“Deep Networks with Stochastic Depth”, Huang et al 2016 (dropout of random layers, which does not just work, but works better and enables >1200-layer residual networks. This might be related to Liao & Poggio 2016 claim that deep residual nets are effectively unrolled RNNs which apply slight variants of an algorithm to many timesteps.)
“The Bitter Fight Over the Benefits of Bilingualism” (So Bialystok is refusing a collaboration because a pre-registered protocol is unscientific and she refuses to work with someone so ‘biased’ that they would damage the non-pre-registered results, and besides, publication bias doesn’t exist—“there is absolutely no evidence”. I see...)
Interview with short seller Jim Chanos (Short sellers are always so interesting. Few people are so motivated to see through the miasma of lies and bias and self-serving optimism in the business and finance world.)
Have you written up somewhere how you stay organized, what software you use, especially with regards to reference management, text editors and works in progress?
I’ve been reading livetweets of the Oracle vs Google trial by jury, presided over by Judge Alsup, who learned to code. In a previous trial, Judge Alsup had declared that the software API in question was uncopyrightable. However, his decision was overturned by the federal court, and the Supreme Court declined to hear the case. So now they are redoing the trial with a new jury, with the premise that an API is copyrightable, in order to decide whether Google’s reimplementation of a Java API from Sun at the time (later bought by Oracle) was “fair use” of copyrightable materials.
The decision in this trial has the potential to impact the future of software development. The 9 billion dollars being asked for by Oracle is actually not the biggest thing at stake in this case. It is being decided by jurors who were picked because they know very little about how software development works.
Short Online Texts Thread
Everything is heritable:
“Genome-wide association study of cognitive functions and educational attainment in UK Biobank (n=112151)”, Davies et al 2016
Continuation of the recent phenome theme of inter-correlations of negative traits and inter-correlation of positive traits:
“Analysis of shared heritability in common disorders of the brain”, Anttila et al 2016
“Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses”, Okbay et al 2016
“Physical and neurobehavioral determinants of reproductive onset and success”, Day et al 2016
“Association between stressful life events and psychotic experiences in adolescence: evidence for gene–environment correlations”, Shakoor et al 2016
“Introducing precise genetic modifications into human 3PN embryos by CRISPR/Cas-mediated genome editing”, Kang et al 2016 (see also Liang et al 2015; both describe the CRISPR state of the art as of early 2014, 2+ years ago)
“Variants near CHRNA3/5 and APOE have age- and sex-related effects on human lifespan”, Joshi et al 2016
“National Happiness and Genetic Distance: A Cautious Exploration”, Proto & Oswald 2015
Friends are as genetically similar as fourth cousins (media)
“Monkey kingdom: China is positioning itself as a world leader in primate research”
Politics/religion:
Book Review: Albion’s Seed
“Peer review: Troubled from the start”
“My Year in San Francisco’s \$2 Million Secret Society Startup”
AI:
“Deep Networks with Stochastic Depth”, Huang et al 2016 (dropout of random layers, which does not just work, but works better and enables >1200-layer residual networks. This might be related to Liao & Poggio 2016 claim that deep residual nets are effectively unrolled RNNs which apply slight variants of an algorithm to many timesteps.)
“Generating Large Images from Latent Vectors”
“Foveation-based Mechanisms Alleviate Adversarial Examples”, Luo et al 2016
“Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing”, Esser et al 2016
Statistics/meta-science:
“A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook”, Gordon et al 2016 (How often does correlation=causation? <50% of the time in Internet advertising, with gross over/underestimates even with detailed covariate & advanced modeling.)
“The Bitter Fight Over the Benefits of Bilingualism” (So Bialystok is refusing a collaboration because a pre-registered protocol is unscientific and she refuses to work with someone so ‘biased’ that they would damage the non-pre-registered results, and besides, publication bias doesn’t exist—“there is absolutely no evidence”. I see...)
“Why the National Institutes of Health Should Replace Peer Review With a Lottery: A new study shows that peer-review scores for grant proposals are random anyway”
“Bandit based Monte-Carlo planning”, Kocsis & Szepesvári 2006
“How Kalman Filters Work, Part 1”
“The importance of ‘gold standard’ studies for consumers of research”
Psychology/biology:
“The Production of Human Capital in Developed Countries: Evidence from 196 Randomized Field Experiments”, Fryer 2016 (education; lots of small effects):
“A Meta-Analysis of Blood Glucose Effects on Human Decision Making”, Orquin & Kurzban 2016
“Is Ego-Depletion a Replicable Effect? A Forensic Meta-Analysis of 165 Ego Depletion Articles”
“Indoor air quality and academic performance”, Stafford 2015 (apropos of harms of CO2 discussion)
“Seven Pervasive Statistical Flaws in Cognitive Training Interventions”, Moreau et al 2016
“The Voyeur’s Motel”
Technology:
“What the game ‘Werewolf’ teaches us about Trust & Security”
The Raising of Chicago
Economics:
“Demographic Consequences of Defeating Aging”, Gavrilov & Gavrilova 2010
Interview with short seller Jim Chanos (Short sellers are always so interesting. Few people are so motivated to see through the miasma of lies and bias and self-serving optimism in the business and finance world.)
Have you written up somewhere how you stay organized, what software you use, especially with regards to reference management, text editors and works in progress?
oops, just posted this AI article in open...
https://aeon.co/essays/true-ai-is-both-logically-possible-and-utterly-implausible
On how whether grains vs. roots were eaten may have determined the success of ancient civilizations.
I’ve been reading livetweets of the Oracle vs Google trial by jury, presided over by Judge Alsup, who learned to code. In a previous trial, Judge Alsup had declared that the software API in question was uncopyrightable. However, his decision was overturned by the federal court, and the Supreme Court declined to hear the case. So now they are redoing the trial with a new jury, with the premise that an API is copyrightable, in order to decide whether Google’s reimplementation of a Java API from Sun at the time (later bought by Oracle) was “fair use” of copyrightable materials.
The decision in this trial has the potential to impact the future of software development. The 9 billion dollars being asked for by Oracle is actually not the biggest thing at stake in this case. It is being decided by jurors who were picked because they know very little about how software development works.
In the courtroom livetweeting the events are Sarah Jeong https://twitter.com/sarahjeong/ and Parker Higgins https://twitter.com/xor/ and a few other people, most of whom are on Sarah Jeong’s list: https://twitter.com/sarahjeong/lists/oracle-v-google or using the hashtag #googacle https://twitter.com/hashtag/googacle