Log in

We are recruiting !

We are looking for a number highly motivated PhD students for the following three funded PhDs over the period 2017-2020. More detail is available below and in the attached documents.

  • PhD1: Robust and light-weight overlay management for decentralized learning

A growing number of companies are extracting value from the digital data produced by our modern society using Machine learning (ML) techniques. Most of these companies rely today on centralized or tightly coupled ML systems hosted in data centers or in the cloud. This is problematic as this concentration poses strong risks to the privacy of users, and limits the scope of ML applications to tightly integrated datasets under unified learning models.

To address these limitations, this PhD proposes to explore an alternative approach inspired by peer-to-peer networks in which users control their own system, and only exchange a limited amount of information to construct local machine learning models. This strategy is more amenable to preserving user privacy, and respecting the constraints possibly imposed on sensitive data-sets (such as health records, or personal financial data), and holds the potential for highly scalable and robust learning systems. This project aims to study the challenges raised by this strategy in terms of distribution and overlay management.

Contact: François Taïani. More details available here. Send us your application (CV, grade transcripts, references) by email.

  • PhD2: Distributed troubleshooting of edge-compute functions (with Technicolor)

The rise of micro-services, fog-, and edge-computing are prompting a fresh rethink of the typical distribution of capabilities between servers and clients in a distributed applications. This is likely to lead to more services and computations being offloaded to geo-distributed devices, in particular within hybrid cloud/edge architectures. In this context, this PhD thesis focuses on the opportunities that recent generations of end-user gateways (or more generally end-user devices) will offer to implement an edge-compute paradigm powered by user-side micro-services (as illustrated by Amazon’s recent announcement of its Greengrass platform).

In order for service and application providers to actually use the system and deploy applications, the system must ensure an appropriate level of reliability, while simultaneously requiring a very low level of maintenance in order to address the typical size and economics of gateway deployments (at least a few tens of million units).

The PhD thesis aims to build and evaluate models, algorithms and data-structures with a sufficient level of abstraction, in order to deliver results that remain generally applicable beyond any particular use case or application. However, we also plan to implement and evaluate on edge-devices some of the primitives designed within the course of the PhD thesis, in order to validate that the proposed primitives are able to execute in a resource-constrained environment.

Contact: François Taïani, Christoph Neumann (Technicolor). More details available here. Send us your application (CV, grade transcripts, references) by email.

  • PhD3: BBDA – Browser Based Data Analytics

Big data has become the new buzzword not only in computer science but in most scientific fields. The ability to store and analyze huge quantities of data has given birth to numerous new applications, jobs, advances in science, and will probably lead to further disruptive changes as data analytics techniques spread to more and more domains. Yet, the ongoing big-data revolution represents a significant challenge on at least two fronts: scalability and privacy. In terms of scalability, the ever-increasing amount of available data requires systems and protocols that can process and extract information quickly, and at a reasonable cost. In terms of privacy, we see a clear conflict between the need to protect personal data and the potential that this data offers for data analysis applications. If today, we have data analytics protocols that crunch our shopping behavior on sites like Amazon.com or at local supermarkets, it is reasonable to expect that these same protocols could operate on more and more sensitive data coming from smart appliances in our homes, or on health monitoring platforms.

This PhD thesis will explore how solutions to both these important big-data challenges can come from a shift from traditional cloud architectures to decentralized or semi-decentralized alternatives based on the edge/fog computing model. In particular, the candidate will explore how end-user devices like web browsers can gather, exchange, and process data in a collaborative manner without the need of central control. This will lead to new protocols for coordination, communication, and data exchange between web browser based on the WebRTC API. The candidate will then explore how these protocols can serve as a basis for novel decentralized solutions for data analytics. Finally, he or she will develop solutions that make it possible to combine this new fog-oriented model with traditional cloud architectures. Throughout the thesis, the candidate will also engineer the proposed protocol into reusable software components. This will lead to a highly scalable data management system that can operate on highly sensitive data without violating the privacy of users.

Contact: Davide Frey. Send us your application (CV, grade transcripts, references) by email.

For Internships, contact Francois.Taiani@irisa.fr

One Response to “Open positions”

  1. […] We are looking for a number highly motivated PhD students for the following two funded PhDs over the period 2017-2020. More detail here. […]