Högskolan i Skövde

Autonomous vehicles promise to revolutionize society and improve the daily life of many, making them a coveted aim for a vast research community. To enable complex reasoning in autonomous vehicles, researchers are exploring new methods beyond traditional engineering approaches, in particular the idea of drawing inspiration from the only existing being able to drive: the human. The mental processes behind the human ability to drive can inspire new approaches with the potential to bridge the gap between artificial drivers and human drivers. In this review, we categorize and evaluate existing work on autonomous driving influenced by cognitive science, neuroscience, and psychology. We propose a taxonomy of the various sources of inspiration and identify the potential advantages with respect to traditional approaches. Although these human-inspired methods have not yet reached widespread adoption, we believe they are critical to the future of fully autonomous vehicles. 

Learning from only real-world collected data can be unrealistic and time consuming in many scenario. One alternative is to use synthetic data as learning environments to learn rare situations and replay buffers to speed up the learning. In this work, we examine the hypothesis of how the creation of the environment affects the training of reinforcement learning agent through auto-generated environment mechanisms. We take the autonomous vehicle as an application. We compare the effect of two approaches to generate training data for artificial cognitive agents. We consider the added value of curriculum learning—just as in human learning—as a way to structure novel training data that the agent has not seen before as well as that of using a replay buffer to train further on data the agent has seen before. In other words, the focus of this paper is on characteristics of the training data rather than on learning algorithms. We therefore use two tasks that are commonly trained early on in autonomous vehicle research: lane keeping and pedestrian avoidance. Our main results show that curriculum learning indeed offers an additional benefit over a vanilla reinforcement learning approach (using Deep-Q Learning), but the replay buffer actually has a detrimental effect in most (but not all) combinations of data generation approaches we considered here. The benefit of curriculum learning does depend on the existence of a well-defined difficulty metric with which various training scenarios can be ordered. In the lane-keeping task, we can define it as a function of the curvature of the road, in which the steeper and more occurring curves on the road, the more difficult it gets. Defining such a difficulty metric in other scenarios is not always trivial. In general, the results of this paper emphasize both the importance of considering data characterization, such as curriculum learning, and the importance of defining an appropriate metric for the task.

The paper investigates the potential of electric bikes as a sustainable alternative for local transportation during vacations in rural areas, replacing mainly transportation by car. Our starting point is that a key to behaviour change in the context of tourism and leisure travels is to make the sustainable travel option the most desirable option for the traveller. We study this by exploring three different electric bike offers in the area of Skaraborg, Sweden, and analysing the experience of the cyclists. 15 participants were invited to rent electric bicycles as mode of transportation for tourism in and around the small towns Lidköping and Skara for one day. Individual semi structured interviews were performed with the participants. Our main result is that there are aspects of electric bikes that make them particularly appropriate as a sustainable mode of transportation for local tourist destinations, and could also lower the threshold for more sustainable behaviours. Travelling by electric bike was experienced as beneficial in several ways; it automatically and effortlessly gives the cyclist access to nature, it constitutes a plausible option for more sustainable transportation at medium distances, and it allows the entire group of travellers, such as a families, to feel as being part of shaping the journey which contributes to engagement and motivation. Although intended for tourists, it also facilitate local people to access and reflect on their local flora and fauna. We also found that digital tools such as maps, information sources, and booking systems have a key part in the deployment when appropriately integrated. In addition, we found that it was important to lower the threshold of trying electric bikes, and an important key for this was personal service from proficient people who could provide a sense of security by giving instructions, answering questions, and support in adjusting the equipment.

Digitalization is forcing the industry to rethink current practices in all business domains, pushing for a digital transformation of business and operations at a high rate and, thus, paving the way for new business models and making others redundant. For small and medium-sized companies (SME), in particular, it is an enormous challenge to keep up with the pace of technological development. Several initiatives have argued the industry’s need for continuous digitalization, innovation, transformation ability, and future skills and competencies development. However, the advancement of the Swedish industry in this area has been uneven, where larger organizations have begun their digital transformation journey to some extent, but SMEs risk falling behind. In addition to the technological transformation, the challenges regarding the industries’ skills supply need to be solved, where a workforce with the right competencies, knowledge, and skill sets are equally, if not more, important for remaining competitive. One of the key elements to face these challenges in the companies will be to recruit knowledgeable employees or re-skill the existing ones. Efficient access to relevant knowledge and skills is still a major concern for companies that will surely affect their competitiveness for a long time to come. This paper elaborates on the opportunities and challenges that Swedish universities face in the context of lifelong learning and education for industry professionals. The paper presents results and experiences gained from a lifelong learning project for industry professionals at the University of Skövde in collaboration with ten industry partners. The results from the project show that in addition to pedagogical methods, current structures and policies within academia need to be further developed to effectively serve industry professionals. The paper also presents a concept of education for industry professionals in the lifelong learning context based on the results and experience gained from the project.

Self-driving cars not only solve the problem of navigating safely from location A to location B; they also have to deal with an abundance of (sometimes unpredictable) factors, such as traffic rules, weather conditions, and interactions with humans. Over the last decades, different approaches have been proposed to design intelligent driving systems for self-driving cars that can deal with an uncontrolled environment. Some of them are derived from computationalist paradigms, formulating mathematical models that define the driving agent, while other approaches take inspiration from biological cognition. However, despite the extensive work in the field of self-driving cars, many open questions remain. Here, we discuss the different approaches for implementing driving systems for self-driving cars, as well as the computational paradigms from which they originate. In doing so, we highlight two key messages: First, further progress in the field might depend on adapting new paradigms as opposed to pushing technical innovations in those currently used. Specifically, we discuss how paradigms from cognitive systems research can be a source of inspiration for further development in modeling driving systems, highlighting emergent approaches as a possible starting point. Second, self-driving cars can themselves be considered cognitive systems in a meaningful sense, and are therefore a relevant, yet underutilised resource in the study of cognitive mechanisms. Overall, we argue for a stronger synergy between the fields of cognitive systems and self-driving vehicles.

We consider the benefits of dream mechanisms - that is, the ability to simulate new experiences based on past ones - in a machine learning context. Specifically, we are interested in learning for artificial agents that act in the world, and operationalize "dreaming" as a mechanism by which such an agent can use its own model of the learning environment to generate new hypotheses and training data. We first show that it is not necessarily a given that such a data-hallucination process is useful, since it can easily lead to a training set dominated by spurious imagined data until an ill-defined convergence point is reached. We then analyse a notably successful implementation of a machine learning-based dreaming mechanism by Ha and Schmidhuber (Ha, D., & Schmidhuber, J. (2018). World models. arXiv e-prints, arXiv:1803.10122). On that basis, we then develop a general framework by which an agent can generate simulated data to learn from in a manner that is beneficial to the agent. This, we argue, then forms a general method for an operationalized dream-like mechanism. We finish by demonstrating the general conditions under which such mechanisms can be useful in machine learning, wherein the implicit simulator inference and extrapolation involved in dreaming act without reinforcing inference error even when inference is incomplete.

This paper presents a novel approach to learning predictive motor control via mental simulations. The method, inspired by learning via mental imagery in natural Cognition, develops in two phases: first, the learning of predictive models based on data recorded in the interaction with the environment; then, at a deferred time, the synthesis of inverse models via offline episodic simulations. Parallelism with human-engineered control-theoretic workflow (mathematical modeling the direct dynamics followed by optimal control inversion) is established. Compared to the latter human-directed synthesis, the mental simulation approach increases autonomy: a robotic agent can learn predictive models and synthesize inverse ones with a large degree of independence. Human modeling is still needed but limited to providing efficient templates for the forward and inverse neural networks and a few other directives. One could consider these templates as the efficient brain network typologies that evolution produced to permit live beings quickly and efficiently learning. The structure of the neural networks both forward and inverse ones; is made of interpretable local models which follows the cerebellar organization (and are also similar to local model approaches known in the literature). We demonstrate the learning of a first-round model (contrasted to Model Predictive Control) for lateral vehicle dynamics. Then, we demonstrate a second learning iteration, where the forward/inverse neural models are significantly improved.

The controller of an autonomous vehicle needsthe ability to learn how to act in different driving scenariosthat it may face. A significant challenge is that it is difficult,dangerous, or even impossible to experience and explore variousactions in situations that might be encountered in the realworld. Autonomous vehicle control would therefore benefitfrom a mechanism that allows the safe exploration of actionpossibilities and their consequences, as well as the ability tolearn from experience thus gained to improve driving skills.In this paper we demonstrate a methodology that allows alearning agent to create simulations of possible situations. Thesesimulations can be chained together in a sequence that allowsthe progressive improvement of the agent’s performance suchthat the agent is able to appropriately deal with novel situationsat the end of training. This methodology takes inspirationfrom the human ability to imagine hypothetical situations usingepisodic simulation; we therefore refer to this methodology asepisodic imagination.An interesting question in this respect is what effect thestructuring of such a sequence of episodic imaginations hason performance. Here, we compare a random process to astructured one and initial results indicate  that a structuredsequence outperforms a random one.

I Högskolan i Skövdes (HS) utvecklingsplan (Högskolan i Skövde, 2017) framhålls vikten av att samtliga utbildningsprogram skall erbjuda möjligheter till utbytesstudier. Historiskt sett har antalet utresande vid HS varit lågt samtidigt som fakta som belyser varför så är fallet främst har baserats på enskilda studenters upplevelser av nominerings- och antagningsprocesser. Dock saknas en mer enhetlig och generaliserbar vy av hur olika funktioner på HS rörande utbytesstudier upplevs, både av studenter som genomfört utbytesstudier samt av studenter som varit nominerade men valt att inte genomföra utbytesstudier.

Den frågeställning som adresserats är: vilka hinder finns för studenter som vill genomföra utbytesstudier? Den metod som tillämpats är en enkät som skickats ut till samtliga studenter som varit nominerade för utbytesstudier under 2017. Enkäten har baserats på en Likert-skala (Bryman och Nilsson) som kombinerats med fritextsvar.

Totalt har 20 studenter från olika utbildningsprogram besvarat enkäten. Studien visar att studenterna rent generellt är nöjda med HS administrativa funktioner vilket också är den kanal som oftast utnyttjas. Andra administrativa funktioner såsom programansvariga och ämnesföreträdare upplevs olika av studenterna beroende på vilket ämnesområde som berörs.

Studenterna redovisar ett antal huvudsakliga skäl som hindrar eller försvårar utbytesstudier. 1) Det läggs ett alltför stort ansvar på studenten i att identifiera kurser som denne är behörig till och som inte överlappar med kurser som läses vid det egna lärosätet. Detta innefattar även val av kurser som skall motsvara obligatoriska kurser på hemmaplan som i sin tur ligger som förkunskapskrav för framtida kurser inom utbildning, såsom examensarbeten. 2) Avsaknad av utbytesavtal med specifika länder samt svårighet att identifiera avtal med lärosäten som matchar studentens utbildningsprofil vid det egna lärosätet upplevs likaså som ett hinder för flertalet studenter. 3) Vidare uppger studenterna att det i många fall är svårt att ta till sig information på olika lärosätens hemsidor för att identifiera passande kurser eftersom den information som finns att tillgå upplevs som mycket heterogen och därmed både svårtolkad och svårnavigerad. Detta kan jämföras mot hur svenska lärosäten publicerar information om kurser vilket av studenterna upplevs som mer enhetligt och standardiserat vilket i sin tur gör olika kurser enklare att jämföra.

Något som efterfrågas av många studenter som varit nominerade för utbytesstudier men som i slutändan valt att inte resa är ”utbytesstudie-charter” där det finns färdiga kurspaket som är granskade och validerade av HS. Även önskemål om mer standardiserade paket för boende och andra praktiska frågor efterfrågas av studenterna.

Den information som enkätstudien har lyft fram har i första hand visat på att HS administrativa funktioner för utbytesstudier fungerar tillfredställande. Likaså verifierar enkätstudien de hypoteser som funnits rörande hindrande faktorer för utbytesstudier i de flesta fall stämmer: osäkerhet kring val av kurser och en oro för hur utbytesstudier kommer att påverka framtida studier på hemmaplan är stora orosmoment som kan få studenter av avstå från utbytesstudier trots ett initialt intresse för sådana.

Referenser

Bryman, A., & Nilsson, B. (2011). Samhällsvetenskapliga metoder. Malmö: Liber.

Högskolan i Skövde. (2017). Utvecklingsplan. Skövde.

I Högskolan i Skövdes (HS) nuvarande utvecklingsplan beskrivs att samtliga utbildningsprogram som ges vid lärosätet skall erbjuda möjligheter till studier utomlands. Som en konsekvens har en strategisk satsning för att stimulera en ökad mobilitet vid HS initierats genom att Institutionen för informationsteknologi (IIT) tillsammans med verksamhetsstödet har deltagit i ett UHR-projekt med fokus på vägledningsprocessen i samband med mobilitet.

Projektet har pågått under 18 månader och ett 10-tal deltagare som representerar de flesta roller som är involverade i internationaliseringsarbetet på IIT och verksamhetsstödet har ingått. Projektet har bedrivits främst som en serie av workshoppar där parallell datainsamling har skett med hjälp av intervjuer och enkätstudier.

I detta bidrag vill vi visa några av de mål som projektet fokuserat på samt syftet med dessa. Projektet har haft följande mål:

Identifiera minst tre partnerlärosäten som passar varje utbildningsprogram - Målet syftar till att kartlägga både existerande och nya partnerlärosäte för att på så sätt sänka tröskeln för studenter som är intresserade av utbyte, men som har svårt att hitta lämpliga alternativ.

Tydliggöra roll- och ansvarsfördelning i mobilitetsprocessen - Detta mål syftar till att utforma processbeskrivningar för att tydliggöra roll- och ansvarsfördelning kring utresandeprocessen för programansvarig, ämnesföreträdare, internationell koordinator, studie- och karriärvägledaren, studenten och partnerlärosätet. Även kommunikationsaspekter och studentperspektiv beaktas i detta mål.

Skapa adekvat vägledning och informationsinsatser gentemot studenterna - Syftet med målet är att utveckla, strukturera och systematisera informationsvägar och kommunikation mellan vägledning, programansvarig, partnerlärosäte och studenter.

Identifiera och utvärdera nyckelfaktorerna för att förbättra stödet till studenterna, undanröja hinder i mobilitetsprocessen samt underlätta programansvarigas och studie- och karriärvägledarnas arbete med utresande studenter - Målet är att identifiera och utvärdera nyckelfaktorer som hindrar mobilitet som kan spridas internt på HS samt externt för att i förlängningen öka mobiliteten bland Sveriges studenter.

Se till att samtliga kandidatprogram på IIT har en termin avsatt för utlandsstudier och undanröja de hinder som finns i befintliga programstrukturer - Syftet är att göra det enklare för studenter att under sin ordinarie studietid genomföra utbytesstudier utan att deras studier vid HS blir drabbade av förkunskapsstrukturer som hindrar fortsatta studier vid hemkomst.

I studien inkludera de studenter som har varit intresserade eller sökt utbytesstudier men som inte kommit iväg på utbyte - Syftet är att skapa en god översikt över vilka skäl denna studentkategori har haft för att avstå utbytesstudier för att på så vis kunna förbättra existerande processer för utbytesstudier och därmed minimera risken att intresserade studenter väljer att avstå från utbytesstudier.

Vid projektets avslut analyserades projektets direkta och indirekta interorganisatoriska effekter tillsammans med aktuell statistik för de studenter som under 2017 nominerats för utbytesstudier. En uppenbar effekt av projektets arbete är att antalet studenter som 2017 vid IIT nominerats för utbytesstudier ökat kraftigt. Likaså observeras ökade kunskaper om processen kring utbytesstudier och ett förbättrat studentperspektiv och bättre kunskaper om studenternas upplevelse av nomineringsprocessen.

Under presentation vill vi visa upp fler detaljer från vår analys samt hur vi planerar att arbeta vidare med de resultat vi fått fram i projektet.