systems

A Security-Constrained Reinforcement Learning Framework for Software Defined Networks

Contact Info

Product

Resources

About