You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Sure. Both approaches can be argued for and we should not take a principled approach. Improving current AI systems may need more innate knowledge or may not. And the best way to demonstrate which approach is most promising is to build more robust and useful AI with it.
right, but the current explorations have been massively weighted towards the blank slate end of the spectrum. maybe time for some balance?
Indeed, if we’d gotten ahead of the evolutionary game as Arthur C Clarke projected, we’d already have our own HAL.pic.twitter.com/EJoxv17twb
Arguably, the correct black-box approach will illuminate the other way. For example, a two-tier ML system can be used to first reduce data to "important but general basics" (generalise), then many specific ML structures can use the output to do, say, specific object recognition.
-so, as long as there is an information bandwidth chokepoint at the output of the 1st tier, and as long as the training set includes ALL the specific objects, labeled and all, then the system cannot help but come up a generalization (in the 1st tier), that is "universally" useful
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
