Why Artificial Intelligence Still Stinks at Pattern Recognition
Artificial intelligence (AI) is not nearly as smart as we want it to be. Because we are not nearly as smart as we want to be.
Artificial intelligence (AI) is not nearly as smart as we want it to be.
That’s the biggest takeaway from a new experiment out of MIT’s Computer Science and Artificial Intelligence Lab. A team of researchers sought to improve its AI’s machine learning skills by watching 2 million videos and predicting what would happen next. “Teaching AI to anticipate the future can help it comprehend the present,” New Scientist reports. That ability to anticipate the future gives AI much needed context for everyday tasks, as researcher Carl Vondrick told New Scientist: “if you’re about to sit down, you don’t want a robot to pull the chair out from underneath you.” That ability to correctly anticipate the future will also “allow machines to not take actions that might hurt people or help people not hurt themselves,” according to VICE.
Here’s how the MIT team did it:
Credit: MIT CSAIL
This technique is an important step forward for making smarter AI, but it’s not the only way artificial intelligence learns. As we’ve said before, the goal of creating artificial intelligence is to create a system that converses, learns, and solves problems on its own. AI first used algorithms to solve puzzles and make rational decisions. Now, it uses deep learning techniques like the one from this MIT experiment to identify and analyze patterns so it can predict real-life outcomes. Yet for all that learning, AI is only as smart as a “lobotomized, mentally challenged cockroach,” as Michio Kaku explained to us here:
As smart as artificial intelligence is becoming, it learns in different ways than people. Generally speaking, the human brain has many different ways to learn the same information, but “the most effective strategies for learning depend on what kind of learning is desired and toward what ends,” according to research out of Stanford University. In short, some ways of learning are better at helping you learn specific kinds of information than others. We’ve got a whole primer on how to learn here, but here’s a quick summary:
Rather than memorize facts, your brain retains the information best if you create explanations for why those facts are true. That gives you context.
Rather than struggle to understand an abstract concept, articulate why it’s difficult. That can help you identify where you’re stuck and help you create a solution.
Rather than cram for a test, learn a little bit of the material over a longer period of time. That helps transfer the information into long-term memory storage.
AI does not learn in any of these ways -- yet. Until it can, we have nothing to fear from it. Except more horrifying short videos.
New research links urban planning and political polarization.
- Canadian researchers find that excessive reliance on cars changes political views.
- Decades of car-centric urban planning normalized unsustainable lifestyles.
- People who prefer personal comfort elect politicians who represent such views.
It turns out the human scalp has an olfactory receptor that seems to play a crucial role in regulating hair follicle growth and death.
- Scientists treated scalp tissue with a chemical that mimics the odor of sandalwood.
- This chemical bound to an olfactory receptor in the scalp and stimulated hair growth.
- The treatment could soon be available to the public.
Science and the squishiness of the human mind. The joys of wearing whatever the hell you want, and so much more.
- Why can't we have a human-sized cat tree?
- What would happen if you got a spoonful of a neutron star?
- Why do we insist on dividing our wonderfully complex selves into boring little boxes
SMARTER FASTER trademarks owned by The Big Think, Inc. All rights reserved.