Titan explores AI with Generative Adversarial Networks.
Addressing the question whether ‘algorithms learn from the designers and designs, and serve as capable assistants augmenting the creative teams’, Titan’s AI experts are exploring the potential of GAN.
Generative Adversarial Networks (GAN) is a Deep Learning neural network that continuously improves over time by playing the dual roles of ‘doer’ and ‘critic’. Titan’s AI team trained the neural network with thousands of images from the company’s stellar collections to create this powerful tool. Generating new designs at the rate of over one per second, this tool enables the watch designer to quickly select and shortlist new watch concepts for further evaluation and development. The tool can either operate in an autonomous mode, or under the control of designers who direct its creativity through a simple, intuitive interface.
With features such as the ability to blend aesthetic elements of successful watches, re-imagine watch forms and generating new concepts quickly, this tool is designed to augment the capabilities of our watch designers. Further developments such as using AI to predict new products success rate, quantify cannibalisation effects with existing designs are currently underway. Post successful completion of the pilot, this tool is now being readied for deployment.