How to collaborate

Applications for SPRING 2024 ARE CLOSED (except for Project 3).

We are excited to announce the list of collaboration opportunities in our Mon(IoT)r IoT lab. These can be  interesting opportunities to get exposed to cutting-edge IoT technologies and understand how they work.

If you are interested in any of these projects, you are a current active student at Northeastern University, and you satisfy the prerequisites for the project you are interested in, please send an email to moniotr@ccs.neu.edu with subject “Spring 2024 Mon(IoT)r Lab collaboration” and a recent resume attached (with GPA), specifying the following:

  • What project you are interested in, specifying (a) why you are interested in that particular project, (b) why you are a fit for that project, (c) how you plan to use your existing experience to contribute to that project, (d) how collaborating to the project aligns with your career goals. If you are interested in more than one project, please rank them starting from the one you are interested the most.
  • The preferred start date and end date for the collaboration, and the total number of average hours you plan to spend per week on this project;
  • Your expected course load for the semester (list of classes and credits);
  • Any other time commitments you have during the semester, for example TA, RA, a co-op, side jobs, or on campus / off campus volunteering activity;
  • Your availability for a volunteer position (we currently do not have any available paid or for-credits positions);
  • The possibility or preference to work on site, remotely, or both.

Please note that, in general, we have a preference for projects that are on-site, that last the whole semester (15 weeks), and for a minimum of 10 hours per week.

We will start reviewing applications for Spring 2024 on January 8, 2024 and keep the projects open until the positions are filled. After we start reviewing applications, it may take us up to two weeks to get back to you, so please be patient if you do not hear back by then.

If you are interested to apply for Summer 2024 or later, we cannot guarantee that this list of projects will still be valid, and therefore we suggest to wait for the projects to be updated before applying (usually two weeks before the semester starts).

Available projects

Project 1. Internet of Things analysis from network traffic

Project 1A is no longer available. Project 1B will be available again in Summer 2024.

Internet of Things (IoT) devices are increasingly found in homes, providing useful functionality for devices such as TVs, smart speakers, and video doorbells. Along with their benefits come potential risks, since these devices can communicate information (audio recordings, video recordings, television viewing habits) about their users to other parties over the Internet. However, understanding these risks is difficult due to heterogeneity in devices’ online behaviors. For example, smart speakers responding to voice commands send very different network traffic than a smart power plug that is activated via a companion app.

The goal of this project is to measure what IoT devices are doing, simply based on the network traffic they generate. For example, we would like to know if a smart speaker is recording audio from users when it should not, and we can automatically infer this if we have a good model and analysis of what normal, expected recording behavior looks like.

This project will have several outcomes, including published source code and data, published research papers in academic venues, and press articles about our findings through our journalist partners at the New York Times and other prominent venues.

Some available research directions for this project that a student can choose are:

  • PROJECT 1A (Matter): Investigate the use of the Matter protocol. More information about Matter.
  • PROJECT 1B (IPv6): Analyze if an IoT device behaves differently when deployed on an IPv6 network with respect to an IPv4 network.

Prerequisites

  1. Familiarity with the most important Internet and networking protocols and measurement tools (Ethernet, TCP/IP, DNS, Wireshark/tshark).
  2. Extensive programming experience (python recommended).
  3. Strong interest in network security (e.g., traffic filtering, man-in-the-middle, intrusion detection)
  4. When applying, specify which subproject you are interested in and why (1A, 1B).

Project 2. Voice Assistant User Profiling

This project is no longer available.

Voice assistants such as Amazon’s Alexa, are becoming increasingly pervasive in our homes. While convenient, these systems also raise important privacy concerns. One of them is understanding to what extent they use or share information from past user interactions.

In this project the student will investigate how voice assistant interactions may affect user profiling activities, what the risks of such profiling activities are, and if such activities have been disclosed on the privacy policy of the devices. For example, if a user frequently asks a voice assistant for recipes, will they receive more suggestions and advertisements that are food-related? If yes, was this disclosed in the privacy policy of the device?

To perform this investigation, the student will focus on the three most popular voice assistants (Amazon Alexa, Google Assistant, Apple Siri), accessing them using any of the interfaces available to us, including smart speakers, mobile apps, integration SDKs.

This project is organized into three research activities: (1) the creation of user profiles, such as a list of questions aimed at simulating particular user profiles, followed by experiments to actually expose the voice assistants to such questions; (2) the creation of a way to probe the voice assistants for evidence of customization, such as curating a list of control questions, followed by experiments to check if the voice assistants exhibit any sign of profiling; (3) search of evidence of profiling also beyond the voice assistants, for example searching for targeted advertising in Amazon and Google searches.

This research will be performed in the Mon(IoT)r Lab (https://moniotrlab.khoury.northeastern.edu/), where the student will have access to all the smart speakers that are needed for this project.

Although not a strict requirement, we recommend basic knowledge of python, bash, and/or Selenium, which will help to automate some of the manual tasks that may be necessary for discovering instances of user profiling.

Project 3. Analysis of Transcription Errors in IoT and Social Media Platforms

Voice-controlled IoT (e.g., smart speakers) and social media platforms are increasingly becoming an important part of everyday life, and so is their functionality and need for accessibility. One of the features these platforms offer is the ability to automatically transcribe spoken material, which allows users to employ them for controlling their IoT devices, for remote work, for remote learning, or simply basic communication. As the adoption of these platforms increases and automatic transcriptions become a necessity, especially for those who are working in their second language or are affected by physical or hearing disabilities, it is important that the transcripts they generate are accurate, no matter the speaker’s gender, race, or other characteristics.

In this project, students will help analyzing IoT and social media platforms by curating and improving existing spoken material, designing and running experiment for selected platforms, and determining transcription errors.

Specifically, students will be offered the opportunity to participate in one or more of these research activities: (1) listening to the spoken material we already have (containing voices with different characteristics) and checking whether it is correct and matches the transcripts we have; (2) playing the spoken material to one or more selected IoT platforms (e.g., Google Assistant, Amazon Alexa, Apple Siri), social media platforms (e.g., YouTube, Facebook, Microsoft Stream), videoconferencing platforms (e.g., Zoom, Google Meet, BlueJeans, Webex) and retrieving generated transcripts; (3) measuring and comparing the transcription errors exhibited by the analyzed platform(s).

This research can be performed in our lab or remotely.

Prerequisites:

  1. Familiarity with command line interfaces and scripting (bash recommended).
  2. Some programming experience (python recommended).
  3. Willing to learn tools for splitting, merging, and playing audio files.