This is a project completed for IC3 2022 Blockchain Camp Hackathon. (2nd Place)

Abstract:

We aim to create a private and traceable application that can be used for patients to track and be notified of contamination of medical products such as medicine, vaccines, medical devices, etc. Current methods are not user friendly and are not effective at notifying all users who may be effected. Our solution uses the secure enclave backed blockchain platform Secret Network to implement a private smart contract that can be used by manufacturers, doctors, pharmacists, patients, regulators, and in neccesary cases the public. We also aim to protect manufacturer interests by only revealing adverse affects that have met the predetermined threshold.

GitHub Code

Abstract:

This project and demo is an implementation of a secret data transfer application using Intel SGX secure enclaves and blockchain based bulletin board. This application tries to solve the problem of secure data recovery by designating a semi-trusted administrator that can verify the user's identity during the recovery process. It employs different strategies to ensure accountability of the administrator. In order to limit the admin's power we implement restrictions on the number and frequency of data retrieval, have a mechanism for users to be notified of and cancel a fraudulent retrieval, and allow for public auditing of all signups and retrievals in the enclave.

GitHub Code

This is a project completed for my CS563 Advanced Computer Security Class.

Abstract:

Blockchains are a powerful technology that can be leveraged to allow for accountability and collaboration between multiple parties. With the popularity of permissioned blockchain systems such as Hyperledger Fabric for enterprise use cases, strong expressive access control policies are a necessary tool to facilitate sharing of data and encourage trust among all participants.

The ability to revoke access to data submitted on the blockchain is desirable due to it's flexibility with respect to security guarantees. We create an implementation of this ability and apply it to a Chinese wall access control model use case. We run benchmarks on this implementation to measure it's usability.

GitHub Code

This is a project completed for my CS 598 ML Machine Learning for Sys, Networks, and Security course.

Abstract:

Studies also show that people find Instagram celebrities' endorsements more trustworthy than traditional celebrity endorsements [1]. Thus, spending on influencer marketing is also a growing industry with $8 billion expected to spent in 2020 [2]. With the popularity of influencer marketing on the rise the Federal Trade Commission (FTC) released guidelines on how these influencers should disclose their connections to brands on their social media posts. These guidelines intend to inform the followers of these influencers when they are being advertised too. The FTC states that they do not monitor any social media websites for these violations but would investigate violations that are brought to their attention. This paper aims to analyze if the public trust in influencers is being exploited by hidden advertisements that are not being disclosed.

In order to accomplish this task the first step is to be able to detect if and when these advertisements are disclosed. We do a qualitative analysis of our dataset and describe some different ways influencers disclose marketing. We also identify two potential obfuscation techniques that are used by influencers. These can be the basis for some user study.

We hypothesize that many brands that are doing influencer marketing require the influencers to say or mention specific aspects of their products. We use similarity metrics to analyze posts mentioning a brand. If a post that does not disclose sponsorship is very similar to a post that does disclose sponsorship we conclude that this may be an FTC violation. We also do a case study on a brand that uses an uncommon form of disclosure.

[1] Chen Lou and Shupei Yuan. 2019. Influencer Marketing: How Message Value and

Credibility Affect Consumer Trust of Branded Content on Social Media.

Journal of Interactive Advertising 19, 1 (2019), 58–73. https://doi.org/10.1080/15252019. 2018.1533501 arXiv:https://doi.org/10.1080/15252019.2018.1533501

[2]L99 Firms. 2020. Influencer Marketing Spend. https://99firms.com/blog/influencer-marketing-spend/