Our lectures with Phoenix Perry discussed the ideas of machine learning, and a program called Interact ML.
The discussion began with a talk about search engines use machine learning inputs to filter searches, and as such, may be vulnerable to certain bias and discrimination. The core idea was that when a ‘man’ or ‘woman’ was searched on Google Images, the primary results contained pictures of Caucasian individuals.
While I understand the idea and viewpoint made, I genuinely believe that there is no reason to try and combat it. If we tried to cater to every single individual on the net with their certain searches, the net would be a much different place and have no semblance of a baseline. Beyond that, as stated these searches are the result of the most common concepts related to certain words, and trying to change them may just end up confusing anyone who is searching for items.
The next part of the discussion was on the usage of InteractML, which is a visual scripting tool designed for Unity to aid in the development of videogames without the usage of traditional coding techniques. The ML stands for the Machine Learning aspect of the program, where it must be taught certain interactions through performing them. After the program learns them however, these interactions can then be coded simply using the visual nodes.
The idea of stylized tools to aid in the development process of projects has always been something I’ve been interested in, having looked into software such as Gravity Sketch VR to create 3D models INSIDE VR for use in projects. In the case of InteractML, I would definitely be interested in giving it a try, with me having already seen the frustration common coding practices can cause for those inexperienced in its quirks.