One of the most pressing scientific challenges today is understanding the fate of our oceans and marine ecosystems under on-going climate change. Unfortunately, anthropogenic stressors act at a rate and magnitude that exceed recent natural variability, making the use of decadal ecological data and time-series insufficient for predictions of future behaviour of marine ecosystems. MICRO2MACRO will reconstruct snapshots of marine pelagic ecosystems between 54 and 32 million years ago (Eocene and early Oligocene), when climate and environmental conditions approximated what we will start to experience in the next century and beyond. Using the microfossil record of planktonic foraminifera (PF), the most complete of any Cenozoic eukaryote, the project will generate the first methodologically controlled (hence reproducible) early Cenozoic global dataset of ecology, abundance, species composition, diversity and biogeography (macroecology) of these prolific pelagic calcifiers. Benefiting from the mole of data generated over the last 15 years, I will learn and apply novel tools in data-science technology to compile ocean temperature and chemistry datasets for the studied time intervals and statistically compare them against the new PF dataset generated with this project. This study will combine the most advanced knowledge in several disciplines (micropaleontology, informatics, statistical ecology) to test for links between time-specific climate (e.g., sea surface temperatures) and ecosystem (e.g., species composition, dominant ecology) configurations, and understand how plankton biogeography was shaped in a warmer world. Hence, MICRO2MACRO will highlight future ecological and evolutionary analogues if the current climate trajectory remains interrupted and we are to hit climate conditions similar to those in the Eocene and Oligocene. Given the uncertainties associated with projections based on modern data this study will represent a major advancement in the field.

The realization of computational models for accomplishing everyday manipulation tasks for any object and any purpose would be a disruptive breakthrough in the creation of versatile, general-purpose robot agents; and it is a grand challenge for AI and robotics. Humans are able to accomplish tasks such as “cut up the fruit” for many types of fruit by generating a large variety of context-specific manipulation behaviors. They can typically accomplish the tasks on the first attempt despite uncertain physical conditions and novel objects. Acting so effectively requires comprehensive reasoning about the possible consequences of intended behavior before physically interacting with the real world. In the FAME project, I will investigate the research hypothesis that a knowledge representation and reasoning (KR&R) framework based on explictly-represented and machine-interpretable inner-world models can enable robots to contextualize underdetermined manipulation task requests on the first attempt. To this end, I will design, implement, and evaluate FAME (Future-oriented cognitive Action Modelling Engine), a hybrid symbolic/subsymbolic KR&R framework that will contextualize actions by reasoning symbolically in an abstract and generalized manner but also by reasoning with “one’s eyes and hands” through mental simulation and imagistic reasoning. Realizing FAME requires three breakthrough research results: (1) modelling and parameterization of manipulation motion patterns and understanding the resulting effects under uncertain conditions; (2) the ability to mentally simulate imagined and observed manipulation tasks to link them to the robot’s knowledge and experience; and (3) the on-demand acquisition of task-specific causal models for novel manipulation tasks through mental physics-based simulations. To assess the power and feasibility of FAME, I will use open manipulation task learning as a benchmark challenge.

The emerging and vibrant area of ontology-based data access (OBDA) is currently establishing itself as an important paradigm for processing incomplete and heterogeneous data. The goal of the CODA project is to make OBDA radically more useful for real-world applications by taking a ground-breaking new perspective on its foundations, algorithms, and tools. The project will rest on an ultimately fine-grained complexity analysis that allows to identify islands of tractability inside practically important ontology and query languages that are otherwise intractable. Based on these islands, novel OBDA querying tools will be developed that are custom-made for ontologies from applications in the sense that high computational cost is incurred only when unavoidable for the concrete ontology used (`pay as you go' behaviour). The key deliverables of the project are a set of tailor-made OBDA querying tools that form a precision tool belt for real-world OBDA applications, theoretical results regarding the structure and computational complexity of important islands of tractability, efficient algorithms that allow to put these to work in practice, and optimization techniques and heuristics that support the algorithms in the tools developed. We will also collect and make available a a library of case studies for evaluating OBDA tools. The project is both timely and essential. It is timely because our economy and society are currently experiencing a revolution in data processing and availability, and dealing with incompleteness and heterogeneity is one of the major arising challenges. The project is essential because it has become apparent now that current OBDA tools cannot satisfy industry requirements. In particular, they do not adequately support the limited use of expressive features (`a little bit of disjunction') which intuitively should not result in high computational cost, but with current technology often does.