Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.
Difference between revisions of "AICE WG/"
m |
m |
||
| Line 83: | Line 83: | ||
=== Mailing list === | === Mailing list === | ||
| − | + | The AICE mailing list can be found at [https://accounts.eclipse.org/mailing-list/aice-wg]. Subscription is open to everyone. | |
| + | |||
| + | Archives can be browsed at [http://www.eclipse.org/lists/aice-wg] | ||
=== Monthly meetings === | === Monthly meetings === | ||
| Line 121: | Line 123: | ||
=== The AURA Healthcare use case === | === The AURA Healthcare use case === | ||
| − | Status: Running, see the dedicated page. | + | Status: Running, see [[AICE WG/UseCases/AURA Healthcare|the dedicated page]] for more details. |
| + | |||
| + | Work with the AURA halthcare association on a ML workflow that detects epileptic seizures before they happen, based on ECGs. | ||
| + | * Work with the team on the portability and industrialisation of the prototype. | ||
| + | * Bring the AURA ML workflow to the AICE OpenLab marketplace to foster exchange and collaboration on their solution. | ||
| + | * Provide a k8s cluster to execute resource-intensive tasks and demonstrate the workflow. | ||
=== Automotive - Tool Ecosystem for testing and increase robustness of AI === | === Automotive - Tool Ecosystem for testing and increase robustness of AI === | ||
| Line 130: | Line 137: | ||
* The tool ecosystem shall help to establish an automotive industry-wide accepted way of validating AI modules in automotive products. | * The tool ecosystem shall help to establish an automotive industry-wide accepted way of validating AI modules in automotive products. | ||
* The tool ecosystem shall be available for use by any interested party. | * The tool ecosystem shall be available for use by any interested party. | ||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
| − | |||
Additional information: | Additional information: | ||
* [https://wiki.eclipse.org/images/3/3a/Open_Source_-_Test_and_robustness_increasing_methods.pdf Link to detailed proposal] | * [https://wiki.eclipse.org/images/3/3a/Open_Source_-_Test_and_robustness_increasing_methods.pdf Link to detailed proposal] | ||
| − | === Health- Federated Learning === | + | === Health - Federated Learning === |
Status: Planned | Status: Planned | ||
Revision as of 11:31, 16 December 2021
The Eclipse AI, Cloud & Edge (AICE) Working Group is where the AI, Cloud & Edge ecosystems meet to innovate and grow with open source.
The aim of the AICE Working Groupe is to accelerate the adoption of AI, Cloud & Edge technologies and standards, through the provision and operation of a collaborative work and test environment for its participants, the engagement with research and innovation initiatives and through the promotion of open source projects to AI, Cloud & Edge developers.
The AICE Working Group manages and operates the AICE OpenLab that provides a set of resources to promote the advancement, implementation, and verification of open source software for AI, Cloud, and Edge computing.
Working Group definition and scope
Definition
The Eclipse AICE OpenLab Working Group drives the development, evolution and broad adoption of best practices for AI, Cloud and Edge.
Our focus is to assemble, test and validate AI, Cloud & Edge solutions using calibrated test tools & datasets. The working group also oversees the development of the necessary blueprints and reference architectures that collaboratively combines open source projects, datasets, configurations, and test beds definitions. Together these blueprints and reference architectures aim to deliver end to end use cases that fulfill best practices for privacy, ethics, security, standardization and interoperability.
More generally, the AICE working group is working on:
- Improved compatibility and interoperability of different technologies
- Validated tests of new AI and cloud projects
- Coordination of component life cycles
- Reproducibility for researchers and developers
- Federation of complementary technologies
- Access to free AI & Cloud computing capability platform for developers
- Definition of open specifications and examples of open implementations
- Meeting point for AI, cloud (and robotics, IoT a.o.) initiatives in Europe
The AICE Working Group does this by:
- Fostering open and neutral collaboration amongst members for the adoption of open source technologies.
- Defining and publishing reference architectures, blueprints and distributions of open source software that have been verified for industry AI, Cloud, and Edge standards, requirements, and use cases.
- Developing and providing open source verification test suites, test tools, calibrated datasets and test infrastructure for industry AI, Cloud, and Edge standards, requirements, and use cases.
- Ensuring that key requirements regarding privacy, security and ethics are integrated into all the OpenLab activities.
- Partnering with industry organizations to assemble and verify open source software for their standards, requirements, and use cases.
- Promoting the AICE OpenLab in the marketplace and engaging with the larger open source community.
- Managing the lab infrastructure resources to support this work.
Scope
In general terms, the development and implementation of new (open) standards are complex, time consuming to organise and test, and costly to execute. With the expected level of AI standardisation coming from European initiatives e.g. AI4EU, Gaia-X, and EBRAINS, activities as training, testing and development can best be shared. Because open standards have proven to provide a boost to new business opportunities and technological innovation, it is logical that leading industry players, knowledge institutes and European initiatives collaborate. As such the required comprehensive, safe and open ecosystem/environment does not yet exist, the Eclipse Foundation has taken the first step towards such an AI, Cloud & Edge OpenLab.
This means it does not:
- Set standards
- Provide AI products
This means it can/must do:
- Provision of, but not necessarily restricted to, specifications, testbeds, reference implementations, open APIs, Cloud infrastructure, calibrated data-sets
- Member participation through contribution of technology and/or projects and/or funding and/or in-kind services e.g. manpower, cloud capacity.
- Safeguard sustainability of operations through member contributions and proper allocation of resources
- Promote and build the ecosystem of organisations and community of developers
As a starting point, the scope of the OpenLab program - in order to be relevant to all stakeholders - will address three principle areas of interest:
Policy adoption
- Improve adoption of european standards for interoperability and portability
- The OpenLab intends to collaborate with existing initiatives e.g. Gaia-X, AI4EU, EBRAINS, as guiding programs on behalf of the European Commission digital strategy. Thus enabling incorporation of new European standards by the AI sector.
- The OpenLab provides its members with an open collaborative platform where stakeholders test new standards and develop new approaches and open source (reference) implementations to comply with new standards. With the OpenLab as the European competency centre for AI Open Source, facilitated by the Eclipse Foundation with its long-standing track record in the field.
- Address privacy, ethics and security concerns
- These are major items on the EC digital agenda to be addressed industry-wide
- The open source approach and the collaborative environment provided by the OpenLab will enable specific audit services to be developed and guaranteed in the context of the OpenLab
Business improvement
- Avoidance of vendor lock-in and enabling of technology lock-out
- This necessitates open source and standards, which best can be provided through a vendor-neutral environment where all players can meet, collaborate and share information, and contribute technology in a safe and open environment (in terms of antitrust, legal, code of conduct).
- The OpenLab can guarantee such an environment through proper proven open principles, governance and (specification a.o.) processes.
- Fragmentation within the industry
- The emerging AI sector is still heavily siloed in terms of technologies, organisations and solutions.
- A vendor-neutral common ground as provided by the OpenLab will facilitate and promote synchronous releases, federation of technologies, complementary solutions.
Technology development
- Speed of innovation:
- This requires permissionless open innovation principles applied to the development, test and implementation cycle.
- The open source nature of the OpenLab practices and principles enable shorter time to market, as a central place where the ecosystem can freely access platforms and testbeds to develop, test and implement AI & Cloud solutions.
- Lack of collaboration between education, research and industry:
- Many new innovations in the AI, Cloud & Edge field derive from universities and research institutes. On one hand, these organisations require validation through industry players eg. start-ups, SME’s and large industry players. On the other hand, they contribute heavily to the capacity building of the AI workforce.
- The OpenLab will provide a comprehensive ecosystem and mix of companies, research institutes to enable transfer of technologies and expertise. Furthermore, items such as provision of performance benchmarks and reproducibility can be addressed.
How to participate
The working group is open to the public. If you would like to more actively engage with the initiative, please contact Gaël Blondelle.
Mailing list
The AICE mailing list can be found at [1]. Subscription is open to everyone.
Archives can be browsed at [2]
Monthly meetings
We also hold monthly meetings to discuss actions, organise events and monitor progress our various tasks.
Meetings are announced on the mailing list.
Events
- 2021-11-24 AICE Meetup in Brussels -- see the dedicated page.
Presentations
- Towards an open source AI initiative at the Eclipse Foundation: [3]
- Political challenges and opportunities in making open source AI mainstream: [4]
- Eclipse Deeplearning4j: How to run AI workloads on Jakarta EE compliant servers: [5]
- Meet MindSpore, the new open source AI framework!: [6]
- Q & A: [7]
- Welcome Message | Gaël Blondelle | Open Source AI Workshop S1E2: [8]
- Trustworthy AI & Open Source | Eclipse Open Source AI Workshop S1E2: [9]
- Introduction to Pixano: an Open Source Tool to Assist Annotation of Image Databases | Open Source AI: [10]
Program & Projects
We are currently in the process of drafting the program for the future Working Group.
The AICE OpenLab
The AICE OpenLab is a platform where partners can discuss and share AI-related resources, experiences and benchmarks.
It relies on AI4EU Experiments to provide a marketplace and visual editor to execute complex AI workflows, and also provides a Kubernetes cluster for the execution, demonstration and benchmarking of workflows.
The first project to use the OpenLab platform is the AURA use case.
The AURA Healthcare use case
Status: Running, see the dedicated page for more details.
Work with the AURA halthcare association on a ML workflow that detects epileptic seizures before they happen, based on ECGs.
- Work with the team on the portability and industrialisation of the prototype.
- Bring the AURA ML workflow to the AICE OpenLab marketplace to foster exchange and collaboration on their solution.
- Provide a k8s cluster to execute resource-intensive tasks and demonstrate the workflow.
Automotive - Tool Ecosystem for testing and increase robustness of AI
Status: Planned
Define and implement a tool ecosystem for testing and to increase robustness of automotive AI applications.
- The tool ecosystem shall help to establish an automotive industry-wide accepted way of validating AI modules in automotive products.
- The tool ecosystem shall be available for use by any interested party.
Additional information:
Health - Federated Learning
Status: Planned
Challenge:
- Healthcare is a complexe ecosystem composed of different partners with each a very specific role but each highly interdependent.
Success Criteria:
- Development of an open-source framework that would allow any 3rd party to dispose of an easily deployable federated learning framework.
- Deployment of the open-source framework in real use-cases research projects with hospitals, pharmaceuticals and research institution partners.
- Deployment in Production of the federated learning framework by industrial partners.