Over the past decade, researchers in Artificial Intelligence (AI) have made ground-breaking progress on long-standing problems. Now that AI is becoming increasingly part of our daily lives, we need to avoid being ruled by machines and their decisions. Hybrid Intelligence (HI) is the combination of human and machine intelligence, expanding human intellect instead of replacing it. It takes human expertise and intentionality into account when making meaningful decisions and perform appropriate actions, together with ethical, legal and societal values. Our goal is to design Hybrid Intelligent systems, an approach to Artificial Intelligence that puts humans at the center, changing the course of the ongoing AI revolution. By providing intelligent artificial collaborators that interact with people we amplify human capacity for learning, reasoning, decision making and problem solving. The challenge is to build intelligent systems that augment and amplify rather than replace human intelligence, that leverage our strengths and compensate for our weaknesses. Such Hybrid Intelligence requires meaningful interaction between artificial intelligent agents and humans to negotiate and align goals, intentions and implications of actions. Developing HI needs fundamentally new solutions to core research problems in AI: current AI technology surpasses humans in many pattern recognition and machine learning tasks, but it falls short on general world knowledge, common sense, and the human capabilities of (i) collaboration, (ii) adaptivity, (iii) explanation and (iv) awareness of norms and values. These challenges will be addressed in four interconnected research lines: Collaborative HI: How to design and build intelligent agents that work in synergy with humans, with awareness of each others strengths and limitations? Adaptive HI: HI systems will need to operate in situations not anticipated by their designers, and cope with variable team configurations, preferences and roles. Explainable HI: Intelligent agents and humans need to be able to mutually explain to each other what is happening (shared awareness), what they want to achieve (shared goals), and what collaborative ways they see of achieving their goals (shared plans and strategies Responsible HI: Values such as transparency, accountability, trust, privacy and fairness be an integral part of the design and operation of HI systems. Applications in healthcare, education and science will demonstrate the potential of Hybrid Intelligence: virtual agents and robots will help children with concentration problems to study better; virtual agents and robots will support children in paediatric oncology wards by providing them with entertainment and information during prolonged hospital stays; virtual agents will collaborate with scientists on large scale analysis of the literature, formulate new hypotheses and help design experiments to test them. The team brings together top AI researchers from across the Netherlands in machine learning, knowledge representation, natural language understanding & generation, multi-agent systems, human collaboration, cognitive psychology, multimodal interaction, social robotics, AI & law and ethics of technology. We will initiate a Hybrid Intelligence Centre (HI Centre) to host joint research facilities, multidisciplinary PhD programs, and training and exchange programs. Sustainability of the HI Centre is ensured through tenure track positions with guaranteed long-term funding from the participating universities.

Robots are increasingly part of our daily lives. They help provide medical care, care for our homes and gardens, and support education and the workplace. They are increasingly able to perform simple tasks and convey messages, but they also lack many skills needed for interaction and communication. Dramaturgy for Devices will develop these skills in collaboration with theater and dance makers. We will show how the performing arts can contribute to innovative design tools and methods, and show the value of the skills and knowledge of theater practice for technological innovation.

The success of the Web of Data (WOD) is based on the thorough understanding of, and agreement upon, the semantics of data and ontologies. But the Web of Data as a whole is complex, and inherently messy, contextualised, opinionated, in short: the Web of Data is a market-place of ideas, rather than a database. Existing paradigms are inappropriate for dealing with this new type of knowledge structure. The urgency of dealing with the non-standard characteristics of the Web of Data has been recognised, and separate initiatives try to tackle its individual manifestations, e.g. inconsistencies, contexts, vagueness, provenance, etc. Tomorrow?s Web requires novel semantics with efficient (generic) implementations to ensure semantic clarity, reuse and interoperability. We recently introduced pragmatic semantics [29] as a new semantic paradigm integrating elements from market theory and classical semantics into a framework of optimisation over partial truth functions, each representing a particular world-view. We propose nature-based algorithms to implement those semantics. The main results will be 1) Well-understood formalisms for pragmatically interpreting semantic data on the Web. 2) Algorithms to calculate the related inferences, such as entailment and truth, and guidelines for using the new paradigm in application domains. 3) A system integrating various implementations of generic and specific multi-optimisation methods, implementations of a variety of partial truth-functions, and representation standards for semantic meta-data. 4) A thorough analytic and empirical evaluation of the proposed semantic paradigm.

This project aims to enhance understanding of how to develop human-centric AI with common sense. It will transform the development of mental modeling, solution verification, figure-of-speech identification, explanation of behaviors through cultural values, and coherent visual reasoning, demonstrating success in controlled lab experiments complemented by human studies in critical domains.

Scientific, technical and medical knowledge is built on research data. It increasingly plays a similar role in the social sciences and humanities. Research datasets are either deposited by researchers or automatically extracted from publications. We propose to create open source search and recommendation solutions for research datasets so as to enable their re-use. The main benefit is that datasets can be more easily found. This way, data re-use is stimulated and redundancy in data collection is avoided. Situated at the interface between the philosophy of science and computer science, the development of innovative algorithmic solutions will be informed by combining three perspectives. First, we will examine the use of datasets in publications, in different disciplines, and for different research tasks, to understand to which extent scientific discovery is based on data-availability and how it is affected by data-sharing cultures. Second, we will contribute semantic technologies to support dataset search, to match research data with user groups, and to generate research dataset search engine result pages. Third, we will develop information retrieval algorithms for unsupervised dataset search and predicting user interactions with dataset search engine results. We will combine these into a self-learning method for searching datasets. Our solutions will be implemented in Elseviers retrieval and recommendation environments. The project will engage the data science community through co-design workshops at critical stages in the research planning, through regular participation in data science and search engine meetups, and by releasing its algorithmic solutions as open source.