This is a linkpost for https://singularityhub.com/2016/12/07/this-one-equation-may-be-the-root-of-intelligence/
Can we not do clickbait titles on linkposts, please? Let's use the Hacker News rule -- default to the article title, but if it's not a good representation of the content of the article (e.g. it's clickbait), change it to something descriptive.
Namely? Bayes? (TBH I wouldn't expect bayes because that'd be wrong, I think - you can have "dumb" intelligence based on reinforcement learning)
This equation is simply the sum of each x = i choose k for k in [ 1, i ].
So what he's saying is that the neural circuits that follow the principles he describes have one neuron to represent every possible combination of on/off states in the set of inputs. It's the most brain-dead way you could possibly implement a classifier system.
Does the magical 2^i-1 equation predict that the human brain with cca 85-86 billion neurons can only contain 36 different concepts?
From a paper by Dr. Tsien, retrieved from http://www.augusta.edu/mcg/discovery/bbdi/tsien/documents/theoryofconnectivity.pdf
Fifth, this power-of-two mathematical logic confines the total numbers of distinct inputs ( i ) coming into a given microcircuit in order to best utilize the available cell resources. For instance, as a result of its exponential growth, at a mere i = 40, the total number of neurons ( n ) required to cover all possible connectivity patterns within a microcircuit would be more than 10^12 (already exceeding the total number of neurons in the human brain). For Caenorhabditis elegans – which has only 302 neurons, limiting i to 8 or less at a given neural node makes good economic sense. Furthermore, by employing a sub-modular approach (e.g., using a set of four or fi ve inputs per subnode), a given circuit can greatly increase the input types it can process with the same number of neurons. '
He also mentions cortical layering. It seems like he's envisioning the brain as a forest of smaller, relatively shallow networks following the principles he describes, rather than one tree where all neurons are wired together in a uniform way.
"In stark contrast, Tsien predicts the brain runs on a series of pre-programmed, conserved networks. These networks are not learned; instead, they’re made up of pre-established neural networks, wired according to a simple mathematical principle.
In other words, at a fundamental level the brain’s wiring is innate — the motifs, established by genetics, underlie our ability to extract features, discover relational patterns, abstract knowledge and ultimately, reason."
Brain Computation Is Organized via Power-of-Two-Based Permutation Logic
http://journal.frontiersin.org/article/10.3389/fnsys.2016.00095/full
" the unifying mathematical principle upon which evolution constructs the brain’s basic wiring and computational logic represents one of the top most difficult and unsolved meta-problems in neuroscience"
"This simple mathematical logic can account for brain computation across the entire evolutionary spectrum, ranging from the simplest neural networks to the most complex."