AI4Life

AI4Life_Logo
Menu

Author: Pasi Kankaanpää

Categories
event report

BioHackathon Europe 2022

Biohackathon europe 2022

Paris, 7-11 November 2022

by Beatriz Serrano-Solano

AI4Life was present this year at the BioHackathon Europe 2022, which took place in November at Campus Des Berges de Seine, near Paris. Since 2018, ELIXIR Europe has been organising the BioHackathon to bring together both ELIXIR and non-ELIXIR members to work on projects aligned with the ELIXIR Platforms, Communities or Focus Groups.

The participation of AI4Life was focused on two projects:
– Project 9: Disseminating FAIR Machine Learning Models via BioModels
– Project 17: Metadata schemas supporting Linked Open Science (with a focus on reproducibility)

Through the interaction with these two projects, we explored the collaboration of the BioImage Model Zoo with databases like BioModels, by discussing the minimum metadata that would allow the interoperability between such resources. We also engaged with more initiatives like Bioschemas, the DOME recommendations, the EDAM ontology and the ELIXIR Machine Learning Focus Group.

We will be happy to continue these interactions, and we look forward to our future collaboration and participation in similar events.

AcknowledgementS
Categories
past events

Workshop: Towards FAIR AI image data

Workshop: Towards FAIR AI image data

Date

Venue

Description

24-25 January, 2023

Online

The BioImage Archive (BIA), EMBL-EBI’s data resource for open life sciences image data, provides general-purpose deposition services for any imaging dataset accompanying a publication, as well as reference image data. The BIA currently supports basic deposition of image annotations together with corresponding images. As part of the recently awarded Horizon Europe project, AI4LIFE, we would like to improve the BioImage Archive’s support for image annotations as part of AI-ready datasets.
In particular, we wish to make deposited annotations as widely usable as possible, adhering to the FAIR principles of Findability, Accessibility, Interoperability and Reusability. This is particularly challenging for image annotations, since there are few standards for representing annotation data that are widely adopted across the community.
Towards this end, we will hold a workshop involving participants from the bioimaging AI community, including data generators, annotators, AI researchers and software/tool developers.

Categories
event report

AI4Life First Hackathon

FIRST AI4LIFE HACKATHON

Heidelberg, 27-28 October 2022

by Beatriz Serrano-Solano

During the kick-off meeting in October 2022, the AI4Life project partners discussed and identified various topics that the team will be tackling over the next 3 years. Following up on the meeting outcomes, a two-day hackathon was organized to start working on those specific topics and challenges that came out of the discussion. Throughout the two-day event, participants were able to make significant progress, and soon after the beginning of the event, the first success stories started to arise. More of these events will follow during the 3 years of the project to keep the momentum and the collaboration going. Thank you to all the participants!

AcknowledgementS
Categories
event report

Kick-off meeting in Heidelberg

Kick-off meeting

Heidelberg, 25-26 October 2022

by Beatriz Serrano-Solano

AI4Life European project is a collaborative effort that brings together 10 partners from 8 different European countries. Those partners include 4 European Research Infrastructures, covering a broad scope of scientific use cases: marine biology (EMBRC), plant phenotyping (EMPHASIS), compound screening (EU-OPENSCREEN) and structural biology (Instruct). The project is funded with a budget of 4 million euros, which will be used to bridge the gap between the computational and life science communities. AI4Life will last for 3 years, during which time the team will work together to achieve the project’s objectives:
  1. Democratized availability of AI-based image analysis methods.
  2. Establish standards for the submission, storage and FAIR access.
  3. Simple model deployment, sharing, and dissemination through a new developer-facing service.
  4. Organize Open Calls and Challenges for image analysis problems.
  5. Empower common image analysis platforms with AI integration.
  6. Organizing outreach and training events i.e. image analysis courses/workshops and participation in international conferences.

AI4Life officially started in September and kicked off at the end of October 2022. During the kick-off meeting, experts from across Europe gathered in Heidelberg for two days, during which the participants actively engaged in discussions about the timeline, milestones, deliverables and work distribution among the partners, and agreed on a plan for the next few months. The team members are also committed to maintaining an open line of communication with the audience and keeping them informed of the project’s progress throughout the duration of the project. 

Overall, the team had a productive meeting, and we are excited to begin working together.  Thank you to all the organizers, speakers, and participants!

Program
Time Topic Speaker/s
09:00 – 09:10 Welcome Rachel Robinson-Lehtinen
09:10 – 09:40 Introduction to AI4Life & Euro-BioImaging ERIC John Eriksson
09:40 – 09:50 Interactions with other Horizon Europe funded projects Antje Keppler
09:50 – 10:55 Interactions with partner RIs Aastha Mathur, Davide de Cioccio, Roland Pieruschka,
Katja Herzog, Silke Schumacher
10:55 – 11:00 Coffee Break
11:00 – 11:30 Message from the European Commission Project Adviser Antonio Ventura
European Research Executive Agency
11:30 – 12:00 EOSC and costing for cloud resources Peter Maccallum
CTO ELIXIR
12:00 – 13:00 Lunch & Group Photo
13:00 – 13:25 Project administration, timelines & reporting (WP1: Project Management) Rachel Robinson-Lehtinen
Appointment of the Executive Board
Appointment of the Open Calls Selection Committee
13:25 – 15:30 Introduction of Scientific Topics Anna Kreshuk & Florian Jug
WP2: User services & computing infrastructures Wei Ouyang
WP3: Direct support Arrate Muñoz
Discussion Anna Kreshuk & Florian Jug
15:30 – 15:40 Coffee Break
15:40 – 17:45 WP4: Contributing services Ricardo Henriques
WP5: Data, model & computing standards Matthew Hartley
WP6: Support for open calls, challenges & new services Florian Jug
WP7: Communication, outreach & training Aastha Mathur, Marianna Childress-Poli
Discussion Anna Kreshuk & Florian Jug
19:15 Dinner
Time Topic
09:00 – 12:30 Workshop on technical discussions / Pressing issues raised in Day 1
AcknowledgementS
Categories
News

Press Release

Press release

25.10.2022

Bridging Artificial Intelligence and Machine Learning with Bioimage Analysis

The AI4Life project launched today, marks an exciting chapter in the computational and life science research communities. The €4 million Horizon Europe funded project aims to create accessible, harmonized, and interoperable AI tools and methods for solving today’s microscopy image analysis problems.

THE GAP

Machine learning (ML) has accelerated frontier research in the life sciences, but democratized access to such methods is not a given. Limited access to necessary hardware/software and expertise combined with insufficiently documented methods hinder life science researchers from harnessing the power of such tools. Furthermore, while modern Artificial Intelligence (AI)-based methods typically generalize well to unseen data, no standard exists for sharing and fine-tuning pretrained models between different analysis tools. Compounding the issue, existing user-facing platforms operate entirely independently, often failing to comply with FAIR data and Open Science standards. Furthermore, the staggering pace of AI and ML development make it impossible for the non-specialist to stay up to date. Hence, urgent services and infrastructures to solve such problems are required to expand cutting edge life science research.

THE BRIDGE

The 10-partner consortium will build an open, accessible, community-driven repository of FAIR pre-trained AI models and develop services to deliver these models to life scientists, including those without substantial computational expertise. AI4Life will provide direct support and ample training activities to prepare life scientists for responsible use of AI methods. Additionally, AI4Life will drive community contributions of new models and interoperability between analysis tools. AI4Life will also facilitate Open calls and public Challenges aimed at providing state-of-the-art solutions to unsolved image analysis problems in life science research. 

AI4Life brings together AI/ML researchers, developers of open-source image analysis tools, providers of European-scale storage and compute services, and European life science Research Infrastructures – all united behind the common goal to enable life scientists to benefit from the untapped, tremendous power of AI-based analysis methods.

THE CORE OBJECTIVES

  • Democratize availability of AI-based image analysis methods
  • Establish standards for the submission, storage, and FAIR access of reference data, reference annotations (ground-truth), trained AI models, and trainable AI methods
  • Simple model deployment, sharing and dissemination through a new developer-facing service
  • Organize Open calls and Challenges for image analysis problems
  • Empower common image analysis platforms with AI integration
  • Organizing outreach and training events

THE TEAM

Our multidisciplinary team of experts in computational and life sciences as well as 4 European Research Infrastructures.

10

PARTNERS

3

YEARS

8

COUNTRIES

4M

EUROS

Categories
Uncategorized

About us

ABOUT US

AI4Life in a nutshell

AI4Life is a Horizon Europe-funded project that brings together the computational and life science communities.

Its goal is to empower life science researchers to harness the full potential of Artificial Intelligence (AI) and Machine Learning (ML) methods for bioimage analysis particularly microscopy image analysis, by providing services, and developing standards aimed at both developers and users.

AI4Life promises to create harmonized and interoperable AI tools & methods via open calls and public challenges and bring these developments to researchers via strategic outreach and advanced training.

The services provided and solutions developed within the AI4Life framework are crucial to solving today’s microscopy image analysis problems and will contribute to boosting the pace of biological and medical insights and discovery in the coming years.

The BioImage Model Zoo and FAIR data principles are core facets of the AI4Life project.

10

PARTNERS

3

YEARS

8

COUNTRIES

4M

EUROS

Objectives

and our goals

1

Democratized availability of
AI-based image analysis methods

2

Establish standards for the submission, storage and FAIR access

3

Simple model deployment, sharing, and dissemination through a new developer-facing service

4

Organize Open Calls and Challenges for image analysis problems

5

Empower common image analysis platforms with AI integration

 

6

Organizing outreach and training events i.e. image analysis courses/workshops and participation in international conferences

COMMUNITY PARTNERS

Our extensive network

PROJECT PARTNERS

Our interdisciplinary team

STRUCTURE

The way we work

Accordion Content

The overall aim of WP1 is to provide project management support to the consortium as a whole as well as to the individual consortium partners in matters linked to AI4Life. Professional project management will establish the conditions for effective and efficient delivery of AI4Life.

One of our major goals is to provide a comprehensive resource website for users and enable non-computational users from the life sciences to easily test-run and evaluate pre-trained models and run computational tools in the web browsers. We will provide a user-friendly UI, a cloud computing infrastructure for running AI models in the web browser and a standard for running tools in our computing infrastructure or connecting web applications and desktop software tools to the website. The computing infrastructure will be made available through providers within the European Open Science Cloud (EOSC).

The overall aim of the WP3 is to make the BioImage Model Zoo realise its full potential by bringing its Deep Learning (DL) tools and the related material to the point that they are easily and conveniently usable by the end-users and the model contributors. Without responsive and efficient user support, and good feedback to the consumer software, the BioImage Model Zoo risks becoming a container of models, data and code lacking real application. On the contrary, by implementing the AI4Life vision, we will achieve widespread use of DL tools in the life sciences community elevating the biological questions that can be answered to the next level. For that, WP3 will focus on consolidating the integration of the DL tools provided by the BioImage Model Zoo into the users’ image processing workflows. Moreover, the necessary mechanisms will be established to assure the correct deployment of the provided DL tools in real applications. We plan to achieve it through direct support for users and contributors of models.

WP4 aims to establish a framework that helps method developers and advanced image analysts to integrate their deep-learning-based algorithms into the AI4Life projects, such as the Bioimage Model Zoo. By doing so, we will empower developers with analytical and computational tools to reach a large audience of non-computational scientists and further their research capacity through accessible machine-learning technology. The WP, in particular, will provide base tools and guidelines, encouraging open contributions crowdsourced by the imaging community in a Findability, Accessibility, Interoperability, and Reusable (FAIR) form.

Accordion Content

This work package will develop and implement support strategies, build and strengthen support networks, and run open calls for supporting challenging analysis problems that (i) are limiting the rate of progress in the field the analysis problems originate from and (ii) are likely to be significantly improved or even solved when employing modern AI-based methods.

D5.1. Annotation standards and software, libraries and reference examples

This work package will develop and implement support strategies, build and strengthen support networks, and run open calls for supporting challenging analysis problems that (i) are limiting the rate of progress in the field the analysis problems originate from and (ii) are likely to be significantly improved or even solved when employing modern AI-based methods.

This work package will develop and implement a communication and outreach strategy as well as training opportunities for scientists from diverse fields, including biology, medicine, physics and informatics, who could benefit from using AI-based analysis tools on their image data. The communication strategy will ensure that key stakeholders (biologists, medical professionals, imaging scientists, (bio-)informaticians, funders, etc., as well as partners affiliated with EOSC-related projects) are aware of the AI4Life project and the opportunities it provides, so that they can interact and engage in our activities.

contact

Get in touch with us

ai4life@eurobioimaging.eu

@AI4LifeTeam