Abstract
In this work, we demonstrate Sterling, a decentralized marketplace for private data. Sterling enables privacy-preserving distribution and use of data by using privacy-preserving smart contracts which run on a permissionless blockchain. The privacy-preserving smart contracts, written by data providers and consumers, immutably and irrevocably represent the interests of their creators. In particular, we provide a mechanism for data providers to control the use of their data through automatic verification of data consumer contracts, allowing providers to express constraints such as pricing and differential privacy. Through smart contracts and trusted execution environments, Sterling enables privacy-preserving analytics and machine learning over private data in an efficient manner. The resulting economy ensures that the interests of all parties are aligned.
For the demonstration, we highlight the use of Sterling for training machine learning models on individuals' health data. In doing so, we showcase novel approaches to automatically appraising training data, verifying and enforcing model privacy properties, and efficiently training private models on the blockchain using trusted hardware.
- Martın Abadi et al. "Deep Learning with Differential Privacy". In: CCS. 2016, pp. 308--318. Google ScholarDigital Library
- Ittai Anati et al. "Innovative technology for CPU based attestation and sealing". In: HASP. Vol. 13. 2013.Google Scholar
- Nicholas Carlini et al. "The Secret Sharer: Measuring Unintended Neural Network Memorization & Extracting Secrets". In: arXiv:1802.08232 (2018).Google Scholar
- Raymond Cheng et al. "Ekiden: A Platform for Confidentiality-Preserving, Trustworthy, and Performant Smart Contract Execution". In: arXiv:1804.05141 (2018).Google Scholar
- D. Dao et al. "DataBright: Towards a Global Exchange for Decentralized Data Ownership and Trusted Computation". In: arXiv:1802.04780 (2018).Google Scholar
- Cynthia Dwork, Guy N. Rothblum, and Salil P. Vadhan. "Boosting and Differential Privacy". In: FOCS. 2010, pp. 51--60. Google ScholarDigital Library
- Lisa Fleischer and Yu-Han Lyu. "Approximately optimal auctions for selling privacy when costs are correlated with data". In: EC. 2012, pp. 568--585. Google ScholarDigital Library
- Justin Hsu et al. "Differential Privacy: An Economic Method for Choosing Epsilon". In: CSF. 2014, pp. 398--410. Google ScholarDigital Library
- Nick Hynes, Raymond Cheng, and Dawn Song. "Efficient Deep Learning on Multi-Source Private Data". In: arXiv 1801.07860 (2018).Google Scholar
- Ruoxi Jia et al. "'How Much is My Data Worth?": Data Valuation with Efficient Shapley Value Estimation". 2018.Google Scholar
- Keystone Project. 2018. URL: https://keystone-enclave.org/.Google Scholar
- Pang Wei Koh and Percy Liang. "Understanding Black-box Predictions via Influence Functions". In: ICML. 2017, pp. 1885--1894.Google Scholar
- Ilia A. Lebedev, Kyle Hogan, and Srinivas Devadas. "Secure Boot and Remote Attestation in the Sanctum Processor". In: IACR. 2018, p. 427.Google Scholar
- Frank McSherry. "Privacy integrated queries: an extensible platform for privacy-preserving data analysis". In: SIGMOD. 2009, pp. 19--30. Google ScholarDigital Library
- Joe Near et al. "Optio: Differential Privacy for Machine Learning Pipelines, Statically and Automatically". 2018.Google Scholar
- Oasis Blockchain. 2018. URL: https://www.oasislabs.com/.Google Scholar
- Olga Ohrimenko et al. "Oblivious Multi-Party Machine Learning on Trusted Processors". In: USENIX Security. 2016, pp. 619--636. Google ScholarDigital Library
- Alvin Rajkomar et al. "Scalable and accurate deep learning for electronic health records". In: arXiv:1801.07860 (2018).Google Scholar
Recommendations
A demonstration of VisDPT: visual exploration of differentially private trajectories
The release of detailed taxi trips has motivated numerous useful studies, but has also triggered multiple privacy attacks on individuals' trips. Despite these attacks, no tools are available for systematically analyzing the privacy risk of released ...
Disclosure Avoidance in the Census Bureau’s 2010 Demonstration Data Product
Privacy in Statistical DatabasesDemonstration Paper: Wirelessly Sensing Medication Administration: Cyber-Physical Event Detection and Notification Utilizing Multi-Element Chipless RFID
WH '14: Proceedings of the Wireless Health 2014 on National Institutes of HealthMedication administration is one pathway by which Adverse Drug Events (ADE) can occur. While Electronic Medical Administration Record (eMAR) systems help reduce the number of ADEs, current eMAR implementations suffer from workarounds that defeat safety ...
Comments