Symbols and AI

Went a little overboard on the discussion last time (don’t worry–more like that will inevitably show up) so today’s topic is a little shorter and sweeter.

I brought up in my first post that the importance of symbols stretches beyond the confines of merely narrative and philosophy. Symbols have also made a great splash in Rhetoric (courtesy of some theories of Kenneth Burke) and my own, personal, homeland: Computer Science.

For this discussion, I’ll be focusing on symbols a la Plato–as a means to convey truths about objects that are more real than the individual can perceive.

Rhetoric loves this idea. It adopts it wholeheartedly and raises it very nearly as one of its own. The art of argumentation is all about perspective–what you choose to highlight will consequently conceal other perspectives. Thus, the selection can often be considered a rhetorical device–a symbol that represents that argument in that moment. Rhetoricians use this to establish commonality between audience and speaker to astounding effect.

Computer Science also adores symbols, though I don’t think many would tell you that in quite that way. In the process of training artificial intelligence, robots learn to recognize patterns by landmark traits. This is generally just called machine learning, but–shocker–it actively utilizes symbols to try and stitch together a picture of what it’s being shown! I’m uncertain to what degree this matters to AI development but I’ve been chomping at the bit to find out.