Post Shake Up, DeepMind Explains It Has the Secret AI Sauce
June 15, 2021
Do you remember the power struggle between the posh DeepMinders and the Mountain View crowd? No. Oh, well, no problem. Here’s the short version: Mountain View triumphed. Is that the Mountain View unit which wrestled with ethical AI? Answer: Yep, so what? Mountain View won. DeepMind lost.
I think I have spotted the first official statement which suggests the direction the post-skirmish Google AI jabber will go. Sure, I may be wrong, but let’s take a look at what’s revealed in “DeepMind Researchers Say Reinforcement Learning Is the Key to Cracking General AI.” [If you want to get some info about reinforcement learning, try this link. For a run down of other AI “religions”, check out this link. ]
Now let’s look at the write up “DeepMind Researchers Say…” passage:
In a new paper submitted to the peer-reviewed Artificial Intelligence journal, scientists at UK-based AI lab DeepMind argue that intelligence and its associated abilities will emerge not from formulating and solving complicated problems but by sticking to a simple but powerful principle: reward maximization.
I ask, “But what if other methods are useful?” The response I hypothesize, “Well, we’re the DeepMind Google.”
There is this statement quoted in the “DeepMind Researchers Say” article:
“Reinforcement learning assumes that the agent has a finite set of potential actions. A reward signal and value function have been specified. In other words, the problem of general intelligence is precisely to contribute those things that reinforcement learning requires as a pre-requisite,” Roitblat said. “So, if machine learning can all be reduced to some form of optimization to maximize some evaluative measure, then it must be true that reinforcement learning is relevant, but it is not very explanatory.”
I have put in bold face and red the operative word in this quasi quantum supremacy type statement from the online ad agency: “If.”
Yep, if, the close cousin of would, coulda, shoulda, and “I apologize.”
Stephen E Arnold, June 15, 2021
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
