4 ms·
This is just bad epistemology. >As a consequence, if you, with your puny human brain, want to understand why AlphaGo chose a particular move, the “explanation”
by amasad 10y ago
This is just bad epistemology.
>As a consequence, if you, with your puny human brain, want to understand why AlphaGo chose a particular move, the “explanation” may well consist of the networks of weighted connections that then pass their outcomes to the next layer of the neural network. Your brain can’t remember all those weights, and even if it could, it couldn’t then perform the calculation that resulted in the next state of the neural network.
Why did you reduce the neural network to the connections and layers but maintained the “brain” at the level of the whole? If you want to compare apples to apples you should also reduce the brain to neurons and connections. When a ball is about to hit your face your brain “can’t remember how to calculate the velocity and acceleration of the ball and predict that in 100ms it will impact your face and thus you need to send a signal for your hand to move to cover your face”. Yet we do it, how? Well, it’s encoded in the complex system of nerves and networks just like the the information about the various potential Go boards and moves is encoded in the neural network.
>And even if it could, you would have learned nothing about how to play Go, or, in truth, how AlphaGo plays Go — just as internalizing a schematic of the neural states of a human player would not constitute understanding how she came to make any particular move.
Similarly if you were to study the neurons, nerves, and networks and their firing patterns that are involved in protecting your face from the ball you wouldn’t learn anything about balls, acceleration or even pain.
- laretluval 10y agoAre you arguing that we "understand" the opaque computations in our brain because they happen in our brain? That is pretty clearly not a useful definition of understanding.
- randyrand 10y agoNot only that, but they're forgetting time. A brain 2x the size does not mean it must learn twice as much, or can understand 2x more complicated things. But may just mean it can learn twice as fast.
- deepsun 10y agoI often hear "brain size" argument, and can't help but remind that crow birds (brain mass: 13g) are smarter than sperm whales (8,000g). By "smarter" here I mean performing human (1,300g) IQ tests better.
- throwaway87423 10y ago> When a ball is about to hit your face your brain “can’t remember how to calculate the velocity and acceleration of the ball and predict that in 100ms it will impact your face and thus you need to send a signal for your hand to move to cover your face”. It's likely doing optical flow calculations (much simpler), rather than integrating an ODE. Richard Dawkins apparently makes the same mistake about "catching a ball" in his book.
- mockery 10y agoBy 'optical flow calculations' do you mean a linear extrapolation of the projected 'image-space' position of the ball? If so, that seems like it would be insufficient since a thrown object doesn't move linearly, let alone its projection. If not, then I'm curious what you're describing, and how it's much simpler than integrating an ODE?
- throwaway87423 10y agoNot quite. If you're interested take a look at, https://youtu.be/eKaYnXQUb2g?t=6m10s https://youtu.be/eKaYnXQUb2g?t=6m10s https://www.youtube.com/watch?v=iz9UVIo_ZUo&list=PL0AknDL1Vt-nEn1U-ztDulFATxiPvrELY&index=10 https://www.youtube.com/watch?v=iz9UVIo_ZUo&list=PL0AknDL1Vt... Camera homography preserves straight-lines, but it also gives a heuristic to avoid obstacles. Apparently this is also how bees and birds navigate.