4 ms·
Most machine learning is assigning weights in a chain of matrix multiplications and normalization functions. There is no known experimentally verifyiable model
by greysphere 4y ago
Most machine learning is assigning weights in a chain of matrix multiplications and normalization functions.
There is no known experimentally verifyiable model of toddlers' brains, let alone one based on matrix multiplication and normalization. Developing such a model would be a noteworthy achievement.
Therefore these are different.
- 9wzYQbTYsAIc 4y agoSome Artificial Neural Networks have been shown to significantly (at least up to 50% concordance) model brain function. Not the mention the laborious work of neuroscientists to build out the connectome of the human brain.
- greysphere 4y agoTwo systems that produce the same output for some set of inputs doesn't show the systems are the same. My phone can produce the same results as my brain for short arithmetic problems. My phone is not a brain. The neuroscientists I know in the field would be among the first to tell you that our ability to model the brain is nearly non existent. In fact we don't even have a great model of a single neuron [1]. This statement doesn't invalidate the work folks are doing to try and reach that goal. Biology is hard. [1] https://en.m.wikipedia.org/wiki/Biological_neuron_model https://en.m.wikipedia.org/wiki/Biological_neuron_model
- 9wzYQbTYsAIc 4y agoSee https://news.ycombinator.com/item?id=33197285 https://news.ycombinator.com/item?id=33197285
- mattkrause 4y agoAs a working neuroscientist, I’ll co-sign this! Understanding 50% of the brain, whatever that would even mean, is an utter fantasy.
- 9wzYQbTYsAIc 4y agoI should have clarified that I was talking about the specific brain function of semantic comprehension. I am not suggesting that we are anywhere near having a complete analytical model of 50% of the brain. I am suggesting that we do have tools to continue answering questions about functional aspects of the brain. Or am I missing something that indicates the non-utility of “function analysis” of biology-based artificial neural networks?
- orbifold 4y agoThese articles use far more cautious language than you suggest and if they don't everyone working in the field is hopefully aware that such claims are the academic equivalent of clickbait at best.