Collectible Card Games (CCG) are one of the most popular types of games in both digital and physical space. Despite their popularity, there is a great deal of room for exploration into the application of artificial intelligence in order to enhance CCG gameplay and development. This paper presents Fifth Aeon a novel and open source CCG built to run in browsers and two A.I applications built upon Fifth Aeon. The first application is an artificial intelligence competition run on the Fifth Aeon game. The second is an automatic balancing system capable of helping a designer create new cards that do not upset the balance of an existing collectible card game. The submissions to the A.I competition include one that plays substantially better than the existing Fifth Aeon A.I with a higher winrate across multiple game formats. The balancer system also demonstrates an ability to automatically balance several types of cards against a wide variety of parameters. These results help pave the way to cheaper CCG development with more compelling A.I opponents.

Recent years have seen several projects across the globe using drones to detect sharks, including several high profile projects around alerting beach authorities to keep people safe. However, so far many of these attempts have used cloud-based machine learning solutions for the detection component, which complicates setup and limits their use geographically to areas with internet connection. An on-device (or on-controller) shark detector would offer greater freedom for researchers searching for and tracking sharks in the field, but such a detector would need to operate under reduced resource constraints. To this end we look at SSD MobileNet, a popular object detection architecture that targets edge devices by sacrificing some accuracy. We look at the results of SSD MobileNet in detecting sharks from a data set of aerial images created by a collaboration between Cal Poly and CSU Long Beach’s Shark Lab. We conclude that SSD MobileNet does suffer from some accuracy issues with smaller objects in particular, and we note the importance of customized anchor box configuration.

In this thesis, we analyze the audio signal of a guitar to extract musical data in real-time. Specifically, the pitch and octave of notes and chords are displayed over time. Previous work has shown that non-negative matrix factorization is an effective method for classifying the pitches of simultaneous notes. We explore the effect of window size, hop length, and other parameters to maximize the resolution and accuracy of the output.Other groups have required prerecorded note samples to build a library of note templates to search for. We automate this step and compute the library at run-time, tuning it specifically for the input guitar. The program we present generates a musical visualization of the results in addition to suggestions for fingerings of chords in the form of a fretboard display and tablature notation. This program is built as an applet and is accessible from the web browser.

<P> Video coding standards have been designed to support many applications such as broadcasting systems, movie industries and media storage. All video coding standards try to reduce data in video sequences as much as possible by exploiting spatial and temporal redundancies. Although those video coding standards are suitable for a wide variety of applications, some applications require low encoder complexity specifically for real time video encoding. Most of the computational complexity of a video encoder can be attributed to the motion estimation function. </p> <p> Motion estimation using multiple reference frames is widely used as the basis for recent video coding standards (eg. H.264/ AVC) to achieve increased coding efficiency. However, this increases the complexity of the encoding process. In this thesis, new techniques for efficient motion estimation are proposed. A combination of multiple reference frame selection and image residue-based mode selection is used to improve motion estimation time. By dynamic selection of an initial reference frame in advance, the number of reference frames to be considered is reduced. In addition, from examination of the residue between the current block and reconstructed blocks in preceding frames, variable block size mode decisions are made. Modified initial motion vector estimation and early stop condition detection are also adopted to speed up the motion estimation procedure. Experimental results compare the performance of the proposed algorithm with state of the art motion estimation algorithms and demonstrate significantly reduced motion estimation time while maintaining PSNR performance. </p> <p> In addition a new side information generation algorithm using dynamic motion estimation and post processing is proposed for improved distributed video coding. Multiple reference frames are employed for motion estimation at the side information frame generation block of the decoder. After motion estimation and compensation, post processing is applied to improve the hole and overlapped areas on the reconstructed side information frame. Both median filtering and residual-based block selecting algorithms are used to deal with hole and overlapped areas, respectively. The proposed side information method contributes to improving the quality of reconstructed frames at the distributed video decoder. The average encoding time of the distributed video coding is shown to be around 15% of H.264 inter coding and 40% of H.264 intra coding. The proposed side generation algorithm is implemented in a frequency domain distributed system and tested throughout various test sequences. The proposed side information based distributed video coding demonstrates improved performance compared with that of H.264 intra coding. </p> <p> Experimental implementations of the proposed algorithms are demonstrated using a set of video test sequences that are widely used and freely available. </p> / Thesis / Doctor of Philosophy (PhD)

Peers are an invaluable resource for students at undergraduate universities. Many factors can impact how students form connections, some of which are tied to the students’ identities. While social networks have been studied in the context of universities, little research has been done specifically about peer collaboration and even less in the context of computer science classes. Our research aims to gain an understanding of how peer networks form in computing classes and the effect of being involved in one on students’ academic performances. We collected survey data (n = 139) about students’ peer collaborative behaviors in computer science classes. Several patterns were observed. Students were more likely to collaborate with students with their same demographic identity, including gender and ethnicity. A student’s demographic identity had no predictive quality on their participation in a peer network. Students who participated in a peer network had a higher grade performance in the class compared to students who did not.

This dissertation addresses the need for using machine learning-based methods rather than traditional rule-based methods for controlling non-playable characters (NPCs). The goal of the Reinforcement Learning Framework for the Unreal Engine is to enable game development studios to create, train, and more easily implement smarter, more compelling AI characters in major video game releases. The framework contains three distinct software libraries: an Unreal Engine reinforcement learning library whose purpose is to enable Unreal Engine levels to act as reinforcement learning environments, a python library which provides convenient abstractions and implementations to the reinforcement learning process, and a flexible connection system responsible for the communication between the two sides of the framework. In this dissertation, I describe the framework in detail, demonstrate the framework’s capability by implementing, training, and evaluating on the cartpole benchmark, and prove the system’s viability by comparing it to similar tools already on the market.

<p>Many computer systems in our daily life require highly available applications (such as medical equipment) and some others run on difficult to access places (such as satellites). These systems are subject to a variety of potential failures that may degrade their performance. Therefore, being able to reason about faults and their impact on systems is gaining considerable attention. Existing work on fault tolerance is mostly focused on addressing faults at the programming language level. In the recent past, significant efforts have been made to use formal methods to specify and verify fault tolerant systems to provide more reliable software. Related with this, some researchers have pointed out that Deontic Logic is useful for reasoning about fault tolerant systems due to its expressive nature in relation to defining norms, used to describe expected behaviour and prescribing what happens when these norms are violated.</p> <p>In this thesis, we demonstrate how Deontic Logic can be used to model an existing real world problem concerning fault tolerance mechanisms. We consider different situations that a vehicle faces on the road and the consequent reactions of the driver or vehicle based on good and bad behaviour. We got the idea and motivation for this case study from the SASPENCE sub-project, conducted under the European Integrated Project PReVENT. This sub-project focuses on a vehicle’s behaviour in maintaining safe speed and safe distance on the road. As our first modelling attempt, we use a Propositional Deontic Logic approach, to justify to what extent we can apply this Logical approach to model a real world problem. Subsequently, we use a First Order Deontic Logic approach, as it can incorporate the use of parameters and quantification over them, which is more useful to model real world scenarios.</p> <p>We state and prove some interesting expected properties of the models using a First Order proof system. Based on these modelling exercises, we acquired different engineering ideas and lessons, and present them in this thesis in order to aid modelling of future fault tolerant systems.</p> / Master of Science (MSc)

<p>With the growing demand for high data rate communication services, multi-carrier communication schemes have started to become the method of choice in many applications. In this thesis, multi-carrier communication systems are studied and various methods for improving their achievable bit rate arc proposed.</p> <p>Filtered multitone (FMT) is a multi-carrier communication scheme which is implemented using a modulated filter bank structure. In this thesis, an efficient design method for the prototype filter of the FMT system is proposed. This design method allows improvement of the achievable hit rate by efficient evaluation of the inherent trade-off between the sub channel spectral containment provided by the prototype filter and the intersymbol interference (lSI) that the filter generates. Numerical results further demonstrat.e the effectiveness of the proposed design method. The insight gained from this design is also used to determine the optimal number of subchannels in FMT systems. Moreover, since the presence of lSI in FMT subchannels outputs renders the conventional water-filling power loading algorithm suboptimal, we propose an efficient power loading algorithm for Fl\IT that enables higher achievable bit rates.</p> <p>Discrete multitone (DMT) is a popular multi-carrier communication scheme, mainly due to its rather low complexity. However. DMT suffers from poor subchannel spectral characteristics. In this thesis, a family of bi-windowed DMT transceivers is proposed that provide both improved sub channel 8pectral containment at the transmitter and improved spectral selectivity at the receiver, without requiring the cyclic prefix to be longer than the order of the channel impulse respon8e. The window8 arc designed in a channel independent manner and are con8trained to produce sub channel outputs that arc free from lSI. Furthermore, the design allows the interisubchannel interference (leI) to be controlled in such a way that it can be mitigated using a relatively simple minimum mean square error (MMSE) successive interference cancellation scheme. Numerical results demonstrate the significant gain in the achievable bit rate obtained Ly the proposed scheme.</p> / Doctor of Philosophy (PhD)

<p>The practice of medicine has evolved from individual practice, mainly led by physicians, towards an interdisciplinary team-based activity that takes into consideration patients’ needs, preferences and values. This shift towards what is known as the patient-centered care (PCC) model requires better communication among medical staff and patients. Feedback is a key factor to improve team learning processes and also to understand patient perspectives. The objective of this thesis was to create a system to facilitate feedback processes within the Toronto General Hospital (TGH). To do so, a kiosk-based system was designed, allowing patients and health care providers to send feedback messages as well as positive recognition messages in appreciation of other people’s commitments and achievements. A pilot test was conducted by placing a kiosk in the General Internal Medicine unit at TGH for seven weeks. During this study, the application running the kiosk registered all user actions, with the objective of understanding how patients and staff interacted with the system. The kiosk was perceived as a positive, easy to use, and valuable tool by patients and providers. The different elements that need to be improved for future upgrades of the system are discussed. Overall, the use of kiosk-based systems to gather feedback should be considered when creating high performing teams in the health care domain.</p> / Master of Science (MSc)

<p>The purpose of this research has been to reassess the Ag-Mg system using the CALPHAD technique. Compared with previous assessments, we carry out the optimization by fitting calculations to the original data instead of second-hand information. Moreover, we use a two sub-lattice model and a four sub-lattice model based on compound energy formalism to simulate both first-order and second-order transformations between the FCC phase and the L1₂ phase. Undoubtedly, the CALPHAD technique has achieved a degree of maturity, but its deficiencies are regularly ignored.</p> <p>In this thesis, we develop an interval method based on Kantorovich’s idea to overcome the shortcomings of the CALPHAD technique. Both advantages and disadvantages of the interval method are discussed. We also present an example of the interval approach on thermodynamic optimization of the Ag-Mg melt. The results suggest that this method would be helpful as a pre-optimization tool.</p> / Master of Applied Science (MASc)