Suicide surpassed homicide to be the second leading cause of death among people 10-24 years old in the United States \cite{1}. This statistic is alarming especially when combined with the more than eight distinctly different types of clinical depression among society today \cite{2}. To further complicate this health crisis, let’s consider the current worldwide isolating pandemic often referred to as COVID-19 that has spanned 12 months. It is more important than ever to consider how we can get ahead of the crisis by identifying the symptoms as they set in and more importantly ahead of the decision to commit suicide. To capitalize on the modern shift to electronic-based interactions \cite{1}, the use of Artificial Intelligence (AI) and Machine Learning (ML) methods to aid in identification have been previously implemented in Virtual Human interviewing platforms. This effort examines these existing approaches and includes an independent survey that is used to solve the gap in early identification of depression and suicidal ideation using a virtual human interviewing platform by soliciting honest, open, and current feedback from Soldiers on how to optimize such a system to encourage its use in the future. Specifically, the analysis of the survey results identify critical gaps from a participants perspective to be security, customization's, and error handling recommended to be included in future development of the EMPOWER (Enhancing Mental Performance and Optimizing Warfighter Effectiveness and Resilience: From MultiSense to OmniSense) platform. These recommendations are provided to the USC-ICT EMPOWER team to be included in the next prototype and system test.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

This dissertation consists of three articles about colorism and its relevance to public administration (PA), with specific focus on local policing. The overarching arguments are: 1) our lack of focus on the nuanced factors related to race have hindered our ability to adequately respond to biased criminal justice (CJ) outcomes; and 2) there are hidden patterns of biased behaviors that originate at the street-level, and these patterns have the propensity to impact every aspect of CJ. Colorism could serve as a more comprehensive approach to addressing racial bias. Colorism is a system of disadvantage and privilege based on skin color, with a bias for lighter skin. Article I introduces colorism to the field of PA, and it uses data on workplace colorism complaints to illustrate how colorism currently intersects with PA. In addition, the article uses scenarios to demonstrate the potential impact colorism might have at the street-level. Article II builds upon the descriptions of colorism provided in Article I by simulating a conceptual model of colorism and local policing. The major finding of this study is that, counter to the expectations of the experiment, those in the middle of the skin color spectrum experienced higher rates of incarceration when aggressive steps were taken to counter colorism. The major contributions from this project include a conceptual model that describes the relationship between the distinct levels of colorism�"individual, interactive and institutional. In Article III, I explore two conceptual models of interactive colorism in a local policing context. In the first model, bias behaviors are less likely to receive a challenge. In the second model, biased behaviors are likely to be challenged by counter-behaviors (e.g., fair policing). Subject-matter experts and non-subject-matter experts were used to select the model that most accurately depicts the phenomenon. I used online focus groups and phone interviews with police officers, theorists (e.g., sociologists, psychologists), and non-subject-matter experts to gather feedback. Majority of the informants recommended that future research on interactive colorism be framed as a bidirectional phenomenon. The informants provided additional considerations for future research, such as the variation in police culture across police departments. / PHD

Yes / Membrane computing is a new natural computing paradigm inspired by the functioning and structure of biological cells, and has been successfully applied to many different areas, from biology to engineering. In this paper, we present kPWorkbench, a software framework developed to support membrane computing and its applications. kPWorkbench offers unique features, including modelling, simulation, agent-based high performance simulation and verification, which allow modelling and computational analysis of membrane systems. The kPWorkbench formal verification component provides the opportunity to analyse the behaviour of a model and validate that important system requirements are met and certain behaviours are observed. The platform also features a property language based on natural language statements to facilitate property specification. / EPSRC

The problem of constructing and analyzing systems of intelligent, autonomous agents is becoming more and more important. These agents may include people, physical robots, virtual humans, software programs acting on behalf of human beings, or sensors. In a large class of multi-agent scenarios, agents may have different capabilities, preferences, objectives, and constraints. Therefore, efficient allocation of resources among multiple agents is often difficult to achieve. Automated negotiation (bargaining) is the most widely used approach for multi-agent resource allocation and it has received increasing attention in the recent years. However, information uncertainty, existence of multiple contracting partners and competitors, agents' incentive to maximize individual utilities, and market dynamics make it difficult to calculate agents' rational equilibrium negotiation strategies and develop successful negotiation agents behaving well in practice. To this end, this thesis is concerned with analyzing agents' rational behavior and developing negotiation strategies for a range of complex negotiation contexts. First, we consider the problem of finding agents' rational strategies in bargaining with incomplete information. We focus on the principal alternating-offers finite horizon bargaining protocol with one-sided uncertainty regarding agents' reserve prices. We provide an algorithm based on the combination of game theoretic analysis and search techniques which finds agents' equilibrium in pure strategies when they exist. Our approach is sound, complete and, in principle, can be applied to other uncertainty settings. Simulation results show that there is at least one pure strategy sequential equilibrium in 99.7% of various scenarios. In addition, agents with equilibrium strategies achieved higher utilities than agents with heuristic strategies. Next, we extend the alternating-offers protocol to handle concurrent negotiations in which each agent has multiple trading opportunities and faces market competition. We provide an algorithm based on backward induction to compute the subgame perfect equilibrium of concurrent negotiation. We observe that agents' bargaining power are affected by the proposing ordering and market competition and for a large subset of the space of the parameters, agents' equilibrium strategies depend on the values of a small number of parameters. We also extend our algorithm to find a pure strategy sequential equilibrium in concurrent negotiations where there is one-sided uncertainty regarding the reserve price of one agent. Third, we present the design and implementation of agents that concurrently negotiate with other entities for acquiring multiple resources. Negotiation agents are designed to adjust 1) the number of tentative agreements and 2) the amount of concession they are willing to make in response to changing market conditions and negotiation situations. In our approach, agents utilize a time-dependent negotiation strategy in which the reserve price of each resource is dynamically determined by 1) the likelihood that negotiation will not be successfully completed, 2) the expected agreement price of the resource, and 3) the expected number of final agreements. The negotiation deadline of each resource is determined by its relative scarcity. Since agents are permitted to decommit from agreements, a buyer may make more than one tentative agreement for each resource and the maximum number of tentative agreements is constrained by the market situation. Experimental results show that our negotiation strategy achieved significantly higher utilities than simpler strategies. Finally, we consider the problem of allocating networked resources in dynamic environment, such as cloud computing platforms, where providers strategically price resources to maximize their utility. While numerous auction-based approaches have been proposed in the literature, our work explores an alternative approach where providers and consumers negotiate resource leasing contracts. We propose a distributed negotiation mechanism where agents negotiate over both a contract price and a decommitment penalty, which allows agents to decommit from contracts at a cost. We compare our approach experimentally, using representative scenarios and workloads, to both combinatorial auctions and the fixed-price model, and show that the negotiation model achieves a higher social welfare.

The way a multi-agent modeller represents an agent not only affects its ability to reason about agents but also the interpretability of its representation space as well as its efficacy on future downstream tasks. We utilize and repurpose metrics from the field of representation learning to specifically analyze and compare multi-agent modellers that build real-valued vector representations of the agents they model. By generating two datasets and analyzing the representations of multiple LSTM- or transformer-based modellers with various embedding sizes, we demonstrate that representation metrics provide a more complete and nuanced picture of a modeller's representation space than an analysis based only on performance. We also provide insights regarding LSTM- and transformer-based representations. Our proposed metrics are general enough to work on a wide variety of modellers and datasets.

Organizations are living in a world today where reorganization is a prerequisite for their survival and adaption to the changes demand in the society. During this process, employees’ learning in the workplace is critical for the progress of both employees and organizations, thus organizations are required to create the proper learning opportunities that promote employees’ learning which is not as easy to implement as planning. The aim of this study is to contribute the knowledge increase about the insurance agents’ learning conditions in connection with the reorganization that caused the new tasks for the insurance agents in The Swedish Social Insurance Agency. Based on the purpose of this study, there are two questions which needs to be asked:1) What activities do managers employ to promote the insurance agents’ learning and what are the challenges that the managers may face during this process?, 2) What can promote and limit the insurance agents' learning? A qualitative case study was performed based on 12 semi-structured interviews with the managers, insurance specialists and insurance agents at Swedish Social Insurance Agency. The collected data has been analyzed through a thematic analysis base on the theoretical framework; adaptive- and developmental learning, performance-oriented and development-oriented leadership, which is built on productive- and developmental logic. The result of the study shows that managers apply activities such as planned and partially planned activities as well as spontaneous activities to promote insurance agents' learning. Depending on the perspective, the insurance agents’ learning is promoted or limited in different situations. This study demonstrates the following learning conditions for the insurance agents’ learning; 1) design of work tasks; 2) management’s support; 3) specify goals and plans; 4) supporting documents; 5) enough time and space to learn; 6) training and internal collaboration; and 7) insurance agents’ competence andcommitments. The conclusion shows that both an adaptive- and a developmental learning exists and there is a dominance of the logic of development and production. A further conclusion is that the implementation of reorganization creates an opportunity for developmental learning with focus on the development logic. Thus, the reorganization promotes the insurance agent’s developmental although the organization shows interest towards adaptive learning with the focus on the logic of production. The summary of this study shows that there is a need of balance between the both logics of leaning (development and production) along with maintaining the balance between the adaptive- and development learning in order to promote the insurance agents’ learning.

The key objectives of this research were to develop an integrated design and analysis methodology for sheet-metal product development based on agent-based technology, feature-based design, optimization and finite element analysis techniques, and to study the performance of prototype systems developed based on such a methodology. To achieve the research objectives, an agent-based framework was proposed for integrating and coordinating activities of participants involved in sheet-metal product development based on the investigation of the industrial requirements and the procedures of the development of sheet-metal products. Prototype systems were developed based on the proposed framework to answer research problems outlined for the design and implementation of agent-based systems, such as agent encapsulation, system architecture, agent communication and agent coordination. The performance of such prototype systems demonstrates that communication and coordination among domain agents can facilitate product development and reduce product cost. An agent-based optimization approach based on an "A-Teams" approach (Talukdar et al, 1996) was proposed for process optimization in the tooling design stage to combine the utilization of the traditional optimization techniques used to solve sheet-metal forming problems and agent-based approaches. Three test cases were used of varying complexity from a rectangular cup to the NUMISHEET'99 automobile front door panel simulation benchmark for the determination of optimal drawbead restraining forces and blankholder forces when designing draw dies for stamped parts. A network of software agents, each implementing a different numerical optimization technique, was used in combination with metal forming simulation software to optimize process variables. It was found that the performance of each agent (and optimization technique) depended strongly on the complexity of the problem. For a given amount of computational effort, a network of collaborating agents using different optimization techniques always outperformed agents using a single technique in terms of both the best solution found and the variance of the collection of best solutions. To provide guidance for the design and implementation of real applications, static and dynamic attributes and metrics of such agent-based collaborative systems, which can be evaluated in the preliminary system design stage and the system implementation stage, were proposed to study the impact of system architectures and coordination strategies on system performance. In addition, real-time system performance was statistically studied based on the data collected by the visualiser agent generated with the agent building toolkit. The results of case studies for system performance evaluation demonstrate the applicability of evaluation strategies proposed and can be used as a reference model for performance and scalability analysis on agent-based sheet-metal product development systems. The proposed evaluation strategies are applicable to general applications for product development by taking into consideration other performance indicators. / Thesis / Doctor of Philosophy (PhD)

Colloidal processing has been widely used for many industrial application. Electrophoretic deposition (EPD) is an important colloidal technique, which plays an important role in the fabrication of organic-inorganic composites. In this work, advanced dispersing agents with excellent adsorption, dispersion and film forming properties have been developed. The adsorption mechanism and the deposition kinetics of EPD have been studied. It is shown that catechol and salicylic groups of dispersants can form chelation bonding with metal atoms on inorganic particle surface and provide particle charging, dispersion and deposition. The aromatic dispersants can adsorb on carbon nanotubes due to π-π interactions. The long chain ionic polymeric dispersants were investigated, which provide efficient electrosteric stabilization. The film forming properties of dispersants allow dispersants with dispersed materials to deposit and form composite films on the substrate. Universal dispersing agents have been developed for the colloidal processing. These dispersants show possibilities to adsorb and disperse various organic and inorganic materials and pave the way for the fabrication of multi-functional materials for various applications. The composite materials, prepared using new dispersants showed promising corrosion protection of metals, flame retardant and energy storage properties. / Thesis / Master of Applied Science (MASc)

This study aims to analyze how the problem of corruption manifests itself in thoroughlycorrupt contexts. Through a theoretical framework inspired by rational choice two theorieswill be examined against each other, collective action and principal agent theory. The studyaims to investigate whether part of the explanation is a misunderstanding of how corruptionmanifests itself in deeply corrupt nations. It does this by comparing two nations that bothshare such characteristics, Tanzania and Zambia with a theory-testing approach. The resultshows that while both theories have some merit, the collective action theory might be morefitting.