Bir Cyborg'un şehir algısı

When the algorithms of reinforcement learning (RL) blend with architecture and urban design, a 'cyborg' construct that entails cybernetic principles of feedback, can be used to find adaptable and responsive designs that are better suited for users, as user preference, user habitual behaviours, and social and environmental conditions are continuously changing. With methodologies motivated by RL algorithms, architects can use them in Unity's ML-Agents toolkit, containing agents that have the capability to learn from their positive and negative actions in environments, in an alike manner to how human beings would, through simulation studies that imitate real-world set-ups that may be too intricate to study using conventional techniques, by examining curiosity-driven behavioural patterns and preferences in agents to discover design solutions that apply to real-life. Using RL algorithms in ML-Agents mimics real-world settings where cyborg explorers, agents with a flâneur attitude on exploration, can use their curiosity about environments to be motivated in navigating and learning about it, offering architects the opportunity to learn about real-world behavioural patterns through the perceptions of a cyborg explorer. This thesis' simulation study examines the influence of curiosity as a motivational drive on a cyborg explorer's behavioural patterns in one of Istanbul's urban settings, with a focus on the Sultan Ahmet Camii neighbourhood.