Hacker News Reader

kristjankalmon Measles cases hit record high in Europe

> The majority of cases we are seeing are in teenagers and young adults who missed out on their MMR vaccine when they were children.

The vaccines->autism paper by Wakefield came out 1998, so teenagers were born after that. Correlation/causation etc, but it'd be surprising if the media frenzy which followed didn't affect the vaccination rate.

Also, according to this report measles diagnosed in Europe: 2016 -- 5,273; 2017 -- 23,927; 2018 -- 41,000 (to date)

Grim.

kristjankalmon Switching Off: Joseph Brodsky and the moral responsibility to be useless

the full transcript of the trial of Joseph Brodsky:

http://soviethistory.msu.edu/1954-2/whats-a-woman-to-think/w...

so absurd and surreal at times it's hard to believe it's not fiction

---

Grudinina: The difference between a parasite and a starting-out poet is that the parasite eats but does not work, while the poet works but may not always eat.

Judge: The court does not appreciate that remark. In our country, a man earns according to his work.

kristjankalmon Facebook urged to make GDPR its “baseline standard” globally

>> with limited resources they wouldn't be able to focus on the critical period of acquiring users and instead would be forced into building compliance features

If you cannot comply with privacy rules you should not do social media, whatever you growth phase

kristjankalmon North Carolina Is Ordered to Redraw Its Congressional Map

it would be relatively trivial to simulate a null distribution -- generate 10^n quasi-random maps which then give you the null distribution of voters across districts. Compare that null distribution with an actual map and if it sits beyond the x-th percentile you can call it 'gerrymandered'.

kristjankalmon Quantum A.I. Publications

The Murray Gell-Mann Amnesia Effect

kristjankalmon Theories of civilisational ‘collapse’ grow less convincing when scrutinised

>> you would be able to find similar letters begging for Romes intervention during Romes reign.

which ones specifically then?

and more importantly, its not just this letter, as i mentioned above the whole gildas’ de excidio goes explaining in quite detail the sudden and devastating effect of the collapse of roman authority in britain.

do you have any evidence to back up your insistence that majority of the population went on like business as usual?

kristjankalmon Theories of civilisational ‘collapse’ grow less convincing when scrutinised

ok, to really flesh it out for you then -- this is a desperate letter by the representatives of roman citizens in britain to Aetius, consul, to send help or do something since barbarians, crop failures, raids etc. it certainly gives the impression of people witnessing and being subject to the collapse of the civic society. if you want more than an anecdote read the whole account of roman collapse in britain by gildas, a contemporary. its pretty desperate in tone and attitude.

kristjankalmon Theories of civilisational ‘collapse’ grow less convincing when scrutinised

>> namely that there wasnt a collapse in the mind of the people back then

my quote was an example how people were quite acutely aware of the collapse around them. wiki link just a reference for the quote.

kristjankalmon Theories of civilisational ‘collapse’ grow less convincing when scrutinised

even that i would dispute in case of Rome

"To Agitius, thrice consul: [...] The barbarians drive us to the sea, the sea drives us to the barbarians; between these two means of death, we are either killed or drowned."

https://en.wikipedia.org/wiki/Groans_of_the_Britons

kristjankalmon Theories of civilisational ‘collapse’ grow less convincing when scrutinised

Decline or collapse, or whatever word one uses -- is not subjective, as this article seems to imply. There are loads of metrics -- levels of urbanisation, existence of monetary system and taxes, levels of literacy, state funded infrastructure projects, literary projects, libraries and institutions of education. All of these pretty much disappeared in western europe between 400 and 700AD or were reduced to tiny isolated examples. Same dynamic with Hittites or Harappan city states. You can call it something else than a collapse but it's not a "subjective value judgement".

kristjankalmon Theories of civilisational ‘collapse’ grow less convincing when scrutinised

" generally weren’t particularly aware that there was a collapse going on"

yes they very much were -- read the history of Gothic wars or Totila's sack of Rome. Even more the sack by Alaric in 410, despite the lack of damage done, was a truly shocking event for contemporary writers and external observers.

kristjankalmon Estonia blocks electronic ID cards over identity-theft risk

http://news.err.ee/640385/government-to-close-id-card-certif...

"Information System Authority (RIA) Director General Taimar Peterkop likewise confirmed that the threat assessment had changed after the research published by the Czech researchers on Monday revealed that the security flaw affecting Estonian ID cards is easier to exploit than previously believed."

is what you are reffering to "the research published by the Czech researchers"?

kristjankalmon iPhone X

> The most powerful and smartest chip ever in a smartphone, with a neural engine that’s capable of up to 600 billion operations per second.

sorry but "neural engine" is just too funny. i mean there's normal levels of bullshit bingo and then there's branding your chip "neural engine".

kristjankalmon Xcode 8.3 produces binaries 3x larger than Xcode 8.2.1

you are obviously right, but i think that in the specific case described above -- computer code -- we have binary digits as final and indivisible units.

kristjankalmon Xcode 8.3 produces binaries 3x larger than Xcode 8.2.1

yes, you are quite right -- i was referring to the binary digit rather than the shannon

kristjankalmon Xcode 8.3 produces binaries 3x larger than Xcode 8.2.1

yes, but your gain is still an integer number. the non-integer ratio is just a statistic (gain per number of units).

kristjankalmon Model-Based Machine Learning

rather than downvoting i'm actually curious why you think unsupervised learning is not ML?

there'd be so much less noise in these comments/discussions if we just did away with vague and illdefined labels such as ML or AI

kristjankalmon This AI Boom Will Also Bust

This is a false dichotomy. Both OLS regression and, say, random decision forest regression have the same objective (predict values) and achieve it with similar means (build a generative model / function). They solve the same problem. Contrastingly, assembler and python are broadly aimed at completely different use cases.

Broadly, whether you should move from OLS to random forest regression = SNR increase / increase in manhours and money spent.

kristjankalmon This AI Boom Will Also Bust

> AI helps provide good analysis without having to cleaning up data manually.

I don't even know where to start. I suppose you don't really think that what separates ML/AI (whatever that means) from your standard OLS regression is that denoising is not done manually?

> Cleaning up data is very expensive. And without that, the analysis is good for nothing.

I hope you are not saying here that linear regression cannot handle noise.

In the end, as this blog post also points out, ML/AI is just a vague blanket term. What you want is a statistical method that captures the signal as fast and efficiently as possible and often the gain from going beyond simple linear models might be marginal.

kristjankalmon How close are we to solving vision?

No, not really. The "structure of the world" such nets learn is based on bottom-up processing of data -- moving from basic features such as orientations and colours to more complex features. As a result it will make famously absurd predictions like mistake a spotty fur coat for an actual leopard, since it has no "model" of a leopard, in the sense people go "this is an absurd place for a leopard to sit so it's most ceratinly a fur coat rather than an animal". Or to use use a more technical term, no prior probability of a leopard given observed data. Hence a standard convnet (if there is such a thing) will massively overestimate the probabilities of such "adversial" stimuli.

kristjankalmon New Startup Aims to Commercialize a Brain Prosthetic to Improve Memory

"An implanted memory prosthetic would have electrodes to record signals during learning, a microprocessor to do the computations, and electrodes that stimulate neurons to encode the information as a memory."

or in other words -- we record "signals", then compute something, and finally "stimulate" the neurons. I appreciate that this isn't a scientific report but this borders on parody.

The magnitude of technical, computational and biological problems to be solved here in order to achieve anything even remotely close to the headline would require 3-4 major scientific breakthroughs. With all respect, you don't publish them in IEEE proceedings.

kristjankalmon The most disruptive technology of the last century was home appliances

I think you severly underestimate the complexity of ironing and cleaning -- I would say wiping dust in a room full of objects (many of them fragile and of complex shape) is a significantly harder task for a robot than steering/parking/stopping a car. And ironing has been solved, though the machines dry cleaners use take too much room for most households to be practical,

kristjankalmon Stan: a probabilistic programming language

some useful books available as a PDF:

David Barber "Bayesian Reasoning and Machine Learning" http://web4.cs.ucl.ac.uk/staff/D.Barber/pmwiki/pmwiki.php?n=... -- Thorough and systematic, accompanied by a Matlab library

Marc Steyvers "Computational Statistics with Matlab" http://psiexp.ss.uci.edu/research/teachingP205C/205C.pdf -- Somewhat unfinished but very good hands-on guide for a total novice

David Draper "Bayesian Modeling, Inference and Prediction" https://users.soe.ucsc.edu/~draper/draper-BMIP-dec2005.pdf -- Quite good if you're already familiar with a field, might be slightly daunting for a beginner

and the classic, David MacKay's "Information Theory, Inference, and Learning Algorithms" http://www.inference.phy.cam.ac.uk/itprnn/book.html

kristjankalmon Technology has created more jobs than it has destroyed, says 140 years of data

can anyone find the link to the original paper? it seems to be by Deloitte but I have found nothing on their website nor just googling authors' names.

kristjankalmon In Zimbabwe, We Don’t Cry for Lions

Anthropocentrism at its purest. Reminds me of someone on BBC talking about "shark infested waters of South Africa".

kristjankalmon How Can There Still Be a Sex Difference, Even When There Is No Sex Difference?

hence, if you sample either tail of the distribution you get on the average more men than women. and sampling one of the tails is exactly what selective schools/companies/sports clubs/etc tend to do.

kristjankalmon Suddenly, a leopard print sofa appears

read the Hinton paper cited in the OP -- no amount of processing power will make CNNs represent structured latent variables

kristjankalmon Introducing Xubuntu core

support for wider range of hardware? backwards compatibility? the former has certainly improved a lot over the last couple of years

kristjankalmon What we know about the EmDrive and Cannae Drive

exactly -- I'd guess not too many non-physicits realise what's a prior for 'conservation of momentum violated'

kristjankalmon What we know about the EmDrive and Cannae Drive

the problem is that the excitement is generated by the use of exotic terminology. you wouldn't get excited if a single study reported that alzheimer's was reversed by copious praying and sprinkling of holy water.