Achieving Trustworthy Biomedical Data
Image source: Facebook Engineering
MotivationPrivacy and trust of data capture and sharing is an issue rising to the center of public attention and discourse. With the advents of large-scale academic, medical, and industrial research initiatives that collect personalized biomedical data from users, methods for providing sufficient privacy in biomedical databases and conveying a sense of trust to the user are crucial for the field of biocomputing to advance with the grace of the public. If the public does not trust innovations in biomedical systems, then research funding of these efforts will not be supported. In an age where privacy of personal data is at the forefront of geopolitical issues and public discourse, all measures must be taken to ease the potential apprehension of patients sharing their data with academic and industry institutions. It is therefore crucial that we confront and begin to address these issues now. We need the best ideas supported by rock solid science.
Session TopicsThe importance of trust in healthcare solutions, especially with machine learning approaches deployed on mobile devices, cannot be overstated. Privacy and trust in big data systems has recently come to the forefront of public and academic discourse across all domains. While all of the traditional issues related to data privacy and sharing apply to biomedical data, there are additional challenges and opportunities within the domain of biocomputing. Research themes in this area include, but are not limited to, sharing of datasets in a privacy-preserved manner, algorithmic techniques for transforming data in a secure manner (e.g., homomorphic encryption), integration of omics data across data sources, human-computer interaction (HCI) studies of biomedical systems which collect data, demonstrations of security flaws in existing systems, ethical issues related to sharing, and development of biomedical systems which intrinsically prioritize trust, transparency, and/or privacy in their design. Examples of perspectives within the scope of this session include:
- Novel algorithms and techniques for enabling secure querying of patient data, including new encryption methods, secure aggregation techniques, and transformations of biomedical data into a format that is secure but still useful to researchers and other stakeholders.
- Integration of multi-omics data across data sources in a secure and trustworthy manner.
- Demonstrations of successful privacy-preserving machine learning techniques and paradigms applied to biomedical domains, such as federated learning and dimension reduction.
- Demonstrations of security and privacy holes in existing biomedical systems and methods.
- Mobile technology for personalization of care and ownership of data.
- Incentive mechanisms for sharing health data, such as payment to owners.
- Novel data augmentation techniques for biomedical datasets.
- Development of biomedical systems which intrinsically prioritize trust, transparency, and/or privacy in their design.
- Human-computer interaction (HCI) studies of user perceptions of and trust in new and established biomedical systems.
- Ethical case studies and perspectives about current data science principles, grounded in theoretical bioethical frameworks.
- Trustworthy use of adaptive learning and human-in-the-loop labeling processes to build appropriate training data and network architectures.
- Any submission within the general scope of PSB which highlights trust, security, and/or privacy as a central aspect.
Dennis P. Wall
Nicholas P. Tatonetti
Mount Sinai Health System
Submission InformationThe submitted papers are reviewed and accepted on a competitive basis. At least three reviewers will be assigned to each submitted manuscript.
August 3, 2020: Call for papers deadline. THIS IS AN ABSOLUTE DEADLINE.
September 14, 2020: Final paper decisions announced by PSB organizers.
October 1, 2020: Camera-ready papers due.
November 15, 2020: Abstract deadline for non-reviewed posters.
January 3-7, 2021: Conference dates.