Dealing with large amounts of unlabeled data is a very challenging task. Recently, many different approaches have been proposed to leverage this data for training many machine learning models. Among them, self-supervised learning appears as an efficient solution capable of training powerful and generalizable models. More specifically, instead of relying on human-generated labels, it proposes training objectives that use ``labels'' generated from the data itself, either via data augmentation or by masking the data in some way and trying to reconstruct it. Apart from being able to train models from scratch, self-supervised methods can also be used in specific applications to further improve a pre-trained model. In this thesis, we propose to leverage self-supervised methods in novel ways to tackle different application scenarios. We present four published papers: an open-source library for self-supervised learning that is flexible, scalable, and easy to use; two papers tackling unsupervised domain adaptation in action recognition; and one paper on self-supervised learning for continual learning. The published papers highlight that self-supervised techniques can be leveraged for many scenarios, yielding state-of-the-art results.

In order to enable more widespread application of robots, we are required to reduce the human effort for the introduction of existing robotic platforms to new environments and tasks. In this thesis, we identify three complementary strategies to address this challenge, via the use of imitation learning, domain adaptation, and transfer learning based on simulations. The overall work strives to reduce the effort of generating training data by employing inexpensively obtainable labels and by transferring information between different domains with deviating underlying properties. Imitation learning enables a straightforward way for untrained personnel to teach robots to perform tasks by providing demonstrations, which represent a comparably inexpensive source of supervision. We develop a scalable approach to identify the preferences underlying demonstration data via the framework of inverse reinforcement learning. The method enables integration of the extracted preferences as cost maps into existing motion planning systems. We further incorporate prior domain knowledge and demonstrate that the approach outperforms the baselines including manually crafted cost functions. In addition to employing low-cost labels from demonstration, we investigate the adaptation of models to domains without available supervisory information. Specifically, the challenge of appearance changes in outdoor robotics such as illumination and weather shifts is addressed using an adversarial domain adaptation approach. A principal advantage of the method over prior work is the straightforwardness of adapting arbitrary, state-of-the-art neural network architectures. Finally, we demonstrate performance benefits of the method for semantic segmentation of drivable terrain. Our last contribution focuses on simulation to real world transfer learning, where the characteristic differences are not only regarding the visual appearance but the underlying system dynamics. Our work aims at parallel training in both systems and mutual guidance via auxiliary alignment rewards to accelerate training for real world systems. The approach is shown to outperform various baselines as well as a unilateral alignment variant.

The ultimate goals of adaptive serious educational games (adaptive SEG) are to promote effective learning and maximising enjoyment for players. Firstly, we develop the SEG by combining knowledge space (learning materials) and game content space to be used to convey learning materials. We propose a novel approach that serves toward minimising experts' involvement in mapping learning materials to game content space. We categorise both content spaces using known procedures and apply BIRCH clustering algorithm to categorise the similarity of the game content. Then, we map both content spaces based on the statistical properties and/or by the knowledge learning handout. Secondly, we construct a predictive model by learning data sets constructed through a survey on public testers who labelled their in-game data with their reported experiences. A Random Forest algorithm non-intrusively predicts experiences via the game data. Lastly, it is not feasible to manually select or adapt the content from both spaces because of the immense amount of options available. Therefore, we apply reinforcement learning technique to generate a series of learning goals that promote an efficient learning for the player. Subsequently, a combination of conditional branching and agglomerative hierarchical clustering select the most appropriate game content for each selected education material. For a proof-of-concept, we apply the proposed approach to producing the SEG, named Chem Dungeon, as a case study to demonstrate the effectiveness of our proposed methods.

In this study we aim to investigate what employees at a callcenter see as prospects, opportunities and barriers of learning within a company. To answer the purpose of this essay, two questions were formulated; "What opportunities for learning does employees experience in the company?" and "How can learning in the workplace improve for the sellers?". First, we present what previous research has highlighted regarding workplace learning and learning within the callcenter industry. Afterwards we present ten semi-structured interviews with leaders and sellers within a telecom company in Stockholm. The material from the interviews was analyzed on the basis of previous research and theories in socio-cultural learning. The result shows that there are many conditions and opportunities for learning within the company and that the learning promoted in the workplace, is largely shaped in adaptive oriented way. To further promote the learning of the employees requires a development of their acting-space and the organizational culture within the business. / I denna studie syftar vi till att undersöka vad medarbetare vid ett callcenter ser för förutsättningar, möjligheter samt hinder för lärande inom ett företag. För att besvara syftet har två frågeställningar formulerats; "Vilka förutsättningar för lärande upplever medarbetare inom företaget?" och "Hur kan lärandet på arbetsplatsen förbättras för säljarna?". Först presenterar vi vad tidigare forskning har belyst angående arbetsplatslärande och lärande inom callcenterbranschen. Vi genomförde därefter tio stycken semistrukturerade intervjuer med ledare och säljare på ett Telecom bolag i Stockholm. Materialet analyserades därefter utifrån tidigare forskning och teorier inom det sociokulturella lärandet. Resultatet visar att det finns många förutsättningar och möjligheter för lärande inom företaget samt att lärandet som främjas på arbetsplatsen är till största del av anpassningsinriktad form. För att ytterligare främja lärandet hos medarbetarna krävs en utveckling av handlingsutrymmet och organisationskulturen inom verksamheten.

Agentes atuando em sociedade devem agir de maneira coerente para atingir um objetivo comum. A coordenação nos sistemas multiagente previne o comportamento caótico dos agentes, permite que o sistema lide com restrições globais e a interdependência entre os agentes, e faz com que o sistema possa ser composto por agentes com diferentes competências. A coordenação pode ser baseada na estrutura organizacional, onde a comunidade de agentes atua a favor de um objetivo comum através da forma como estão organizados. Em ambientes dinâmicos a organização dos agentes deve se adaptar a mudanças nos objetivos do sistema, na disponibilidade de recursos, nos relacionamentos entre os agentes, e assim por diante. Esta flexibilidade é um problema chave nos sistemas multiagente e está relacionada a modelos de adaptação como os observados nos insetos sociais. O presente trabalho propõe uma abordagem para a geração e adaptação da organização de um sistema multiagente, em tempo de execução, utilizando como base os modelos teóricos de organização das colônias de insetos sociais. Esta abordagem enfoca a alocação e o escalonamento dinâmicos de tarefas distribuídos entre agentes com diferentes competências e em ambientes de larga escala. Dois cenários principais são utilizados para experimentar e validar a abordagem proposta. Estes cenários estão baseados em problemas de pesquisa operacional denominados Resource-Constrained Project Scheduling Problem (RCPSP) e o Generalized Assignment Problem (GAP). Este trabalho contribui para o avanço do estado-da-arte no estudo e desenvolvimento de sistemas multiagente e na modelagem e aplicação de técnicas de inteligência de enxames em problemas computacionais. A abordagem proposta para coordenação de agentes em cenários complexos é nova, eficaz e robusta. De maneira geral, esta abordagem contribui para busca da solução de problemas de coordenação de sistemas multiagente aplicados a problemas reais. / A community of individual agents must work in a coherent manner to reach some common goal. The coordination process in multiagent systems prevents chaotic behavior of agents, makes the system able to deal with global constraints and inter-agents dependencies, and allows the system to be composed of agents with different capabilities. This process is normally based on the organizational structure, where the community of agents works towards the system goal through the manner they are organized. However, in dynamic environments, agents must be able to adapt to the changing goals of the system, to the resources available, to their relationships with another agents, to changes on the environment and so on. This problem is a key one in multiagent systems and relates to models of adaptation, such as those observed among social insects. This work proposes a new approach to generate and adapt the multiagent organization on the fly based on the theoretical models of social insects colonies organization. This approach focuses on distributed dynamic scheduling and task allocation using agents with different capabilities in large scale environments. Two main scenarios have been used to experiment and validate the proposed approach: the Resource-Constrained Project Scheduling Problem (RCPSP) and the Generalized Assignment Problem (GAP). This work contributes to advancing the state-of-the-art in the study and development of multiagent systems and in the modeling and application of swarm intelligence techniques. The proposed approach to coordinate agents in complex scenarios is novel, effective and robust. This approach contributes to the search of coordination solutions to multiagent systems real applications.

Agentes atuando em sociedade devem agir de maneira coerente para atingir um objetivo comum. A coordenação nos sistemas multiagente previne o comportamento caótico dos agentes, permite que o sistema lide com restrições globais e a interdependência entre os agentes, e faz com que o sistema possa ser composto por agentes com diferentes competências. A coordenação pode ser baseada na estrutura organizacional, onde a comunidade de agentes atua a favor de um objetivo comum através da forma como estão organizados. Em ambientes dinâmicos a organização dos agentes deve se adaptar a mudanças nos objetivos do sistema, na disponibilidade de recursos, nos relacionamentos entre os agentes, e assim por diante. Esta flexibilidade é um problema chave nos sistemas multiagente e está relacionada a modelos de adaptação como os observados nos insetos sociais. O presente trabalho propõe uma abordagem para a geração e adaptação da organização de um sistema multiagente, em tempo de execução, utilizando como base os modelos teóricos de organização das colônias de insetos sociais. Esta abordagem enfoca a alocação e o escalonamento dinâmicos de tarefas distribuídos entre agentes com diferentes competências e em ambientes de larga escala. Dois cenários principais são utilizados para experimentar e validar a abordagem proposta. Estes cenários estão baseados em problemas de pesquisa operacional denominados Resource-Constrained Project Scheduling Problem (RCPSP) e o Generalized Assignment Problem (GAP). Este trabalho contribui para o avanço do estado-da-arte no estudo e desenvolvimento de sistemas multiagente e na modelagem e aplicação de técnicas de inteligência de enxames em problemas computacionais. A abordagem proposta para coordenação de agentes em cenários complexos é nova, eficaz e robusta. De maneira geral, esta abordagem contribui para busca da solução de problemas de coordenação de sistemas multiagente aplicados a problemas reais. / A community of individual agents must work in a coherent manner to reach some common goal. The coordination process in multiagent systems prevents chaotic behavior of agents, makes the system able to deal with global constraints and inter-agents dependencies, and allows the system to be composed of agents with different capabilities. This process is normally based on the organizational structure, where the community of agents works towards the system goal through the manner they are organized. However, in dynamic environments, agents must be able to adapt to the changing goals of the system, to the resources available, to their relationships with another agents, to changes on the environment and so on. This problem is a key one in multiagent systems and relates to models of adaptation, such as those observed among social insects. This work proposes a new approach to generate and adapt the multiagent organization on the fly based on the theoretical models of social insects colonies organization. This approach focuses on distributed dynamic scheduling and task allocation using agents with different capabilities in large scale environments. Two main scenarios have been used to experiment and validate the proposed approach: the Resource-Constrained Project Scheduling Problem (RCPSP) and the Generalized Assignment Problem (GAP). This work contributes to advancing the state-of-the-art in the study and development of multiagent systems and in the modeling and application of swarm intelligence techniques. The proposed approach to coordinate agents in complex scenarios is novel, effective and robust. This approach contributes to the search of coordination solutions to multiagent systems real applications.

Agentes atuando em sociedade devem agir de maneira coerente para atingir um objetivo comum. A coordenação nos sistemas multiagente previne o comportamento caótico dos agentes, permite que o sistema lide com restrições globais e a interdependência entre os agentes, e faz com que o sistema possa ser composto por agentes com diferentes competências. A coordenação pode ser baseada na estrutura organizacional, onde a comunidade de agentes atua a favor de um objetivo comum através da forma como estão organizados. Em ambientes dinâmicos a organização dos agentes deve se adaptar a mudanças nos objetivos do sistema, na disponibilidade de recursos, nos relacionamentos entre os agentes, e assim por diante. Esta flexibilidade é um problema chave nos sistemas multiagente e está relacionada a modelos de adaptação como os observados nos insetos sociais. O presente trabalho propõe uma abordagem para a geração e adaptação da organização de um sistema multiagente, em tempo de execução, utilizando como base os modelos teóricos de organização das colônias de insetos sociais. Esta abordagem enfoca a alocação e o escalonamento dinâmicos de tarefas distribuídos entre agentes com diferentes competências e em ambientes de larga escala. Dois cenários principais são utilizados para experimentar e validar a abordagem proposta. Estes cenários estão baseados em problemas de pesquisa operacional denominados Resource-Constrained Project Scheduling Problem (RCPSP) e o Generalized Assignment Problem (GAP). Este trabalho contribui para o avanço do estado-da-arte no estudo e desenvolvimento de sistemas multiagente e na modelagem e aplicação de técnicas de inteligência de enxames em problemas computacionais. A abordagem proposta para coordenação de agentes em cenários complexos é nova, eficaz e robusta. De maneira geral, esta abordagem contribui para busca da solução de problemas de coordenação de sistemas multiagente aplicados a problemas reais. / A community of individual agents must work in a coherent manner to reach some common goal. The coordination process in multiagent systems prevents chaotic behavior of agents, makes the system able to deal with global constraints and inter-agents dependencies, and allows the system to be composed of agents with different capabilities. This process is normally based on the organizational structure, where the community of agents works towards the system goal through the manner they are organized. However, in dynamic environments, agents must be able to adapt to the changing goals of the system, to the resources available, to their relationships with another agents, to changes on the environment and so on. This problem is a key one in multiagent systems and relates to models of adaptation, such as those observed among social insects. This work proposes a new approach to generate and adapt the multiagent organization on the fly based on the theoretical models of social insects colonies organization. This approach focuses on distributed dynamic scheduling and task allocation using agents with different capabilities in large scale environments. Two main scenarios have been used to experiment and validate the proposed approach: the Resource-Constrained Project Scheduling Problem (RCPSP) and the Generalized Assignment Problem (GAP). This work contributes to advancing the state-of-the-art in the study and development of multiagent systems and in the modeling and application of swarm intelligence techniques. The proposed approach to coordinate agents in complex scenarios is novel, effective and robust. This approach contributes to the search of coordination solutions to multiagent systems real applications.

The PhD thesis is composed of three chapters and discusses the policy choice under uncertainty and learning in the context of climate change.

No description available.