A common way to evaluate the performance of players in two-player games is to have them play against other players. If the player wins more games than other players, then it is said to be more capable; in other words, the strength of a player is measured relatively. In this project, we seek a way to evaluate the performance of players in terms of absolute. In recent years, self-play reinforcement learning has given rise to capable game-playing agents in a number of complex domains such as Go and Chess. We perform an analysis of a self-play agent using scaled-down versions of Go on a generic platform to measure the strength of the agent via our developed methods.

Homeopathy is a holistic natural system of medicine and helps patients recover from all types of illnesses naturally, while strengthening their immune system and increasing their energy and vitality. The Hygieia Homeopathy Clinic provides basic knowledge of homeopathy to their patients. Patients use their time-tested methods to trust in their own body’s recovery functionality. The main problem of the website is their patient’s inability to search the website for knowledge and protocols about homeopathy. Other problems with the website include the ability of patients to view and make appointments, purchase vitamins and supplements, and payment information. The Smart Homeopathy Doctor App Senior Design 2021 Team’s goal is to provide their clients a fully functional mobile app for easier content viewing, appointment making, shop, and patient messaging. Furthermore, the website needs to facilitate easy communication between the doctor and patients. The Smart Homeopathy Doctor App is a mobile application. Its primary function is to allow users to query a server-hosted database. The content of the database includes publicly available, non-sensitive data such as FAQs pertaining to homeopathy. The administrator performs database CRUD operations. Over the course of the project, our team has refined our time management skills and honed our Peer Review skills. We communication better by not only updating others on our progress but also asking members for help. We also learned Ionic Framework with Angular for our front-end user experience and learned to store our database in Firebase.

Our goal is to create a user friendly dashboard with data related to the current COVID-19 pandemic. This includes an interactive map, charts, and numbers presented to the user in a simplified manner. The data spans every county in the United States. Beyond just being a COVID-19 Tracker, our tool will be available as an API that can be used with any other state and county specific data.

Identifying new and cutting-edge investment strategies is a crucial step in establishing any large business within its relative industry. Fort Capital, whose primary investment focus is on industrial-grade buildings, is taking an innovative and insightful approach to geographic understanding. Fort Capital aims to identify trade routes used by major market players, such as Amazon and Walmart, to find the areas where industrial warehouses and large-scale distribution centers are in highest demand. To locate such trade routes, identifying the main travelers on these routes is essential, and Truck Detective aims to do exactly that. Using machine learning and artificial intelligence models such as a deep neural network, Truck Detective enables Fort Capital to detect, with high accuracy, the location of big rig trucks, and can additionally help identify where they came from or where they are heading. This, in turn, illuminates geographically important areas with promising investment opportunities for Fort Capital.

Vehicle Re-identification, which aims to retrieve matching vehicles across different cameras, is a challenging problem in Intelligent Transport System due to different factors such as illumination conditions, occlusions, and video resolution. Numerous studies are proposing the use of Deep Neural Networks, a recent advance in Artificial Intelligence, thanks to their exceptional feature embedding extraction. However, Deep Neural Networks perform poorly on cross-domain settings. Furthermore, vehicle re-identification training data is relatively limited because public videos are only accessible to the authority only. Our study tackles the above challenges by utilizing several state-of-the-art techniques on domain learning to expand the model's generalization capability. Our research shows that we can outperform other state-of-the-art models by large margins on popular vehicle re-identification benchmarks.