via:phnk   50

Archive ouverte HAL - The Great Regression. Machine Learning, Econometrics, and the Future of Quantitative Social Sciences
"What can machine learning do for (social) scientific analysis, and what can it do to it? A contribution to the emerging debate on the role of machine learning for the social sciences, this article offers an introduction to this class of statistical techniques. It details its premises, logic, and the challenges it faces. This is done by comparing machine learning to more classical approaches to quantification – most notably parametric regression– both at a general level and in practice. The article is thus an intervention in the contentious debates about the role and possible consequences of adopting statistical learning in science. We claim that the revolution announced by many and feared by others will not happen any time soon, at least not in the terms that both proponents and critics of the technique have spelled out. The growing use of machine learning is not so much ushering in a radically new quantitative era as it is fostering an increased competition between the newly termed classic method and the learning approach. This, in turn, results in more uncertainty with respect to quantified results. Surprisingly enough, this may be good news for knowledge overall."

--- The correct line here is that 90%+ of "machine learning" is rebranded non-parametric regression, which is what the social sciences should have been doing all along anyway, because they have no good theories which suggest particular parametric forms. (Partial exceptions: demography and epidemiology.) If the resulting confidence sets are bigger than they'd like, that's still the actual range of uncertainty they need to live with, until they can reduce it with more and better empirical information, or additional constraints from well-supported theories. (Arguably, this was all in Haavelmo.) I look forward to seeing whether this paper grasps these obvious truths.
to:NB  to_read  regression  social_science_methodology  machine_learning  via:phnk  econometrics  to_be_shot_after_a_fair_trial 
august 2018 by cshalizi
A dive into the making of Immersion - Ctrl-Alt-Test
"Iteration time is king. The faster you can iterate, the more you can experiment, the more variations you can explore, the more you can refine your vision and increase the overall quality."
art  via:phnk 
april 2018 by rmsaksida
High-end art is one of the most manipulated markets in the world — Quartz
The nature of art as a commodity inherently makes efficient prices, meaning prices that reflect all available information about value, impossible. Value is subjective; the intrinsic value of a painting is paint and canvas—beyond that value is often a matter of taste.

This is why the industry has developed an intricate signaling process where the approval of a handful of galleries, collectors and museums, determines what is good and valuable. Dealers who own and work at art galleries invest many resources in building the artist’s brand.

But artists often take years to mature and have uneven periods, so any perception that an artist is over-hyped or overpriced can be anathema to his career. Value in art can be arbitrary but brands are fragile.
2018:art  nice-thinking  via:phnk 
january 2018 by mozzarella
Names of b.....s badder than Taylor Swift, a class in women's studies? · Maëlle
lol yesssss the function calling up quick women's history lessons at the end.
text-mining  via:phnk 
december 2017 by mozzarella
Machine learning fundamentals via linear regression
finally, a gentle one-sentence description of gradient descent in ML
Gradient descent enables a model to learn the gradient or direction that the model should take in order to reduce errors (differences between actual y and predicted y).
statistics  machine-learning  via:phnk 
december 2017 by mozzarella
LOL Nothing Matters: A defense of the internet’s absence of meaning | New Republic
I tweet with the best of them, and I like reading the hard stuff. I have a phone filled with novels, even some experimental ones. But the reality is that the most profound feeling of cultural participation for me comes from trawling databases.

a new old search engine to try:
history  web:social-networks  web:architecture  at_a_loss_for_tags  nice-thinking  via:phnk 
november 2017 by mozzarella
Wealth and Secular Stagnation: The Role of Industrial Organization and Intellectual Property Rights: RSF: Vol 2, No 6
"Changes in firm strategy and structure partially explain the sources and consequences of rising wealth inequality in America. Combining use of state-created monopolies around intellectual property rights (IPRs) for profitability and firm-level strategies to transform their industrial organization by pushing physical capital and noncore labor outside the boundaries of the firm leads to rising levels of wealth and income inequality among firms as well as individuals. Income inequality among firms in turn reduces growth in productive investment and thus in aggregate demand. Slower growth reflexively deters firms from new investment, aggravating the shortfall in aggregate demand. Decreased protection for IPRs and increased protection for subcontracted workers would help increase aggregate demand and thus push growth back to its prior level, as well as reducing wealth and income inequality among individuals."
to:NB  economics  inequality  class_struggles_in_america  market_failures_in_everything  via:phnk 
may 2017 by cshalizi
Toronto Rave Mixtape Archive
The point of this site is to bring together the great sets of the past and help to keep the vibes that make our city the place to be alive. From Trance to Rotterdam we have been working hard to bring you a bit of everything that the Toronto rave scene was about.
via:phnk  Toronto  music  history 
january 2017 by pgslr
Surfeit and surface | Big Data & Society
This is awesome. (But it's also completely compatible with causal inference!) Also, the cultural references will probably require footnotes in just 10 years.
social_science_methodology  sociology  data_mining  levi.john_martin  have_read  via:phnk  to_teach:undergrad-ADA  to_teach:data-mining  re:any_p-value_distinguishable_from_zero_is_insufficiently_informative  to:blog 
april 2016 by cshalizi

related tags

2018:art  @followup  abandoned  academia  academic-culture  academic-publishing  algorithms  archaeology  architecture  art  articles  artifacts  at_a_loss_for_tags  bad_data_analysis  bad_science  blogging  books  class_struggles_in_america  classifiers  community_discovery  cultural-norms  culture  data:science  data:visualization  data_journalism  data_mining  datasci-v5  datavis  design  e-books  econometrics  economic_history  economic_policy  economics  education  ellis.warren  emacs  epic_fail  estimation  ethics  explor  exploration  finance  financial_crisis_of_2007--  fraud  funny  funny:geeky  futurology  gisting  gives_economists_a_bad_name  glossing  great-works  have_read  healy.kieran  his_plan_was_sheer_elegance_in_its_simplicity  history  in_nb  inequality  interface_design  interview  jupyter  king.gary  levi.john_martin  literature  machine-learning  machine_learning  maps  market_failures_in_everything  memory  microblogging  minimalism  misspecification  model_checking  mongol_empire  moocs  moral_philosophy  mozz-v3  music  nanohistory  network_data_analysis  nice-thinking  nostalgia  ok_surprisingly_fragile_data_analysis  opinion  peer-review  peer_review  photography  political_economy  political_science  politics  programming  public_opinion  publishing  python  quality-control  r  re:adafaepov  re:any_p-value_distinguishable_from_zero_is_insufficiently_informative  reading  regression  relevance-theory  reputation  research  scholarly_communication  science_as_a_social_process  science_fiction  social_networks  social_science_methodology  socialmedia  socialnetworks  sociology  something_about_america  stata  statistics  surveys  tech:security  text-mining  theme  to:blog  to:nb  to_be_shot_after_a_fair_trial  to_read  to_teach  to_teach:baby-nets  to_teach:data-mining  to_teach:statcomp  to_teach:undergrad-ada  toread  toronto  track_down_references  trust  truth  us_politics  usa:surveillance  user-experience  vis  visual_display_of_quantitative_information  web:architecture  web:social-networks  what_is_best_in_life  why_oh_why_cant_we_have_a_better_press_corps  word-processing  workflow  writing 

Copy this bookmark: