"Bust the Ghost" is an interactive game that combines probabilistic inference and engaging gameplay mechanics. The objective is to find and capture a ghost hidden in a 9x12 grid using sensor readings and Bayesian inference. Objectives and Game Mechanics
Grid Layout: The ghost is placed randomly in the grid according to a uniform prior distribution. Sensor Readings: Players click on grid cells to receive color-coded feedback: Red: Ghost is in the clicked cell. Orange: Ghost is 1-2 cells away. Yellow: Ghost is 3-4 cells away. Green: Ghost is at least 5 cells away. Gameplay: The game continues until the player locates the ghost or exhausts their credits or bust attempts.
Bayesian Inference: The game updates the probability of each cell containing the ghost based on sensor readings using Bayesian methods. Probability Updates: When a sensor reading is received, probabilities are recalculated to reflect the new information.
Interactive Components: Grid Display: Click to receive sensor feedback. Peep Toggle Button: Displays current ghost location probabilities. Bust Button: Allows players to guess the ghost's exact location. Score and Bust Attempts: Tracks remaining credits and guesses.
JavaScript & React: For dynamic and responsive UI development. Redux: For efficient state management, ensuring consistent state transitions and strategic gameplay. Custom Utility Functions: For Bayesian updates and probabilistic modeling, enhancing decision-making and gameplay dynamics.
Integration of a direction sensor providing directional hints, improving the gaming experience by combining distance and directional data for more accurate probability updates.
"Bust the Ghost" showcases the application of Bayesian inference in gaming, creating a strategic and engaging experience. The use of advanced web development technologies like React and Redux ensures a responsive and interactive UI, while probabilistic modeling adds depth to the gameplay.
| Pipeline Step | What I Did | Azure Equivalent |
|---|---|---|
| Input & State Mgmt | Captured user interaction and game state | Azure Logic Apps (event-driven trigger) |
| Processing | Applied Bayesian inference for probability updates | Azure Databricks (rule-based logic) |
| Modeling/Decision | Recalculated probabilities dynamically | DBT (calculated models/views) |
| Output | Visual feedback and decision system (UI in React) | Power BI / Web Frontend |
| Automation | Used Redux actions to trigger logic | Logic Apps / Function Apps |
| Versioning | Stored and tracked with GitHub | Azure DevOps Git |