Characters in games and virtual worlds continue to gain improvements in both their visual appearance and more human-like behaviours with each successive generation of hardware. One area that seemingly would need to be addressed if this evolution in human-like characters is to continue is in the area of characters with emotions. To begin addressing this, the thesis focuses on answering the question “Can an emotional architecture be developed for characters in games and virtual worlds, that is built upon a foundation of formal psychology? Therefore a primary goal of the research was to both review and consolidate a range of background material based on the psychology of emotions to provide a cohesive foundation on which to base any subsequent work. Once this review was completed, a range of supplemental material was investigated including computational models of emotions, current implementations of emotions in games and virtual worlds, machine learning techniques suitable for implementing aspects of emotions in characters in virtual world, believability and the role of emotions, and finally a discussion of interactive characters in the form of chat bots and non-player characters. With these reviews completed, a synthesis of the research resulted in the defining of an emotion architecture for use with pre-existing agent behaviour systems, and a range of evaluation techniques applicable to agents with emotions. To support validation of the proposed architecture three case studies were conducted that involved applying the architecture to three very different software platforms featuring agents. The first was applying the architecture to combat bots in Quake 3, the second to a chat bot in the virtual world Second Life, and the third was to a web chat bot used for e-commerce, specifically dealing with question and answers about the companies services. The three case studies were supported with several small pilot evaluations that were intended to look at different aspects of the implemented architecture including; (1) Whether or not users noticed the emotional enhancements. Which in the two small pilot studies conducted, highlighted that the addition of emotions to characters seemed to affect the user experience when the encounter was more interactive such as in the Second Life implementation. Where the interaction occurred in a combat situation with enemies with short life spans, the user experience seemed to be greatly reduced. (2) An evaluation was conducted on how the combat effectiveness of combat bots was affected by the addition of emotions, and in this pilot study it was found that the combat effectiveness was not quite statistically reduced, even when the bots were running away when afraid, or attacking when angry even if close to death. In summary, an architecture grounded in formal psychology is presented that is suitable for interactive characters in games and virtual worlds, but not perhaps ideal for applications where user interaction is brief such as in fast paced combat situations. This architecture has been partially validated through three case studies and includes suggestions for further work especially in the mapping of secondary emotions, the emotional significance of conversations, and the need to conduct further evaluations based on the pilot studies.

This dissertation reports five experiments exploring the use of AI-based smart agents to support physician-patient interactions. In each experiment, a sample of female participants evaluates video tapes of simulated physician-patient interactions in a setting involving early stage breast cancer diagnosis. Experiment 1 manipulates communication style (empathetic/impassive) for both a human physician (played by an actor) and an avatar that mimics the human. Empathetic styles elicit more liking and trust from patients and are also more persuasive. The avatar loses less than the human physician on desirable patient outcomes when communication style changes from empathetic to impassive. A mediation analysis shows that the communication style and physician type effects flow serially through liking and trust to persuasion. Experiment 2 reports an extended replication, adding a new avatar with less resemblance to the human physician. The findings match those of Experiment 1: both avatars have similar effects on liking, trust, and persuasion and are similarly anthropomorphized. Experiment 3 examines whether the patient's mindset (hope/fear about the cancer prognosis) influences likely patient outcomes. The mindset manipulation does not influence patient outcomes, but we find support for the core serial mediation model (from liking to trust to persuasion). Experiment 4 explores whether it matters how the avatar is deployed. Introducing the avatar as the physician's assistant lowers its evaluations perhaps because the patients feel deprioritized. The human physician is evaluated significantly higher on all outcome dimensions. Experiments 1-4 focused on the first phase of a standard three-phased physician-patient interaction protocol. Experiment 5 examines communication style (empathetic/ impassive) and physician type (human/avatar) effects across the three sequential phases. Patient outcomes improve monotonically over the three interaction phases across all study conditions. Overall, our studies show that an empathetic communication style is more effective in eliciting higher levels of liking, trust, and persuasion. The human physician and the avatar elicit similar levels of these desirable patient interaction outcomes. The avatar loses less when communication style changes from empathetic to impassive, suggesting that patients may have lower expectations of empathy from avatars. Thus, if carefully deployed, smart agents acting as physicians' avatars may effectively support physician-patient interactions. / Doctor of Philosophy / Healthcare professionals often have the difficult task of breaking bad news to patients. Research has shown that physician's communication style influences patient outcomes (liking, trust, persuasion, and compliance). Some physicians may adopt an impassive communication style to avoid emotional involvement with patients and some others may be overly empathetic and are prone to be perceived as inauthentic. These deficiencies persist despite an emphasis on developing physician communication skills. As in other service domains, a new generation of humanoid service robots (HSRs) offers potential for supporting physician-patient interactions. The effectiveness of such Artificial Intelligence (AI)/smart agent supported physician-patient interactions will rest, in part, on the communication style designed into the smart agents. A patient interacting with a smart agent emulating a human physician may assess different cognitive capabilities (knowledge and expertise), attribute different motivations, and make different socio-cultural appraisals than when they interact with the physician in-person. This research examines whether communication style (empathetic versus impassive) implemented via facial expression and vocal delivery elicits different patient responses when interacting with a smart agent (a physician' avatar) versus the physician in person. Findings suggest that, an empathetic (vs impassive) communication style elicits more positive patient responses, avatar physicians fare at par or better than the human physician and lose less on the patient outcomes when the communication style changes from empathetic to impassive. The avatars' appearance does not play a role in persuasion. Avatars were similarly anthropomorphized and participants' mindset (Hope/Fear) did not influence the outcomes. However, if the avatars are introduced as assistants (versus standalone physicians) there is a possibility that patients may feel downgraded/deprioritized, leading to lower evaluations for the avatars than the human physician. The contrast created when the human physician introduces the avatar may have unintended consequences that lower the avatar's evaluation. Without a direct contrast, patients may be more receptive to avatar interactions, particularly as they become more familiar in service environments. Our findings suggest that, if carefully deployed, smart agents acting as physicians' avatars may effectively support physician-patient interactions. Indeed, patients may have lower expectations of empathy from an avatar versus a human physician. This can facilitate more effective physician-patient interactions and elicit positive downstream effects on patient liking, trust and compliance.

<p dir="ltr">Chemical science can be imagined as a universe of information in which individual galaxies, solar systems, stars, and planets are compounds, reactions, biomolecules, etc. which need to be discovered, researched, and documented. The problem with this is that the universe of chemical science is potentially vaster than the one in which we live, and we are exploring it in a relatively inefficient manner. There is a scene in one of my favorite television shows, Futurama, which paints a picture of traditional chemical exploration. Taking place in the 30^th century, the main character Fry loses his robot friend Bender in outer space and resorts to using a giant telescope in the Himalayan mountains to randomly search through points in space to try to find him. After days of searching nonstop, he gives up noting that it is an impossible task because space is so vast in size, and he is searching so inefficiently. While human exploration of chemistry may not be as inefficient, there are a lot of steps which are driven by trial and error and educated guesswork which ultimately introduce major inefficiencies into scientific discovery. While we don’t live in the 30^th century yet, we do have access to 21^stcentury technology which can assist in exploring chemistry in a more directed manner. This mainly involves using machine learning, search algorithms, and generative powered exploratory AI to serve as a force multiplier which can serve to assist human chemists in chemical exploration. To shamelessly compare this with another space-based sci-fi reference, this would be akin to deploying hundreds or thousands of automated space probes to search unexplored planets, akin to how the empire found the rebellion on Hoth in the Empire Strikes Back.</p><p dir="ltr">The journey to integrate AI with chemical exploration starts with the important concept of standardization and how to apply it to chemically relevant data. To easily organize, store, and access relevant aspects of small molecules, macromolecules, chemical reactions, biological assays, etc. it is imperative that data be represented in a standard format which accurately portrays necessary chemical information. This becomes especially relevant as humans aggregate more and more chemical data. In this thesis, we tackle a subset of standardization in Chapter 2 involving benchmarking sets for comparative evaluation of docking software. One major reason why standardization is so important is that standardization promotes ease of access to relevant data, regardless of if this access is attempted by human or computational means. While improving data access for humans is beneficial, computationally it is a game changer when datamining training data for machine learning (ML) applications. Having standardized data readily available for computational access allows for software to rapidly access and preprocess relevant data boosts efficiency in ML model training. In Chapter 4 of this thesis, the central database of the CIPHER close-loop system is standardized and integrated with a REST API, allowing for rapid data acquisition via a structured URL call. Having database standardization and a mechanism for easy data mining makes a database “ML ready” and promotes the database for ML applications.</p><p dir="ltr">Build upon data standardization and training ML models for chemical applications, the next step of this journey revolves around a concept known as a “chemical space” and how chemists can approximate and sample chemical spaces in a directed manner. In the context of this thesis, a chemical space can be visualized in the following manner. Start by envisioning any chemical relationship between some inputs and outputs as an unknown mathematical function. For example, if one is measuring the assay response of a specific drug at a certain concentration, the input would be the concentration, and the output would be the assay response. Then the bounds of this space are set by determining the range of input values and this forms a chemical space which corresponds to the chemical problem. Chemists sample these spaces every day when they go into the lab, run experiments, and analyze their data. While the example described above is relatively simple in scope, even if the relationship is very complex techniques such as ML can be used to approximate the relationship. An example of this approximation is shown in Chapter 3 of this thesis, where normalizing flow architecture is used to bias a vector space representation of molecules with chemical properties, creating a space which correlates compound and property and can be sampled to provided compounds with specific values of trained chemical properties. Training individual models is important, but to truly emulate certain chemical processes multiple models may need to be combined with physical instrumentation to efficiently sample and validate a chemical space. Chapter 4 of this thesis expands upon this concept by integrating a variety of ML modules with high-throughput (HT) bioassay instrumentation to create a “close loop” system designed around discovering, synthesizing, and validating non-addictive analgesics.</p><p dir="ltr">The final step of this journey is to integrate these systems which sample chemical spaces with AI, allowing for automated exploration of these spaces in a directed manner. There are several AI frameworks which can be used separately or combined to accomplish this task, but the framework that is the focus of this thesis is AI agents. AI agents are entities which use some form of AI to serve as a logical processing center which drives their exploration through a problem space. This can be a simple algorithm, some type of heuristic model, or an advance form of generative AI such as an LLM. Additionally, these agents generally have access to certain tools which serve as a medium for interaction with physical or computational environments, such as controlling a robotic arm or searching a database. Finally, these agents generally have a notion of past actions and observations, commonly referred to as memory, which allows agents to recall important information as they explore. Chapter 5 of this thesis details a custom agentic framework which is tailored towards complex scientific applications. This framework builds agents from source documentation around a specific user defined scope, provides them with access to literature and documentation in the form of embeddings, has custom memory for highly targeted retention, and allows form agents to communicate with one another to promote collaborative problem solving. Chapter 6 of this thesis showcase an application of a simpler agentic framework to an automated lipidomic workflow which performs comparative analysis on 5xFAD vs. WT mice brain tissue. The group of AI agents involved in this system generate mass spectrometry worklists, filter data into categories for analysis, perform comparative analysis, and allow for the user to dynamically create plots which can be used to answer specific statistical questions. In addition to performing all these operational and statistical analysis functions, the system includes an agent which uses document embeddings trained on curated technical manuals and protocols to answer user questions via a chatbot style interface. Overall, the system showcases how AI can effectivity be applied to relevant chemical problems to enhance speed, bolster accuracy, and improve usability.</p>