neuron

Plot your filters, visualize what each neuron is learning. Watch the rate of change of your cost function.

By "turning it on or off" you can lose or find a whole table row or hash value, but surely a "single neuron" can't "store" a generalised "memory".

This is the guy that publicly fired someone during a conference call without a moment's thought, at the snap of a neuron, and without whatever procedures were in place at his company. You should expect this kind of behavior from him 24/7.

And this is a stretch, but it is known that FFI is caused by a malformed prion protein, which is a protein that is deeply involved in various neuron-related functions. It seems possible that this protein is involved in the cleansing process in some way.

But still, nature is able to somehow make different celltypes, such as a skin cell, blood cell, neuron, you get the point. \nThe way how cells "know" which proteins to make is that the hard drive doesn't only hold encoded strings but also some byte sequences in between that don't encode any strings.

I look at job ads for very difficult undertakings using the SV buzzwords of the day - pivot, agile, and so on, and my neuron starts throbbing. Moreover, I rather worry about all the people coming up in the bubble, thinking it is bedrock reality.

Neural networks do not have a rigid delineation between instruction-storage, data-storage, and CPU. Every neuron wears all three hats and intertwingles those concepts.

Let's destroy each neuron immediately after taking it's state and pushing it into the stream, so that the whole process of uploading and death are completely synchronized and nothing of the patient's consciousness remains afterward. I'm sorry, but I still do not think that the patient will experience being uploaded into a computer.

While lower layers look for low level aspects of an image, each layer progressively looks for higher and higher level phenomena--for instance, when you hear about "the cat neuron" it comes from probing a neuron in the higher levels of a convolutional net and finding that pictures of cats happen to turn this neuron "on" while pictures of anything else don't. What is "high level" exactly?

In the case of computer vision it can be quite easy to visualise the features that each neuron in the hidden layer is looking for. In cases where the data is more arbitrary, it can be much harder to reason about, so if your training algorithm isn't performing as you'd like, it can be difficult to understand why it isn't working, even if you already understand that the basic principle of a feed-forward network is just a bunch of simple algebra.

Current efforts involving simulated brains rely both on older, less precise versions of what this nature article talks about - let's call that the roadmap, as it is a map of where all the neurons sit and where they connect to. This project drastically increases the resolution of this roadmap. However the additional information needed is what traffic is carried by each road. The brain as we understand it today consists not just of the connections between the neurons but also that each neuron, and each different network of the brain, secretes it's own neurotransmitter.

The idea that "quantum vibrations" in neuronal "microtubules" lead to decidedly macroscopic effects like detectable EEG signals is absurd on many levels, including spatial and temporal scope. There's also the small problem of neuron behavior modeling. Our existing biochemical and information-theoretical models are a very good match for the behavior of neurons.

By modelling at the neuron level, the only inputs can be at the neuron level. That means you have to physically model the entire mouse's body to connect all the inputs. Even if you had a realistic body simulation, you also need to run the system for an extended period of time. Intelligence is an emergent behavior, and it needs time to work and learn. Since you have to model every single neuron firing at the same time, the speed at which you can run this model is unreasonably slow.