Simple-CV

Slide 1
• Science & Nonduality conference – California (2016)

• Marie Curie Actions

Sky ain't the limit!
DeepDream is a psychophysical AI experiment that visualizes the patterns learned by a convoluted neural network. Similar to when a perceptually naive child watches clouds and tries to interpret random shapes, DeepDream over-interprets and enhances specific salient statistical patterns it detects in an image. It does so by forwarding an image through the Bayesian network, then calculating the gradient of the image with respect to the activations of a particular neighbouring layer. The image is then transformed to increase these activations, enhancing and perturbing the patterns seen by the network, and resulting in a dream-like image. This process was dubbed “Inceptionism” (a reference to InceptionNet, and the movie Inception). InceptionNet: https://arxiv.org/pdf/1409.4842.pdf
```import tensorflow as tf
import numpy as np
import matplotlib as mpl
import IPython.display as display
import PIL.Image
name = url.split('/')[-1]
image_path = tf.keras.utils.get_file(name, origin=url)
img = PIL.Image.open(image_path)
if max_dim:
img.thumbnail((max_dim, max_dim))
return np.array(img)

# Normalize an image
def deprocess(img):
img = 255*(img + 1.0)/2.0
return tf.cast(img, tf.uint8)

# Display an image
def show(img):
display.display(PIL.Image.fromarray(np.array(img)))

# Downsizing the image makes it easier to work with.
show(original_img)
display.display(display.HTML('Image cc-by: Von.grzanka'))

base_model = tf.keras.applications.InceptionV3(include_top=False, weights='imagenet')

# Maximize the activations of these layers
names = ['mixed3', 'mixed5']
layers = [base_model.get_layer(name).output for name in names]

# Create the feature extraction model
dream_model = tf.keras.Model(inputs=base_model.input, outputs=layers)

def calc_loss(img, model):
# Pass forward the image through the model to retrieve the activations.
# Converts the image into a batch of size 1.
img_batch = tf.expand_dims(img, axis=0)
layer_activations = model(img_batch)
if len(layer_activations) == 1:
layer_activations = [layer_activations]

losses = []
for act in layer_activations:
loss = tf.math.reduce_mean(act)
losses.append(loss)

return  tf.reduce_sum(losses)

```
This tutorial contains a minimal implementation of DeepDream, as described by Alexander Mordvintsev.

Namarupa
Panta Rhei

A powerful & flexible methodological alternative to mindless orthodox (paralogistic)
statistical null hypothesis significance testing (NHST) based on the fallacious/invalid logic associated with p-values and fixed α-levels.
Bayesian à posteriori parameter estimation
via Markov chain Monte Carlo simulations

Mindless statistical rituals in science
A critique of Fisherian p-values
"Few researchers are aware that their own heroes rejected what they practice routinely. Awareness of the origins of the ritual and of its rejection could cause a virulent cognitive dissonance, in addition to dissonance with editors, reviewers, and dear colleagues. Suppression of conflicts and contradicting information is in the very nature of this social ritual.”
(Gigerenzer, 2004, p. 592)
***
Gigerenzer, G. (2004). Mindless statistics. The Journal of Socio-Economics, 33(5), 587–606. https://doi.org/10.1016/j.socec.2004.09.033

Display R code & references
Associated R codeBayesian paramter estimation via Markov chain Monte Carlo methods
```#clears all of R's memory
rm(list=ls())
# Get the functions loaded into R's working memory
# The function can also be downloaded from the following URL: # http://irrational-decisions.com/?page_id=1996
source("BEST.R")
dataexp2 <-
# Specify data as vectors
y1 = c(dataexp2\$v00)
y2 = c(dataexp2\$v01)
# Run the Bayesian analysis using the default broad priors described by Kruschke (2013)
mcmcChain = BESTmcmc( y1 , y2 , priorOnly=FALSE ,
numSavedSteps=12000 , thinSteps=1 , showMCMC=TRUE )
postInfo = BESTplot( y1 , y2 , mcmcChain , ROPEeff=c(-0.1,0.1) )

#The function “BEST.R” can be downloaded from the CRAN (Comprehensive R Archive Network) repository under https://cran.r-project.org/web/packages/BEST/index.html
```
Kruschke, J. K., & Liddell, T. M.. (2018). The Bayesian New Statistics: Hypothesis testing, estimation, meta-analysis, and power analysis from a Bayesian perspective. Psychonomic Bulletin & Review, 25(1), 178–206.

Plain numerical DOI: 10.3758/s13423-016-1221-4
DOI URL
Amazonian Rainforest (Indigenous Quechuan language: "Pachamama")
It's like a jungle sometimes -
it makes me wonder how I keep from going under...
00:00 / 00:00

The Amazon deforestation rate in 2020 is the greatest of the decade.

I suggest that humanity should focus on saving the rainforest and not on fuzzy, propagandistic and theory-laden concepts such as "global warming". The rainforest is a clear quantitative indicator of "planetary health". It is thus a statistical criterion which can be measured with great accuracy. Complex systems thinking is needed. Interconnectivity is a key principle of life. The rainforest demonstrates this principle, i.e., the rainforest is an intelligent super-organism.

The Big Apple (almost eaten...)
NYC
It's like a jungle sometimes -
it makes me wonder how I keep from going under...
00:00 / 00:00
HORIZONS Conference 2018
New York
00:00 / 00:00
Twin Tower memorial - New York City
*in memoriam of the satanic Deep State
"Here are enshrined the longing of great hearts and noble things that tower above the tide, the magic word that winged wonder starts, the garnered wisdom that never dies."
~Roscoe C. Brown
Science and Nonduality Conference
Dolce Hayes Mansion
in Californias Silicon Valley
2016
Thompson, C., & Williams, M. L.. (2022). Review of the physiological effects of Phyllomedusa bicolor skin secretion peptides on humans receiving Kambô. Toxicology Research and Application, 6, 239784732210857.

Plain numerical DOI: 10.1177/23978473221085746
DOI URL
Schmidt, T. T., Reiche, S., Hage, C. L. C., Bermpohl, F., & Majić, T.. (2020). Acute and subacute psychoactive effects of Kambô, the secretion of the Amazonian Giant Maki Frog (Phyllomedusa bicolor): retrospective reports. Scientific Reports, 10(1), 21544.

Plain numerical DOI: 10.1038/s41598-020-78527-4
DOI URL
Phyllomedusa bicolor aka. Kambô
Play Kambô Call
00:00 / 00:00
Thompson, C., & Williams, M. L.. (2022). Review of the physiological effects of Phyllomedusa bicolor skin secretion peptides on humans receiving Kambô. Toxicology Research and Application, 6, 239784732210857.

Plain numerical DOI: 10.1177/23978473221085746
DOI URL
Schmidt, T. T., Reiche, S., Hage, C. L. C., Bermpohl, F., & Majić, T.. (2020). Acute and subacute psychoactive effects of Kambô, the secretion of the Amazonian Giant Maki Frog (Phyllomedusa bicolor): retrospective reports. Scientific Reports, 10(1), 21544.

Plain numerical DOI: 10.1038/s41598-020-78527-4
DOI URL
Phyllomedusa bicolor aka. Kambô
Play Kambô Call
00:00 / 00:00
~Jungle fever~
Iquitos
Peru
2022
♫ Marinera
Nature One - 2013
Raketenbasis 'Pydna' Kastellaun
PORTUGAL - Algarve - 2021
9th International Quantum Interactions conference (QI15)
Switzerland, Filzbach 2015
Iquitos 2022
Peruvian Amazon
Maloca in the Amazonian Rainforest (2022)
Mauritia flexuosa
aka. Aguaje
Uncaria tomentosa
aka. Uña de Gato
Ethnobotany & Ethnomedicine
Big Pharma
Straight
from the Jungle
Vario of Iquitos (2022)
Manipal University Jaipur
India 2016

In anthropology (and the social and behavioral sciences) the terms emic and etic refer to two different types of field research. Emic is a perspective from within the social group (from the perspective of the subject). Per contrast, etic refers to an outside-perspective (from the perspective of the observer). In my opinion both are complementary to each other (in the quantum-physical sense of complementarity; cf. bistable perception).

Psychophysics meets Quantum Cognition

This presentation focuses on the role of noncommutativity in visual/perceptual decision-making.
First, some historical background information is provided.
Then, the interrelated concepts of complementarity and superposition are briefly delineated and two paradigmatic visual illusions are demonstrated.
Next, several pertinent empirical results are discussed in the theoretical framework of quantum cognition.
Finally, the interdisciplinary scope of the topic will be adumbrated in the context of "cognitive innovation" (CogNovo).

Switch to fullscreen-mode

An animated semantic MindMap
00:00 / 00:00
Elliptical insights:
Visualising data in 3-dimensional Cartesian space
Expand this overlay to display additional information
Alexander von Humboldt on geometrical cognitionAssociated R codeR package on CRAN'Elliptical Insights: Understanding Statistical Methods through Elliptical Geometry
Whatever relates to extent and quantity may be represented by geometrical figures. Statistical projections which speak to the senses without fatiguing the mind, possess the advantage of fixing the attention on a great number of important facts. ~ Alexander von Humboldt (1811)
```scatterplot3d(x, y=NULL, z=NULL, color=par("col"), pch=par("pch"),
main=NULL, sub=NULL, xlim=NULL, ylim=NULL, zlim=NULL,
xlab=NULL, ylab=NULL, zlab=NULL, scale.y=1, angle=40,
axis=TRUE, tick.marks=TRUE, label.tick.marks=TRUE,
x.ticklabs=NULL, y.ticklabs=NULL, z.ticklabs=NULL,
lab.z=mean(lab[1:2]), type="p", highlight.3d=FALSE,
mar=c(5,3,4,3)+0.1, bg=par("bg"), col.axis=par("col.axis"),
col.grid="grey", col.lab=par("col.lab"),
cex.symbols=par("cex"), cex.axis=0.8 * par("cex.axis"),
cex.lab=par("cex.lab"), font.axis=par("font.axis"),
font.lab=par("font.lab"), lty.axis=par("lty"),
lty.grid=par("lty"), lty.hide=NULL, lty.hplot=par("lty"),
log="", asp=NA, ...)
```
Friendly, M., Monette, G., & Fox, J.. (2013). Elliptical Insights: Understanding Statistical Methods through Elliptical Geometry. Statistical Science

Plain numerical DOI: 10.1214/12-STS402
DOI URL

The Möbius band is an extraordinary geometrical figure. The band is eponymously named after the German mathematician August Ferdinand Möbius who described it in 1885, contemporaneously with another German mathematician named Johann Benedict Listing. It is a so called ruled surface with only one side and one boundary and it possesses the mathematical property of non-orientability (viz., a non-orientable manifold). In fact, the Möbius band is the simplest possible non-orientable surface. A Gedankenexperiment is helpful to understand this property intuitively: Imagine walking on the surface of a giant Möbius band. If you would travel long enough you would end up at the very starting point of the journey, only mirror-reversed. This journey can be repeated - ad infinitum. I argue that the Möbius band provides a reasily accessible metaphor for dual-aspect monism, a theory which challenges the predominant view that mind & matter (i.e., psyche & physis) are two fundamentally different substances. More information can be found on my eponymous websites.

Science and Nonduality Conference
California, San Jose (Silicon Valley)
2016
सरस्वति नमस्तुभ्यं वरदे कामरूपिणि
Visualisation of various MCMC convergence diagnostics for μ1

Trace plot: In order to examine the representativeness of the MCMC samples, we first visually examine the trajectory of the chains. The trace plot indicates convergence on θ, i.e., the trace plot appears to be stationary because its mean and variance are not changing as a function of time.

Shrink factor (Brooks-Gelman-Rubin statistic). Theoretically, the larger the number of iterations T, the closer 𝑅 should approximate 1 (as T → ∞, 𝑅→ 1).

Autocorrelation (effective sample size/EES)

The density plot entails the 95% HDI and it displays the numerical value of the Monte Carlo Standard Error (MCSE) of 0.000454. The Monte Carlo Error (MCSE) is the uncertainty which can be attributed to the fact that the number of simulation draws is always finite. In other words, it provides a quantitative index that represents the quality of parameter estimates. The MCSE package in R provides convenient tools for computing Monte Carlo standard errors and the effective sample size (Gelman et al., 2004).

Expand overlay
Associated R codeBayesian paramter estimation via Markov chain Monte Carlo methods
```#clears all of R's memory
rm(list=ls())
# Get the functions loaded into R's working memory
# The function can also be downloaded from the following URL: # http://irrational-decisions.com/?page_id=1996
source("BEST.R")
dataexp2 <-
# Specify data as vectors
y1 = c(dataexp2\$v00)
y2 = c(dataexp2\$v01)
# Run the Bayesian analysis using the default broad priors described by Kruschke (2013)
mcmcChain = BESTmcmc( y1 , y2 , priorOnly=FALSE ,
numSavedSteps=12000 , thinSteps=1 , showMCMC=TRUE )
postInfo = BESTplot( y1 , y2 , mcmcChain , ROPEeff=c(-0.1,0.1) )

#The function “BEST.R” can be downloaded from the CRAN (Comprehensive R Archive Network) repository under https://cran.r-project.org/web/packages/BEST/index.html
```
Kruschke, J. K., & Liddell, T. M.. (2018). The Bayesian New Statistics: Hypothesis testing, estimation, meta-analysis, and power analysis from a Bayesian perspective. Psychonomic Bulletin & Review, 25(1), 178–206.

Plain numerical DOI: 10.3758/s13423-016-1221-4
DOI URL
Kruschke, J. K.. (2011). Bayesian Assessment of Null Values Via Parameter Estimation and Model Comparison. Perspectives on Psychological Science, 6(3), 299–312.

Plain numerical DOI: 10.1177/1745691611406925
DOI URL
Accumulation of evidence in favor of H₁
Sequential Bayes Factor analysis depicting the flow of evidence for various priors as n accumulates over time.
Bayes Factor robustness check for various Cauchy priors
In this example the maximum Bayes Factor was obtained at r ≈ 0.28 (max BF₁₀ ≈ 12.56).
Bayes Factor analysis: Prior and posterior plot
For this pairwise comparison we obtain a Bayes Factor of BF₁₀ ≈ 9.12 indicating that the data are circa 9 times more likely under H1 than under H0, i.e., P(D│H₁) ≈ 9.12.
Model comparison using Bayes Factor analysis
Expand overlay to display additional information
JASP is based on the ‘BayesFactor’ R package which can be downloaded from the following URL: https://cran.r-project.org/web/packages/BayesFactor/index.html
JASP is an open-source statistics program that is free, friendly, and flexible. Armed with an easy-to-use GUI, JASP allows both classical and Bayesian analyses. URL: https://jasp-stats.org
Wagenmakers, E. J., Love, J., Marsman, M., Jamil, T., Ly, A., Verhagen, J., … Morey, R. D.. (2018). Bayesian inference for psychology. Part II: Example applications with JASP. Psychonomic Bulletin and Review

Plain numerical DOI: 10.3758/s13423-017-1323-7
DOI URL

This animation which was created using R provides an intuitive explanation of what a 95% confidence interval really means: If we would repeat the exact same experiment 50 times, then 95 % of the time the 50 confidence intervals would contain the true mean. Visualisations are a powerful aid to understanding statistics. This makes sense from an evolutionary point of view as abstract symbolic cognition is phylogenetically much more recent than visual reasoning.

The American Psychological Association recommend confidence intervals as the "new statistics" in order to counteract the problems associuated with p-values. However, research clearly shows that confidence intervals are also widely misunderstood and they are therefore no real solution to the statistical crisis.
Out of 118 researcher only 3% were able to give the correct answer.
http://www.ejwagenmakers.com/inpress/HoekstraEtAlPBR.pdf

Visualising and understanding confidence intervals
Expand overlay
Associated R codeBayesian paramter estimation via Markov chain Monte Carlo methods
```> library(animation)
> conf.int
function (level = 0.95, size = 50, cl = c("red", "gray"), ...)
{
n = ani.options("nmax")
d = replicate(n, rnorm(size))
m = colMeans(d)
....
```
Empirical research has shown that confidence are widely misunderstood. As with p-values most professional researchers do not understand the logic behind confidence intervals and their misinterpretation is a robust finding.
Hoekstra, R., Morey, R. D., Rouder, J. N., & Wagenmakers, E.-J.. (2014). Robust misinterpretation of confidence intervals. Psychonomic Bulletin & Review, 21(5), 1157–1164.

Plain numerical DOI: 10.3758/s13423-013-0572-3
DOI URL
The animate package in R is a great way to visualise and understand the logic behind confidence intervals intuitively. The idea behind this simulation is simple: draw samples (random numbers) from the population which follows N(0, 1), and calculate confidence intervals (CI) based on these samples respectively.
Thesis abstract
Quantum cognition is an interdisciplinary emerging field within the cognitive sciences which applies various axioms of quantum mechanics to cognitive processes. This thesis reports the results of several empirical investigations which focus on the applicability of quantum cognition to psychophysical perceptual processes. Specifically, we experimentally tested several a priori hypotheses concerning 1) constructive measurement effects in sequential perceptual judgments and 2) noncommutativity in the measurement of psychophysical observables. In order to establish the generalisability of our findings, we evaluated our prediction across different sensory modalities (i.e., visual versus auditory perception) and in cross-cultural populations (United Kingdom and India). Given the well-documented acute “statistical crisis” in science (Loken & Gelman, 2017) and the various paralogisms associated with Fisherian/Neyman-Pearsonian null hypothesis significance testing, we contrasted various alternative statistical approaches which are based on complementary inferential frameworks (i.e., classical null hypothesis significance testing, nonparametric bootstrapping, model comparison based on Bayes Factors analysis, Bayesian bootstrapping, and Bayesian parameter estimation via Markov chain Monte Carlo simulations). This multimethod approach enabled us to analytically cross-validate our experimental results, thereby increasing the robustness and reliability of our inferential conclusions. The findings are discussed in an interdisciplinary context which synthesises knowledge from several prima facie separate disciplines (i.e., psychology, quantum physics, neuroscience, and philosophy). We propose a radical reconceptualization of various epistemological and ontological assumptions which are ubiquitously taken for granted (e.g., naïve and local realism/cognitive determinism). Our conclusions are motivated by recent cutting-edge findings in experimental quantum physics which are incompatible with the materialistic/deterministic metaphysical Weltanschauung internalised by the majority of scientists. Consequently, we argue that scientists need to update their nonevidence-based implicit beliefs in the light of this epistemologically challenging empirical evidence. [/responsivevoice]

X

MindMap

This is my beautiful "steel tongue drum" (a fantastic handmade percussion instrument). It creates amazing sounds (the harmonic scale is D-minor consisting of the pitches 𝄞 D, E, F, G, A, B♭). The tonality deeply penetrates the mind. Listen to it for yourself and enjoy the vibrations... 🎜

00:00 / 00:00

This is the flower of Albizia julibrissin, the Persian silk tree. I took this picture in 2016 in Plymouth, English Garden, Mount Edgcumbe. Not to be confused with Calliandra angustifolia aka. Bobinsana (a "healing plant" used in shamanism) which is in the Fabaceae family, too.

is virtually transformed by a "psychedelically inspired" Bayesian computer algorithm.
A magnificant flower
DeepDream is a "creative" computer algorithm developed by Google. According to the developers it is "psychedelically inspired".

I created this website because cognitive liberty (freedom of thought) is a fundamental human right which is currently under heavy attack. The website focuses on neuropolitics, history, social psychology, and cognitive psychology, inter alia.

Blue Room ~
Jaipur, India
Kathakali
Kerala
AI generated
AI generated

Sanskrit etymology:
nata = actor, dancer, mime
raja = king
asana = posture, seat

Nataraja is another name for Shiva and his dance symbolizes cosmic energy. It is a challenging balancing asana (cerebellum) which trains focus and concentration (self-control/prefrontal executive functions). It is an extroverted asana full of creative energy.

Click to close this text

The word psyche is etymologically derived from the ancient Greek ψυχή (psukhḗ, which translates into “mind/soul/spirit/breath”). The suffix "logia" (λογία) can be translated as "the study of" (cf. lógos/λόγος which can be be translated, inter alia, as "subject matter Hence, psychology literally means "the study of mind/soul/spirit/breath”. However, many psychologists are utterly nescient of this etymological definition and are not very comfortable with these profound philosophical concepts. There are some laudable exceptions, for instance, the Swiss depth-psychologist C.G. Jung who wrote extensively on Indian psychology and pranayama (viz., psycho-spiritual breathing-exercises). Jung also coined the terms introversion/extroversion which are today widely utilised in mainstream psychology (the 5-factor model of personality). It should be noted that there is no science without philosophy!The notion that science can be seperated from philosophy is a naïve positivistic illusion which completly neglects the history of science. Until quite recently science was philosophy. The terminological dichotomisation is a modern invention.Daniel Dennett formulated the following concise statement: “There is no such thing as philosophy-free science; there is only science whose philosophical baggage is taken on board without examination." — Daniel Dennett, Darwin's Dangerous Idea, 1995

R code

Keywords: Statistical computingOpen-source softwareBayesian analysisMarkov chain Monte Carlo methodsData visualisationLogical inferenceHypothesis testingDeductive reasoningAbductionCredibility intervalsProbability theoryDecision algorithmsNew statisticsReplication crisisCreative statistics

Search R documentation

Frequentist inference
Fixed-effects ANOVAGeneral linear models: mixing continuous and categorical covariatesOutput plot as PDFSimple linear regressionSkewness & Kurtosis
```data(ToothGrowth)

## Example plot from ?ToothGrowth

coplot(len ~ dose | supp, data = ToothGrowth, panel = panel.smooth,
xlab = "ToothGrowth data: length vs dose, given type of supplement")
## Treat dose as a factor
ToothGrowth\$dose = factor(ToothGrowth\$dose)
levels(ToothGrowth\$dose) = c("Low", "Medium", "High")

summary(aov(len ~ supp*dose, data=ToothGrowth))

#install.packages("xtable")
library(xtable)
xtable(x, caption = NULL, label = NULL, align = NULL, digits = NULL,
display = NULL, auto = FALSE, ...)

print(xtable(d), type="html")
print(xtable(d), type="latex") # anova table to latex
#https://cran.r-project.org/web/packages/xtable/index.html
#https://rmarkdown.rstudio.com/

```
```data(ToothGrowth)

# model log2 of dose instead of dose directly
ToothGrowth\$dose = log2(ToothGrowth\$dose)

# Classical analysis for comparison
lmToothGrowth <- lm(len ~ supp + dose + supp:dose, data=ToothGrowth)
summary(lmToothGrowth)
```
```x<- (1:5)
y<- (1:111)
pdf(file=file.choose())
hist(x)
plot(x, type='o')
dev.off()
```

Notes
`file.choose()` is a very handy command which saves the work associated with defining absolute and relative paths which can be quite cumbersome.
`list.files` for non-interactive selection.
`choose.files` for selecting multiple files interactively.
More
https://www.rdocumentation.org/packages/base/versions/3.5.2/topics/file.choose

```# http://wiki.math.yorku.ca/index.php/VR4:_Chapter_1_summary
x <- seq( 1, 20, 0.5)   # sequence from 1 to 20 by 0.5
x                       # print it
w <- 1 + x/2            # a weight vector
y <- x + w*rnorm(x)     # generate y with standard deviations
#    equal to w

dum <- data.frame( x, y, w )     # make a data frame of 3 columns
dum                     # typing it's name shows the object
rm( x, y, w )           # remove the original variables

fm <- lm ( y ~ x, data = dum)   # fit a simple linear regression (between x and y)
summary( fm )                   # and look at the analysis

fm1 <- lm( y ~ x, data = dum, weight = 1/w^2 )  # a weighted regression
summary(fm1)

lrf <- loess( y ~ x, dum)    # a smooth regression curve using a
#   modern local regression function
attach( dum )                # make the columns in the data frame visible
#   as variables (Note: before this command, typing x and
#   y would yield errors)
plot( x, y)                  # a standard scatterplot to which we will
#   fit regression curves
lines( spline( x, fitted (lrf)), col = 2)  # add in the local regression using spline
#  interpolation between calculated
#  points
abline(0, 1, lty = 3, col = 3)    # add in the true regression line (lty=3: read line type=dotted;
#   col=3: read colour=green; type "?par" for details)
abline(fm, col = 4)               # add in the unweighted regression line
abline(fm1, lty = 4, col = 5)     # add in the weighted regression line
plot( fitted(fm), resid(fm),
xlab = "Fitted Values",
ylab = "Residuals")          # a standard regression diagnostic plot
# to check for heteroscedasticity.
# The data were generated from a
# heteroscedastic process. Can you see
# it from this plot?
qqnorm( resid(fm) )               # Normal scores to check for skewness,
#   kurtosis and outliers.  How would you
#   expect heteroscedasticity to show up?
search()                          # Have a look at the search path
detach()                          # Remove the data frame from the search path
search()                      ```