very nice resource for those interested in explainable AI/ML: https://christophm.github.io/interpretable-ml-book/
sharing this because the meme that "we don't understand DNNs" is less and less true with current research trends and needs to DIE.
If I get interest, I can also share my collection on papers doing more theoretical analysis, proving bounds etc
#pettyBourgoisPrivilege: paying off all of your obligations for 2018 before the end of 2018.
Not being ironic either. Loads of people can't, a lot of them due to no fault of their own. Which sucks and we all need to continue working towards changing that.
I am really grateful for everyone who helps me live a life as comfortable as I do (even if it's nothing fancy in Instagram standards) and hope I help others as well
Everyone involved in science should join the fight for default open access. Science is payed for mainly by taxes and do the results belong to the citizens at minimum, to all of us in the long run
I think this will be part of my focus in 2019:http://nautil.us/issue/46/balance/darwin-was-a-slacker-and-you-should-be-too
Posting this both to multiply and to remind me to take a detailedly look later
Powerful video, but content warning: suicide, self harm, mental illness
C++20 looks delicious for the performance nerd in me 🤑 I hope rust gets something similar to nano-coroutines
but while I know very little, as best as I can tell "the cracks" mean that as long as you are and *individual* you can get away with quite a lot, and apparently there seems to be a "fuck you I've got mine" mentality left over similar to soviet Russia and over corrupt states ? Seems like an interesting dichotomy and maybe part of why the theories have started to orthogonalize the two concepts. (4/4,fin)
Meanwhile in the US you have a larger pressure to conformity because your job means much more, since it's tied to health insurance and the social safety network is much weaker. Similarly China is cast as more collectivist than Europe, (3/x/)
But even if you accept them as a two pole continuum, there is a pretty weird conflict between stated or attributed values *in Identity* vs *observed* values. More collectivist cultures. E.g. European culture is usually described as more collectivist than American, but I'd argue the average and median individual has more actual freedom to express their individuality in Europe because things like healthcare etc are taken care of. [2/x]
Thought I have been carrying around with me: Max Weber's and Hofstede's concept of an individualism/collectivism axis in cultural values has taken some hold in the popular discourse as far as I can observe, even though the modern theories seem to view them as orthogonal concepts (https://en.m.wikipedia.org/wiki/Collectivism) [1/X]
I came here to test and update my views on the communities attitude to AI, A(G)I risk and superintelligence to make sure I wasn't straw-manning them and I'm very happy actually to find some new info to update to.
Thank you https://www.fhi.ox.ac.uk/team/ben-garfinkel-bio/ for giving me an update, AI is actually not that big a chunk of the "EA portfolio". Less than about 5% of grants and 15% of donor focus is pretty clear.
Based on discussions and recommendations I had developed the assumption that it (in the form of superintelligence AI risk) was much more prominent.
Oh fuck yes, viva Europa: mandatory open access
Via hackernews, cool list of patterns: http://gordonbrander.com/pattern/
Interesting collection of links on how a carbon tax is needed and (according to the numerous sources cited) apparently good for the economy
Sharing interesting finds I encounter during my research. Topics include: ML/AI,psychology,neuroscience, neuromorphic memristives, economics, consciousness, complex systems
As a general instance, we are not centered on a specific theme, or a specific language. Everyone is welcome as long as they follow the few rules we have.
We always run the latest stable version of Mastodon, providing you a stock experience, with the following modifications:
Although we are a medium+ sized instance, we'd like to keep the feeling that you are at home and safe here.