JOINT HUMANE AI NET AND NAIXUS CALL FOR COLLABORATIVE PROJECTS
BACKGROUND
HumanE AI net has been running a program of micro-projects and there is a potential to link this with the Network for AI and Knowledge for Sustainable Development (NAiXUS) established jointly by International Research Centre on AI under the Auspices of UNESCO, the DataPop Alliance, The Knowledge 4 All Foundation, ELLIS Alicante UNIT and Regional Center for Studies on the Development of the Information Society (Cetic.br). The Humane AI Network of Excellence has funds reserved to finance the involvement of external partners and this call is concerned with micro-projects that would like to leverage these funds to include NAiXUS partners.
REQUIREMENTS
Project proposals should have a small setup in which 1 or more researchers are financed through HumaneAI micro-project funds with the involvement of at least one external partner drawn from a NAiXUS node. The team should work together to produce a paper, dataset or proof of concept system, preferably co-located or at least partially co-located. The proposal should specify a timeframe for the start date of the project and a completion deadline of at most six months from the project start.
FOCUS
The focus of the call is on ‘Collaborative Artificial Intelligence’ as described in Section 3.1 of the updated Humane AI Research Agenda (see HAI-Net-Deliverable-D6.1). We are interested in projects that develop both theoretical models and practical demonstrations of AI systems able to interact among themselves or with humans through representations that are understandable and meaningful to humans. We anticipate that this will require unpacking the black box to provide a compositional structure that can be used to ‘explain’ or ‘justify’ the opinion expressed by the AI system.
Such representations might be structured in terms of so-called ‘narratives’ that can link elements of a situation to previously encountered themes or schemas but could be simpler structures or be based on natural language. While this is not the traditional approach taken towards explainable AI (XAI), it clearly places XAI as an enabler of collaborative AI, a focus that gives a different emphasis and aim to XAI.
Projects could also explicitly develop new theoretical models for multi-agent collaboration and/or compositionality of learning systems. This could include ways of representing and using uncertainty estimates in order to combine evidence for a particular hypothesis that is communicable. Collaboration of any kind requires communication between agents in order that they can coordinate their roles and the division of tasks as well as conveying information pertinent to the execution of those tasks. While not restricted to this topic preference will be given to proposals that are able to advance our understanding of how artificial systems can coordinate their actions among themselves and/or with humans in a human-understandable fashion.
SUBMISSION
Applications are approved on a first-come-first-served basis and the call will be open as long as there is funding available during the lifetime of the projects. However, we encourage early applications as we are planning to hold a hybrid workshop/panel on September 23, 2022, as part of an AIDA ICT48 event and would ask approved projects to give short presentations and participate in the workshop.
Applications should be made to both networks using the same title and description and referencing that the request falls under the joint Humane AI / NAiXUS call. The Humane AI application should be made through this form. If you have any queries concerning the call, please contact John Shawe-Taylor on jst@cs.ucl.ac.uk (Humane AI and NAiXUS).
Applications should be made to both networks using the same title and description and referencing that the request falls under the joint Humane AI / NAiXUS call. The Humane AI application should be made through this form. If you have any queries concerning the call, please contact John Shawe-Taylor on jst@cs.ucl.ac.uk (Humane AI and NAiXUS).
PROJECTS FOR MSc in AI
UCL has recently established two new MSc programmes in:
Each year the students on these courses will undertake projects that consider the application of AI to relevant application domain problems and datasets with a view to developing innovative solutions and novel core technologies. We are looking for external collaborators who are interested in defining challenging projects for our students and co-supervising their projects together with UCL academics, potentially also hosting the students at their institutions. The projects need to be proposed in the early autumn for students to make their choices over the new year and begin background research in the first months of the year with the main work of the project taking place between June and August.
AWS CREDITS PROGRAMME
As part of a broader partnership, the AWS Credit Program provides access to credits in AWS Promotional Credit, helping to offset costs for NAIXUS members with implementing cloud-based solutions. This helps meet mission goals without upfront investment in physical infrastructure.
IRCAI RESEARCH AND ACADEMIC VISITORS SCHEME
IRCAI offers a small number of opportunities for short and longer term research visits to the Centre. Our academic visitors scheme aims is to further enrich our research culture, develop and promote links with other universities and scholars, provide the staff and research students at IRCAI with the opportunity to learn from and work with scholars outside IRCAI, and offer external academics the opportunity to undertake research at IRCAI as well as initiate and continue collaborations with our staff. We welcome applications from PhD students, early (including postdoctoral researchers), mid and senior career academics from a variety of disciplines. We expect that all visiting fellows and scholars will be actively researching the “UNESCO mission”. An internal selection committee, will review and decide applications received. We will prioritise academic excellence, the quality of the research plan, and fit of the applicant’s research to IRCAI‘s expertise and interests. We aim to ensure a spread of visitors across disciplines and all five United Nations geographical regions: Africa, Europe and Americas, Asia and the Pacific, and the Middle East, as well as gender and seniority.
TRAVEL SCHEME
The meeting will be held at the Deep Learning Indaba in Accra Ghana most likely between the 8th and 9th of September, however the conference is from the 3rd to the 9th of September 2023. Deep Learning Indaba is the annual meeting of the African machine learning and AI community with the mission to Strengthen African AI.