History

The Knowledge 4 All Foundation was founded in 2009 by lead partners in PASCAL and PASCAL2 research networks of Excellence.

Networks as vehicles of collaboration

After three iterations of research projects starting from NeuroCOLT and NeuroCOLT2, PASCAL and PASCAL2 coming to a successful end in November 2013, the European Commission announced plans to establish a different funding scheme, namely Horizon 2020 and Erasmus+ which were not to include Networks of Excellence, therefore effectively stopping the funding of large scale scientific and research communities. The Commission requested from the partners to create a continuation of the PASCAL’s best practices, therefore Knowledge 4 All Foundation became PASCAL’s legacy organisation.

PASCAL2 success story in AI

The key successes of PASCAL2 have been to make a big impact in research direction in AI and machine learning in a positive way, and encourage proactive and forward-looking research. To empower young people and help them feel that they can make up their own minds about their own research programme. PASCAL created a new breed of young, hungry researchers, which was great to see. And to be a bit sentimental we’ve found a really nice group of people, a nice atmosphere. The way people worked together has been really positive.

Continuing with the model

The Knowledge 4 All Foundation is the legacy foundation. The idea was to create a permanent presence in Europe. To achieve this we established a legal entity that could carry forward some of the activities of PASCAL. We founded it as a UK company, registered as a charity. The main focus at the beginning was on the VideoLectures.Net website and the field of open education, and the PASCAL2 network of machine learning experts, this quickly extended into running large-scale AI projects and supporting marginalized AI communities across the world.

PASCAL2 major achievements

The PASCAL2 Network achieved significant advancements across various domains of machine learning, artificial intelligence, and their applications. Here are the primary achievements:

1. Expansion of Machine Learning Research and Collaboration

Distributed Institute: PASCAL2 continued and expanded the distributed institute model initiated by its predecessor, enabling robust collaboration across Europe and globally.
Encouragement of Young Researchers: The network empowered a new generation of researchers, fostering creativity and independence in their research trajectories.
Interdisciplinary Collaboration: PASCAL2 brought together cognitive scientists, machine learning specialists, and mathematicians to address complex, multidisciplinary problems.

2. Technological Contributions

Machine Learning Techniques: Advanced research on pattern analysis, statistical modeling, and adaptive systems significantly improved the field’s understanding and application.
Challenges and Competitions: Initiatives like the Visual Object Classes Challenge and Recognizing Textual Entailment pushed the boundaries of AI and natural language processing.

3. Applied Innovations

Media Analysis: The creation of tools like NOAM (News Outlets Analysis and Monitoring System) enabled large-scale media data analysis, identifying patterns and trends in sentiment and reporting.
Disaster Victim Identification: The Bonaparte system, leveraging Bayesian networks, revolutionized victim identification in disasters by improving DNA analysis.
Robotics Learning: Developed robots capable of dynamic learning, exemplified by a robot learning table tennis skills through reinforcement learning, bridging theoretical machine learning and real-world applications.

4. Outreach and Global Impact

Educational Initiatives: PASCAL2 organized workshops, summer schools, and boot camps, such as those in Ghana, to educate and develop global expertise in machine learning.
Knowledge Dissemination: VideoLectures.net and open-access initiatives democratized access to educational content and research outputs.

5. Infrastructure and Funding

Pump Priming and Harvest Programs: Funded exploratory research and short-term projects, leading to impactful spin-offs and practical applications like machine translation tools and sentiment analysis frameworks.
Resource Development: Created and maintained datasets and software libraries for broader research use.

6. Legacy and Sustainability

Knowledge for All Foundation: Established to ensure the continuity of the network’s goals post-project, focusing on open access to academic resources and fostering global research collaboration.

7. Recognition and Influence

Scholarly Contributions: Significant presence in top conferences like ICML, NIPS, and ECML, with many PASCAL-related papers becoming foundational works in their fields.
Practical Applications: Contributions influenced various sectors, including healthcare, natural language processing, and robotics, demonstrating the versatility of machine learning techniques.

PASCAL2 has left an indelible mark on the research landscape, merging theoretical advancements with practical, real-world solutions.

Library

Dedicated to the freeing of the refereed research literature generated through our projects through author/institution self-archiving, we have created a collection of a few hundred papers preserved in our Eprints archive. The main achievement was to match the production of European high-quality papers at relevant computer science conferences like NIPS, now NeurIPS conference.