[cs-talks] Upcoming Seminars: BUSec (Weds) + Danna Gurari's PhD Proposal (Thurs)
Conroy, Nora Mairead
conroynm at bu.edu
Thu Jan 15 10:02:48 EST 2015
Practice talk: Sorting and Searching Behind the Curtain
Foteini Baldimtsi, BU
Wednesday January 21, 2015 at 10am in MCS 180 - Hariri
Abstract: We study the problem of private outsourced sorting of encrypted data. We start by proposing a novel sorting protocol that allows a user to outsource his data to a cloud server in an encrypted form and then request the server to perform computations on this data and sort the result. To perform the sorting the server is assisted by a secure coprocessor with minimal computational and memory resources. The server and the coprocessor are assumed to be honest but curious, i.e., they honestly follow the protocol but are interested in learning more about the user data. We refer to the new protocol as private outsourced sorting since it guarantees that neither the server nor the coprocessor learn anything about user data as long as they are non-colluding. We formally define private outsourced sorting and provide an efficient construction that is based on semi-homomorphic encryption. As an application of our private sort, we present MRSE: the first scheme for outsourced search over encrypted data that efficiently answers multi-term queries with the result ranked using frequency of query terms in the data, while maintaining data privacy. To construct MRSE we use searchable encryption techniques combined with our new private sort framework. Finally, although not discussed in this work, we believe that our private sort framework can turn out to be an important tool for more applications that require outsourced sorting while maintaining data privacy, e.g., database queries. Joint work with Olga Ohrimenko
PhD Thesis Proposal
Danna Gurari, BU
IVC Research Group
Thursday, January 22, 2015 at 9am in MCS 148
Abstract: Advances in image acquisition and storage technologies have led to many image-based experiments that are designed to systematically study basic science processes. This thesis centers on the design of image annotation methods to accelerate the transition from data collection to scientific discovery. The focus is on biomedical images in an effort to contribute to research that addresses society’s health care problems. The first half of the work includes a detailed analysis of the relative strengths and weaknesses of three different approaches to demarcate object boundaries in images: by experts, by crowdsourced laymen, and by automated computer vision algorithms. Results revealed that popular computer vision algorithms and state-of-art human computation crowdsourcing systems come with unique strengths but both methods were insufficient alone to yield the desired accuracy and scale to support quantitative analyses. These results inspired a research trajectory for gaining traction on this problem by examining how to effectively integrate computer vision algorithms and crowdsourced laymen in order to achieve expert-level annotation at scale. Results from two implemented hybrid crowdsourcing-computer methods are compared to annotations of experts, two pure crowdsourcing methods, and two pure algorithmic methods on 405 objects in everyday and biomedical images. Experiments revealed that a hybrid system design yielded the most accurate results among the six system configurations and it yielded an accuracy that is statistically similar to segmentations created by biomedical experts. The final chapter of this proposal includes a case study that will underscore the potential for this hybrid algorithm-crowdsourcing approach to accelerate big data computational biology analyses. A system will be built that incorporates this approach and show how it creates an expert-quality 3D volume reconstruction of fish structures. This work will support analyses that link biological structure and function. To encourage community-wide effort to continue working on developing methods and systems for image-based studies which can have real and measurable impact that benefit society at large, datasets and code are publicly-shared (http://www.cs.bu.edu/~betke/BiomedicalImageSegmentation/).
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the cs-talks