newsletter/2015/12 (Link Bibliography)

“newsletter/​2015/​12” links:

  1. https://gwern.substack.com

  2. 11

  3. Changelog

  4. https://www.patreon.com/gwern

  5. Longevity#metformin

  6. Timestamping

  7. https://www.nature.com/mp/journal/vaop/ncurrent/full/mp2015184a.html

  8. 2015-johnson.pdf

  9. 1969-jensen.pdf: ⁠, Arthur R. Jensen (1969-05-01; iq):

    Arthur Jensen argues that the failure of recent compensatory education efforts to produce lasting effects on children’s IQ and achievement suggests that the premises on which these efforts have been based should be reexamined. He begins by questioning a central notion upon which these and other educational programs have recently been based: that IQ differences are almost entirely a result of environmental differences and the cultural bias of IQ tests. After tracing the history of IQ tests, Jensen carefully defines the concept of IQ, pointing out that it appears as a common factor in all tests that have been devised thus far to tap higher mental processes. Having defined the concept of intelligence and related it to other forms of mental ability, Jensen employs an analysis of model to explain how IQ can be separated into genetic and environmental components. He then discusses the concept of “heritability”, a statistical tool for assessing the degree to which individual differences in a trait like intelligence can be accounted for by genetic factors. He analyzes several lines of evidence which suggest that the heritability of intelligence is quite high (ie., genetic factors are much more important than environmental factors in producing IQ differences). After arguing that environmental factors are not nearly as important in determining IQ as are genetic factors, Jensen proceeds to analyze the environmental influences which may be most critical in determining IQ. He concludes that prenatal influences may well contribute the largest environmental influence on IQ. He then discusses evidence which suggests that social class and racial variations in intelligence cannot be accounted for by differences in environment but must be attributed partially to genetic differences. After he has discussed the influence on the distribution of IQ in a society on its functioning, Jensen examines in detail the results of educational programs for young children, and finds that the changes in IQ produced by these programs are generally small. A basic conclusion of Jensen’s discussion of the influence of environment on IQ is that environment acts as a “threshold variable.” Extreme environmental deprivation can keep the child from performing up to his genetic potential, but an enriched educational program cannot push the child above that potential. Finally, Jensen examines other mental abilities that might be capitalized on in an educational program, discussing recent findings on diverse patterns of mental abilities between ethnic groups and his own studies of associative learning abilities that are independent of social class. He concludes that educational attempts to boost IQ have been misdirected and that the educational process should focus on teaching much more specific skills. He argues that this will be accomplished most effectively if educational methods are developed which are based on other mental abilities besides IQ.

  10. http://aje.oxfordjournals.org/content/156/11/985.full

  11. ⁠, Sira Karvinen, Katja Waller, Mika Silvennoinen, Lauren G. Koch, Steven L. Britton, Jaakko Kaprio, Heikki Kainulainen, Urho M. Kujala (2015-12-15):

    Observational studies report a strong inverse relationship between leisure-time physical activity and all-cause mortality. Despite suggestive evidence from population-based associations, scientists have not been able to show a beneficial effect of physical activity on the risk of death in controlled intervention studies among individuals who have been healthy at baseline. On the other hand, high cardiorespiratory fitness is known to be a strong predictor of reduced mortality, even more robust than physical activity level itself.

    Here, in both animals and/​​​​or human twins, we show that the same genetic factors influence physical activity levels, cardiorespiratory fitness and risk of death. Previous observational follow-up studies in humans suggest that increasing fitness through physical activity levels could prolong life; however, our controlled interventional study with laboratory rats bred for low and high intrinsic fitness contrast with these findings. Also, we find no evidence for the suggested association using pairwise analysis among monozygotic twin pairs who are discordant in their physical activity levels.

    Based on both our animal and human findings, we propose that genetic pleiotropy might partly explain the frequently observed associations between high baseline physical activity and later reduced mortality in humans.

    [See also author’s later thesis, ⁠.]

  12. 2015-rottensteiner.pdf: ⁠, Mirva Rottensteiner, Tuija Leskinen, Eini Niskanen, Sari Aaltonen, Sara Mutikainen, Jan Wikgren, Kauko Heikkilä, Vuokko Kovanen, Heikki Kainulainen, Jaakko Kaprio, Ina Tarkka, Urho Kujala (2015; genetics):

    Purpose: The main aim of the present study (FITFATTWIN) was to investigate how physical activity level is associated with body composition, glucose homeostasis, and brain morphology in young adult male monozygotic twin pairs discordant for physical activity.

    Methods: From a population-based twin cohort, we systematically selected 10 young adult male monozygotic twin pairs (age range, 32–36 yr) discordant for leisure time physical activity during the past 3 yr. On the basis of interviews, we calculated a mean sum index for leisure time and commuting activity during the past 3 yr (3-yr LTMET index expressed as MET-hours per day). We conducted extensive measurements on body composition (including fat percentage measured by dual-energy x-ray absorptiometry), glucose homeostasis including homeostatic model assessment index and insulin sensitivity index (Matsuda index, calculated from glucose and insulin values from an oral glucose tolerance test), and whole brain magnetic resonance imaging for regional volumetric analyses.

    Results: According to pairwise analysis, the active twins had lower body fat percentage (p = 0.029) and homeostatic model assessment index (p = 0.031) and higher Matsuda index (p = 0.021) compared with their inactive co-twins. Striatal and prefrontal cortex (subgyral and inferior frontal gyrus) brain gray matter volumes were larger in the nondominant hemisphere in active twins compared with those in inactive co-twins, with a statistical threshold of p < 0.001.

    Conclusions: Among healthy adult male twins in their mid-30s, a greater level of physical activity is associated with improved glucose homeostasis and modulation of striatum and prefrontal cortex gray matter volume, independent of genetic background. The findings may contribute to later reduced risk of type 2 diabetes and mobility limitations.

  13. https://decorrespondent.nl/3789/Operation-Easy-Chair-or-how-a-little-company-in-Holland-helped-the-CIA-bug-the-Russians/106823277-f4300cc3

  14. https://web.archive.org/web/20131229234049/http://squid314.livejournal.com/338026.html

  15. https://web.archive.org/web/20131229234045/http://squid314.livejournal.com/338607.html

  16. https://www.theatlantic.com/international/archive/2012/10/perplexed-perplexed-on-mob-justice-in-nigeria/264006/

  17. http://mlg.eng.cam.ac.uk/yarin/blog_3d801aa532c1ce.html

  18. ⁠, Juergen Schmidhuber (2015-11-30):

    This paper addresses the general problem of reinforcement learning (RL) in partially observable environments. In 2013, our large RL recurrent neural networks (RNNs) learned from scratch to drive simulated cars from high-dimensional video input. However, real brains are more powerful in many ways. In particular, they learn a predictive model of their initially unknown environment, and somehow use it for abstract (e.g., hierarchical) planning and reasoning. Guided by algorithmic information theory, we describe RNN-based AIs (RNNAIs) designed to do the same. Such an RNNAI can be trained on never-ending sequences of tasks, some of them provided by the user, others invented by the RNNAI itself in a curious, playful fashion, to improve its -based world model. Unlike our previous model-building RNN-based RL machines dating back to 1990, the RNNAI learns to actively query its model for abstract reasoning and planning and decision making, essentially “learning to think.” The basic ideas of this report can be applied to many other cases where one RNN-like system exploits the algorithmic information content of another. They are taken from a grant proposal submitted in Fall 2014, and also explain concepts such as “mirror neurons.” Experimental results will be described in separate papers.

  19. 2004-duncan.pdf

  20. Correlation

  21. 2008-gonzalezalvarez.pdf: “Science in the 21<sup>st<  /​ ​​ ​sup> century: social, political, and economic issues”⁠, Juan R. González-Álvarez

  22. ⁠, Martin E. Héroux, Janet L. Taylor, Simon C. Gandevia (2015-11-13):

    The magnitude and direction of reported physiological effects induced using transcranial magnetic stimulation (TMS) to modulate human motor cortical excitability have proven difficult to replicate routinely. We conducted an online survey on the prevalence and possible causes of these reproducibility issues. A total of 153 researchers were identified via their publications and invited to complete an anonymous internet-based survey that asked about their experience trying to reproduce published findings for various TMS protocols. The prevalence of questionable research practices known to contribute to low reproducibility was also determined. We received 47 completed surveys from researchers with an average of 16.4 published papers (95% 10.8–22.0) that used TMS to modulate motor cortical excitability. Respondents also had a mean of 4.0 (2.5–5.7) relevant completed studies that would never be published. Across a range of TMS protocols, 45–60% of respondents found similar results to those in the original publications; the other respondents were able to reproduce the original effects only sometimes or not at all. Only 20% of respondents used formal power calculations to determine study sample sizes. Others relied on previously published studies (25%), personal experience (24%) or flexible post-hoc criteria (41%). Approximately 44% of respondents knew researchers who engaged in questionable research practices (range 32–70%), yet only 18% admitted to engaging in them (range 6–38%). These practices included screening subjects to find those that respond in a desired way to a TMS protocol, selectively reporting results and rejecting data based on a gut feeling. In a sample of 56 published papers that were inspected, not a single questionable research practice was reported. Our survey revealed that approximately 50% of researchers are unable to reproduce published TMS effects. Researchers need to start increasing study sample size and eliminating—or at least reporting—questionable research practices in order to make the outcomes of TMS research reproducible.

  23. https://www.nytimes.com/2012/10/28/magazine/safety-lessons-from-the-morgue.html?_r=0&pagewanted=all

  24. ⁠, Anna Johnston, Michael Gradisar, Hayley Dohnt, Michael Billows, Stephanie McCappin (2010):

    Fluid intelligence involves novel problem-solving and may be susceptible to poor sleep. This study examined relationships between adolescent sleep, fluid intelligence, and academic achievement. Participants were 217 adolescents (42% male) aged 13 to 18 years (mean age, 14.9 years; SD = 1.0) in grades 9–11. Fluid intelligence was predicted to mediate the relationship between adolescent sleep and academic achievement. Students completed online questionnaires of self-reported sleep, fluid intelligence (Letter Sets and Number Series), and self-reported grades. Total sleep time was not statistically-significantly related to fluid intelligence nor academic achievement (both p > 0.05); however, sleep difficulty (eg. difficulty initiating sleep, unrefreshing sleep) was related to both (p < 0.05). The strength of the relationship between sleep difficulty and grades was reduced when fluid intelligence was introduced into the model; however, the z-score was not to confirm mediation.Nevertheless, fluid intelligence is a cognitive ability integral in academic achievement, and in this study has been shown it to be susceptible to sleep impairments (but not duration) in adolescents.

  25. https://www.theatlantic.com/technology/archive/2012/10/scientists-recover-the-sounds-of-19th-century-music-and-laughter-from-the-oldest-playable-american-recording/264147/

  26. http://www.cs.uic.edu/~i377/MacKenzie-Computer-related-accidental-death-an-empirical-exploration.pdf

  27. http://richardcarrier.blogspot.com/2007/07/experimental-history.html

  28. http://www.ffi.no/no/Rapporter/14-02234.pdf

  29. http://openborders.info/blog/how-persuasive-are-open-borders-advocates-the-case-of-bryan-caplan/

  30. 1951-borges-theargentinewriterandtradition.pdf: ⁠, Jorge Luis Borges (1951; borges):

    Borges considers the problem of whether Argentinian writing on non-Argentinian subjects can still be truly “Argentine.” His conclusion: …We should not be alarmed and that we should feel that our patrimony is the universe; we should essay all themes, and we cannot limit ourselves to purely Argentine subjects in order to be Argentine; for either being Argentine is an inescapable act of fate—and in that case we shall be so in all events—or being Argentine is a mere affectation, a mask. I believe that if we surrender ourselves to that voluntary dream which is artistic creation, we shall be Argentine and we shall also be good or tolerable writers.

  31. https://www.poetryfoundation.org/poem/248068

  32. http://www.mcsweeneys.net/articles/back-from-yet-another-globetrotting-adventure-indiana-jones-checks-his-mail-and-discovers-that-his-bid-for-tenure-has-been-denied

  33. http://craftsmanship.net/the-bonsai-kid/

  34. Movies#star-wars-the-force-awakens

  35. Anime#how-the-grinch-stole-christmas

  36. Anime#fsn-unlimited-blade-works

  37. Anime#brave

  38. Anime#a-charlie-brown-christmas