Data Science Projects by Sai Sreehas
Uploaded by
NIKHIL NIKHIData Science Projects by Sai Sreehas
Uploaded by
NIKHIL NIKHICommon questions
Powered by AIThe Maze Generator project implemented DFS, BFS, and A* algorithms in Python for maze solving. DFS (Depth-First Search) explores the maze paths deeply, returning to the last node when a path ends. BFS (Breadth-First Search) explores all neighbor nodes level by level. A* uses heuristics like the Manhattan distance to prioritize paths that appear closer to the target, aiming for more efficient pathfinding .
The AI-based Tic-Tac-Toe Playing Bot was designed using game-tree search algorithms, which are effective in evaluating every possible move and selecting the best outcomes. Given that Tic-Tac-Toe is a solved game, these algorithms ensure that the AI plays optimally, minimizing mistakes and leading to the best possible outcomes (win or draw).
Auxiliary loss in the Conditional Molecule generation project provided an additional training signal to guide the model toward more desirable outcomes, such as generating molecules with specific properties. It is important because it helps refine the model's predictions, improving accuracy and effectiveness in targeting diseases like Alzheimer’s .
Samudrala Sai Sreehas built a financial data platform that automated the extraction of stock data from Screener.in using Selenium. This platform stored the extracted data in a structured database and incorporated a Django-based web dashboard, improving decision-making speed and reducing manual analysis time .
Samudrala Sai Sreehas utilized his machine learning skills in the Conditional Molecule generation project by applying advanced models like diffusion models with cross-attention mechanisms. For web development, his skills were leveraged in creating the Hostel Management System and the Django-based dashboard for financial data processing, showcasing his ability to integrate programming with practical applications .
The Amphibious Car was designed to operate efficiently both on land and water, which allowed it to replace two separate devices with a single one. This innovation substantially reduced costs and energy requirements, offering a practical and economical solution for dual-terrain navigation .
Selenium was used to automate the extraction of stock data from Screener.in, significantly improving the efficiency of data collection by replacing what would otherwise be a slow and error-prone manual process. This automation not only speeds up data retrieval but also enhances the reliability and accuracy of the gathered data for further analysis .
Cross-attention-based property conditioning in diffusion models allows the incorporation of specific molecular properties into the generation process, ensuring that the generated molecules not only conform to the diffusive stochastic process but also align closely with the desired chemical properties. This significantly enhances the applicability of the generated molecules in targeted applications like drug discovery .
The Hostel Management System was a web-based project designed to handle operations like room allocation, billing, and complaint tracking. It reduced manual effort required for these tasks and significantly improved communication between students and hostel staff by providing a centralized platform for managing operations .
The Conditional Molecule generation project applied a diffusive model for drug-like molecule generation targeting Alzheimer's treatment. The model incorporated cross-attention-based property conditioning and auxiliary loss for better guidance in the generation process. This approach enabled the generation of molecules that met predefined properties related to Alzheimer's treatment effectively .
