Few shot adversarial learning neural talking heads - Blog
A blog I wrote as part of coursework for the Deep Learning module in my MSc.
A blog I wrote as part of coursework for the Deep Learning module in my MSc.
This article discusses and reviews the research done in the design and development of virtual agents with a standalone integrated perception systems independent of the environmental and external factors — software and/or hardware — with a major focus on visual perception, and in that too, active visual perception rather than passive visual perception. Most of the agents, methods and techniques discussed in this paper are virtual and exist in a virtual environment. Previous research has been studied, critiqued, further thought upon and ultimately connected leading to a thought experiment that puts into perspective how giving the power of sight to intelligent virtual agents affects their behaviour compared to agents that do not “see”.
For my dissertation titled "Behavioural Differences in Intelligent Virtual Agents With and Without Active Vision Systems", I trained a completely clueless virtual hummingbot on a virtual island to locate and drink nectar from flowers using only it's eyes (RGB camera, active vision) and then compared the behaviour to a hummingbot with access to location and orientation of the flowers and its own position relative to the environment.
Read the thesis here.
Offline Website Builder