Fingerprinting Skills on Smart Speakers using Machine Learning
Main Author: | |
---|---|
Language: | English |
Published: |
University of Cincinnati / OhioLINK
2020
|
Subjects: | |
Online Access: | http://rave.ohiolink.edu/etdc/view?acc_num=ucin158400122512448 |
id |
ndltd-OhioLink-oai-etd.ohiolink.edu-ucin158400122512448 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-OhioLink-oai-etd.ohiolink.edu-ucin1584001225124482021-08-03T07:13:57Z Fingerprinting Skills on Smart Speakers using Machine Learning Naraparaju, Shriti Computer Science Fingerprinting Skills The popularity of smart speakers with virtual personal assistants such as Amazon Alexa and Google Assistant has been growing. However, the security and privacy of users using these devices have not been thoroughly examined which can lead to severe security and privacy concerns. In this thesis, we perform fingerprinting attack of skills on Amazon Echo by analyzing the encrypted network traffic and show the potential privacy leakage of users using skills on smart speakers. Skills are third-party applications that can be accessed through Amazon Alexa. Amazon Echo, the most dominant smart speaker in the current market provides the feature of Skills & Games which can be accessed through the smart speaker. For example, by enabling the BBC skill, you can hear the live streaming news from BBC on your smart speaker. Skills increase the diversity of using Amazon smart speaker by providing a variety of applications that can be used from a single platform just with the use of voice commands without accessing phone or a computer. In this fingerprinting attack, an adversary can eavesdrop on the network traffic of the smart speaker and predict the skill that is used by the user without decrypting the network traffic. For this attack, we collect the dataset consisting of 10,000 traces of network traffic related to 100 popular skills. We perform feature selection to investigate the top features that contribute towards the users privacy leakage. We implement the attack by leveraging popular machine learning models which have been used in previous network traffic analysis works. Among the different machine learning models we implement, Random Forest Classifier performs the best by achieving an accuracy of 69.85%. The experimental results show that there is a need for developers to focus on the privacy leakages of users. Our results can also be used to detect malicious attacks on smart speakers and can contribute to encrypted traffic analysis and privacy preserving at large. 2020-06-15 English text University of Cincinnati / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=ucin158400122512448 http://rave.ohiolink.edu/etdc/view?acc_num=ucin158400122512448 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws. |
collection |
NDLTD |
language |
English |
sources |
NDLTD |
topic |
Computer Science Fingerprinting Skills |
spellingShingle |
Computer Science Fingerprinting Skills Naraparaju, Shriti Fingerprinting Skills on Smart Speakers using Machine Learning |
author |
Naraparaju, Shriti |
author_facet |
Naraparaju, Shriti |
author_sort |
Naraparaju, Shriti |
title |
Fingerprinting Skills on Smart Speakers using Machine Learning |
title_short |
Fingerprinting Skills on Smart Speakers using Machine Learning |
title_full |
Fingerprinting Skills on Smart Speakers using Machine Learning |
title_fullStr |
Fingerprinting Skills on Smart Speakers using Machine Learning |
title_full_unstemmed |
Fingerprinting Skills on Smart Speakers using Machine Learning |
title_sort |
fingerprinting skills on smart speakers using machine learning |
publisher |
University of Cincinnati / OhioLINK |
publishDate |
2020 |
url |
http://rave.ohiolink.edu/etdc/view?acc_num=ucin158400122512448 |
work_keys_str_mv |
AT naraparajushriti fingerprintingskillsonsmartspeakersusingmachinelearning |
_version_ |
1719457065004433408 |