Recent progress in artificial intelligence (AI) has been mostly due to machine learning and, in particular, deep artificial neural networks (ANNs). Deep learning has an increasing presence in everyday life, including critical applications such as medical diagnosis, transportation, and energy distribution. In response to this, the field of Explainable AI (XAI) has generated much effort in terms of techniques and algorithms to address this problem. However, there is still no consensus on a suite of technology to address these challenges, progress has been extremely limited, and the formal properties of such systems are under-studied. On the other hand, computational neuroscience (CNS) aims to discover the principles behind biological neural networks that enable the brain to support cognition, perception, and action. This project will employ the latest approaches and techniques used in the field of CNS to develop the field of XAI. Specifically, the first major goal will be to employ the methods of representational geometry and neural encoding manifolds (both proven to be effective in revealing meaningful neural relationships in previous studies) to reveal how activations of collections of artificial neurons in hidden layers are associated with the decision-making process of deep networks. Second, the same methodology will be used to reveal novel insights from a variety of existing large-scale biological datasets. Finally, we will compare and contrast the encoding strategies of neural populations found various deep learning architectures with those observed in biological networks. A better understanding of the inner-workings of biological models could directly inform researchers on how to build novel artificial models that are more accurate, robust, and even economical during both training and inference in terms of data, time, and energy consumption.

When standard medications are not appropriate for a patient’s needs, e.g., due to their allergies, retail pharmacies must rely on compounding pharmacies to blend raw ingredients and produce a personalized medication to order. Given the growth in demand for personalized medications and the legal mandate in some countries that compel pharmacies to provide such medications to patients, the efficient operation of compounding pharmacies is critical to timely access to medications. However, there are complex operational dynamics when sequencing production driven by differences in medications and the need to prevent cross-contamination that lead to production delays for patients. Inspired by discussions with the management team of a compounding pharmacy, this project aims to improve operational efficiency and reduce delays by developing a dynamic production control algorithm that sequences medication production. The challenge of identifying optimal policies for multi-product production systems in the presence of set-up times has lead researchers to focus on heuristics, however, existing policies do not account for the sequence-dependent set-up times or batch processing in this setting. We propose and evaluate a theory-driven heuristic based on a novel modification of an optimal control engineering technique, known as the State-Dependent Riccati Equation approach. The performance of this heuristic is to be evaluated both theoretically as well as numerically relative to alternative heuristic policies via simulation. This contributes to the operations management and management literature, through the development and analysis of an innovation in production governance, a critical component of the industrial value chain in this setting. Moreover, the proposed algorithm can be modified for use in other complex production settings where optimal policies are intractable, and decision-makers must rely on heuristics.

People are increasingly moving across national borders, where many of these people are refugees fleeing natural disasters, war, and persecution. This proposed project will investigate the impact of refugee-run small-scale businesses on resilience of marketing systems within refugee settlements, where resilience is the ability of a system to recover in the face of disturbances. This project will specifically seek to understand the resource sets in the marketing system in relation to resilience through employing an integrated capitals framework that includes nine resources (financial, physical, social, natural, human, cultural, public, political, and health) and bridges seven existing capitals frameworks from academia and practice. This work will be qualitative in nature: The researcher will engage in longitudinal fieldwork to collect observational and interview-based data in a refugee settlement in Europe and inductively analyse this data towards the creation of theoretical models. The resulting theory will further academic understanding of resilience, marketing systems, and involved resources, while aiming to create actionable models that can be used to strengthen economic activities related to refugees. This work aligns with the European Commission’s (EC) 2018-2020 Work Programme research priorities related to social and economic effects of migration (SC6). Further, the running of small businesses is considered by the EC to be of key importance towards supporting refugees’ integration in Europe, while resilience is named as a part of the United Nation’s Sustainable Development Goals related to communities. As such, furthering our understanding of refugee-run businesses in relation to resilience within marketing systems is important to the refugees themselves, their host countries, and the public and private sector, and thus to the future of Europe and for the wider global community.

Among the lessons learned from the recent pandemic is the importance of embracing the unconventional. Indeed, atypical approaches have proved to be a key strategic resource for coping with unpredictable events, promoting socio-cultural transformation, and expanding the capacity to deliver valued goods and services. However, although a departure from what is a standard positioning within socio-cultural domains may open up pathways to exceptional impact, existing research shows that atypical ideas, products, or actors generate mistrust and encourage rejection as they are subject to skepticism and face a higher risk of being considered ambiguous or illegitimate. This project's aim is to study how narratives can support the reception of atypicality and stimulate its diffusion. Narratives are powerful means of sustaining the alignment between actions and established conventions, especially in the face of situations that depart from expectations or definitions of what is considered contextually normal. NARR-ATY-VES (Narratives for Atypicality) aims to 1) develop an understanding of the role of narratives' content and structure in promoting and supporting the evaluation of atypicality and 2) introduce an innovative computational approach based on cutting-edge Natural Language Processing (NLP) techniques to model two fundamental dimensions of narratives, namely their contextual embeddedness and their sequential unfolding based on cutting-edge Natural Language Processing (NLP) techniques, to map narrativity, i.e., the set of narrative elements that characterize the complex nature of narratives. Through this hands-on 2-year research project, the Researcher's knowledge and skill will be enhanced enormously, and he shall be en route to becoming an expert in this field.