I'd still like to see some human evidence of this effect to be fully convinced, but still that this story fits in this broader context does help convince me that this is likely true in humans.
128 karma · joined July 17, 2013
I'd still like to see some human evidence of this effect to be fully convinced, but still that this story fits in this broader context does help convince me that this is likely true in humans.
I would wait until replication in primates (human or non-human) before drinking ayahuasca to grow brain cells.
It's hard to answer about overarching purpose. The habits I highlighted each help towards some purpose (organizing my life, learning languages), but only form some portion of that goal.
https://www.beeminder.com/lambdaloop/mustdo - doing 1 admin task per day (slowed down after 3 months, stopped after 6 months)
https://www.beeminder.com/lambdaloop/anki - reviewing some number of anki cards each day (lapsed after ~2 months, recovered but lapsed after 2 months again and stopped)
(goal was private when archived, but yah similar trajectory) - doing weekly reviews (lapsed after ~5 months)
I did use beeminder for 3-4 months at a time, with great benefit! But looking back at every goal I set, it starts slipping past the 3-4 month mark. I got similar results with other systems (like the Lights spreadsheet from ultraworking or keeping a journal in text files), although without the nice beeminder charts..
I'm not sure, do other people see these patterns with their goals? Are there any tricks for fighting this late slippage?
For any system I've tried, it consistently falls after 3-4 months of use (this is beeminder, spreadsheets, text files, among others).
Does anyone have suggestions for how to keep going after a setback?
I've been blocking out time on my calendar for a year and a half now. It definitely adds value, as I know what I should be working on in order. Doing it on a digital calendar also means that I can move stuff around or add tasks quickly as plans (invariably) change.
When trying it as my only method of planning, I found that I was losing the bigger picture and losing track of key tasks. For instance, if I put some task to do for Friday but don't do it and forget to reschedule, it's just lost in my calendar. However, it works quite well with an external todo list, as a way to make it more concrete.
It can add or remove stress depending on how you use it. At first I blocked out too many things per day and would get stressed from not finishing them. Now, I block out maybe 2-3 main tasks which I can usually finish and feel good about spending time for myself in the evening.
Also, you don't need to use a whole street, you can encode a lot in just a room or two!
-- Relatedly, I know some people say that they don't have visual imagery: https://en.wikipedia.org/wiki/Aphantasia . I'm really curious how navigating spaces works for them. Do they have a problem with directions? Could they use memory spaces, but without visual imagery?
Recently, I've been reading "The memory code" by Lynne Kelly and found that most of the ancient cultures faced the problem of memory and used memory spaces (aka "method of loci").
I tried it out recently to memorize the timeline of evolutionary history while walking in city streets. I found I could memorize 1-3 items per minute, which very high recall hours and days later. The structure between items was readily apparent, so that I could, for instance, easily estimate that there were ~650 million years between the first terrestrial eukaryotes and first appearance of sharks.
The streets have acquired new meaning and the process was really quite enjoyable. With this tool, I actually want to learn the structure of many more things.
I tried a couple of mechanical keyboards (including ergodox) but I found the clacking sounds, the long key travel time, and just the bulkiness of the keyboards to be uncomfortable.
I found that there's a bunch of keyboards on Amazon that replicate the small keys. I'm currently using a wireless Jelly Comb keyboard but there's a bunch of similar ones. It's been my favorite so far!
Essentially, the hypothesis is that dreams are some kind of regularization that the brain performs on learning to prevent overfitting. The authors propose a couple of different arguments for why this may be useful (from deep learning) and why dreams in particular is where this regularization might happen (from neuroscience).
I don't think anyone would object that overfitting is bad and regularization is useful. I might even agree with the author that some kind of anti-Hebbian or noise corruption may be happening in the brain to help with this. However, the implied analogies in this paper between the tricks used in deep learning and dreams feels really "just so" to me.
I think their main evidence is that:
1. Dreams facilitate learning
2. The dreams we experience generally reflect the repeated task we experienced during the day
3. We often hallucinate variations of environments, not the exact environment we experienced
I think overall from this, I feel it is fair to say that the brain does something at night to help improve learning. It's not immediately clear that this is analogous to dropout, training augmentation, or generative models used in deep learning. It is even less clear whether dreams are truly needed to perform this. The authors don't really dissociate dreaming from sleeping in most of the literature that they cite. Remember that your experience and the weight changes in your brain may not be that highly correlated. Many neural changes may happen during a dream which are not correlated with your conscious perceptions/recollection of that dream.I find that as we gain new tools to study the nervous system more specifically, both data and models of how neurons are organized at the circuit level become more important. To advance on an analogy in the article, it's like trying to explore the dynamics of NYC without a map. For instance, it's hard to tell how/why people interact with central park if you don't even know where they live. The more specifically you are able to pin down people, the more it matters where exactly they live to understand.
Granted, the fly is much simpler than humans or even mice, and it will likely take decades and new tools for us to study humans in this way. However, when we get there, mapping out the brain connections will be crucial to make sense of it all.
Note: AUR isn't really that bad, as long as you install a package manager that handles it (like yaourt).
If there's a way to completely remove the graphical interface, it's definitely harder than in Linux, and not as usable afterwards.
You might argue that these options are not needed for a typical OS X user, and perhaps you're right. But you cannot say it really gives you "full control of the system".
(Via reddit: http://en.reddit.com/r/Scholar/comments/1kc8mw/request_linea...)
The Gallant lab at UC Berkeley did something somewhat similar about 2 years ago. See here: https://www.youtube.com/watch?v=KMA23JJ1M1o
From what I understand, both reconstructions involve setting up models of brain activity for vision, learning the parameters by machine learning from patients, and then using Bayesian inference to determine what is being seen.
While incredibly cool, we are still a long way from reading thoughts, and even longer if we're not allowed to learn the parameters for that subject first. Right now, we can only kinda reconstruct what someone is seeing, but that's really not much better than a camera.