Project Partner: ALGOLYSIS

Ma(R)s

Ma(R)S, aims to engage a PhD Student (PhDS) in conducting state-of-the-art research and industrial development in the fundamental and highly relevant area of Strongly Consistent Distributed Shared Memory Systems (DSM).
Supervised by leading researchers in the field, both from academia (University of Cyprus) and industry (Algolysis Ltd), and leveraging on her involvement in the development of a Proof of Concept DSM (under the two partners), the PhDS will work in two fronts: (i) Theory: to advance the knowledge in the field of DSM by exploring efficient, robust, and practical solutions in highly dynamic environments, and (ii) Practice: develop a Memory-as-a-Service (MaaS) platform for facilitating the deployment, management, and usage of DSM for data sharing and for supporting the next-generation of (collaborative) distributed applications. The entire framework will be deployed and validated in a relevant environment (TRL6). Specifically, the platform will be used to deploy DSM algorithms over a set of stationary servers and over a set of less-powerful peer-to-
peer devices.

https://projects.algolysis.com/mars/

Smart Parts Detector in Manufacturing using Deep Learning

We are designing and implementing SmartPartsDetector, a new component for the SHOP4CF EU project, that enables manufacturing SMEs of small parts to optically identify their parts by leveraging state-of-the-art convolutional neural networks (CNN). This fundamental capability can then be used to assist and empower humans by devising a wide range of digital tools, e.g. production quality control software, staff training applications, etc.

KeepA(n)I: A Methodological Approach for Identifying Social Stereotypes in Artificial Intelligence Applications

KeepA(n)I (pronounced “Keep an eye and/or Keep AI ”) aspires to play a key role in aiding towards achieving Trustworthy AI, promoting a novel solution that will lay the groundwork for detecting bias in AI applications. The project considers high-risk applications of AI and aims at creating an active approach in detecting social stereotypes in the input and output of the AI application, as to inform the developer of the social behavior of the developed application. Consortium partners will perform state-of-the-art research, positioning them at the forefront of the area of FAIR / Trustworthy AI, and addressing the challenges brought about by the “democratization of AI” within Europe.

Utilizing Efficient Reads for ATomic Objects in 3D Networked Virtual Environments

The advent of fast highly-available network connectivity in combination with affordable 3D hardware (GPUs, VR/AR HMDs, etc.) has enabled making Networked Virtual Environments (NVEs) possible and available to multiple simultaneous end-users beyond the confines of expensive purpose-built 3D facilities and laboratories. However, the algorithms making possible the NVEs of today are already reaching their limits, proving unreliable, suffer asynchronies and deployed over an inherently fault-prone network infrastructure. Thus, new scalable, robust, and responsive strategies that build on top of unreliable, asynchronous, and fault-prone network infrastructure, must be devised in order to support the needs of the NVEs of the future.
Current developments of distributed architectures handle concurrency by either providing weak consistency guarantees (e.g. eventual consistency), or by relying on the bounded life span of inconsistent states. However, recent scientific works are shifting the viewpoint around the practicality of strongly consistent distributed storage spaces by proposing latency-efficient algorithms of atomic read/write Distributed Shared Memory (DSM) with provable consistency guarantees.

Our approach is to combine a DSM implementation of ERATO with a hybrid approach leveraging the simplicity of the client-server architecture. This, allows us to take the best features of the Multi-Server architecture and combine them with a low-level Distributed Shared Memory algorithm offering (A) ease of integration with existing VEs and implementation of new ones, since the DSM promises to take care of synchronization and consistency across servers, without application-specific interventions at client nodes and no special hand-off mechanisms between servers for users joining and leaving; and (B) reduced system complexity, and in turn faster development time and lower deployment and maintenance costs.