# ai: apropos of black box approach in machine learning algorithms

<< A black box is a machine learning program that does not explain how it reaches its conclusions, either because it is too complicated for a human to understand or because its inner workings are proprietary. In response to concerns that these types of models may include unjust inner workings—such as racism—another growing trend is to create additional models to "explain" these black boxes. >>

<<  Even when so-called explanation models are created, (..) decision-makers should be opting for interpretable models, which are completely transparent and easily understood by its users. >>

Ken Kingery. Stop gambling with black box and explainable models on high-stakes decisions. Duke University.  May 14, 2019.

https://m.techxplore.com/news/2019-05-gambling-black-high-stakes-decisions.html  

Cynthia Rudin. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nature Machine Intelligence. volume 1, pages 206–215 May 13, 2019

https://www.nature.com/articles/s42256-019-0048-x   

# behav: intriguing wolves, prosociality among wolves (compared to dogs)

<<  Prosociality is important for initiating cooperation. (..) In a prosocial choice task, wolves acted prosocially to in-group partners; providing significantly more food to a pack-member compared to a control where the partner had no access to the food. Dogs did not. Additionally, wolves did not show a prosocial response to non-pack members, in line with previous research that social relationships are important for prosociality.  >>

Rachel Dale, Sylvain Palma-Jacinto, et al. Wolves, but not dogs, are prosocial in a touch screen task. PLoS ONE 14(5): e0215444. doi: 10.1371/journal.pone.0215444  May 1, 2019.

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0215444

Wolves more prosocial than pack dogs in touchscreen experiment. Findings support idea that dogs helping pack members is ancestral tendency, and not due to domestication. May 1, 2019.

https://www.eurekalert.org/pub_releases/2019-05/p-wmp042419.php

Also

'wolves'

https://flashontrack.blogspot.com/search?q=wolves

# soc: computational socioeconomics, a brief manifesto.

AA << will make a brief manifesto about a new interdisciplinary research field named Computational Socioeconomics, followed by detailed introduction about data resources, computational tools, data-driven methods, theoretical models and novel applications at multiple resolutions, including the quantification of global economic inequality and complexity, the map of regional industrial structure and urban perception, the estimation of individual socioeconomic status and demographic, and the real-time monitoring of emergent events. >>

Jian Gao, Yi-Cheng Zhang, Tao Zhou.
Computational Socioeconomics. arXiv:1905.06166v1 [physics.soc-ph].  May 15, 2019.

https://arxiv.org/abs/1905.06166

download (free, ~11Mb):    https://arxiv.org/pdf/1905.06166

# tech: apropos of peripheral neural interfaces; the AlterEgo 'second self'

<< AlterEgo is a  non-invasive, wearable, peripheral neural interface that allows humans to converse in natural language with machines, artificial intelligence assistants, services, and other people without any voice-without opening their mouth, and without externally observable movements-simply by articulating words internally. >>

<< AlterEgo seeks to combine humans and computers—such that computing, the Internet, and AI would weave into human personality as an internal "second self" and augment human cognition and abilities. >>

Jimmy Day. Project AlterEgo. MIT Media Lab.

https://www.media.mit.edu/projects/alterego/overview/

Nancy Cohen. AlterEgo opens silent spring of computer connections via wearable. Phys.org. May 20, 2019.

https://m.techxplore.com/news/2019-05-alterego-silent-wearable.html

# tech: coloured walls and cities by nano pixels (also for immediate hack poetry)

<< Plasmonic metasurfaces are a promising route for flat panel display applications due to their full color gamut and high spatial resolution. >>

AA << present scalable electrically driven color-changing metasurfaces constructed using a bottom-up solution process that controls the crucial plasmonic gaps and fills them with an active medium. >>

<< ... which are a hundredfold thinner than current displays. >>

Jialong Peng, Hyeon-Ho Jeong, et al. Scalable electrochromic nanopixels using plasmonics. Science Advances   Vol. 5, no. 5, eaaw2205  doi: 10.1126/sciadv.aaw2205. May 10, 2019.

https://advances.sciencemag.org/content/5/5/eaaw2205

Smallest pixels ever created could light up color-changing buildings. University of Cambridge. May 10, 2019.

https://m.phys.org/news/2019-05-smallest-pixels-color-changing.html

FonT

2107 - coll' anse d' apostrofo. Jan 14, 2007

<< traccia coll' anse d' apostrofo un virtuosistico //
irriproducibile tagete // >>

https://inkpi.blogspot.com/2007/01/2107-coll-anse-d-apostrofo.html

# psych: exploring trust among politicians

<< The new research, (..) tested a vulnerability-centered definition of trust-meaning, defining trust as a willingness to be vulnerable to the actions of another. The results revealed three assessments that lead to one trusting in the government: whether it has the ability to do its job, the benevolence to care about its people and the integrity to generally do the right thing. >>

Do you trust politicians? Depends on how you define trust. Michigan State University. May 15, 2019

https://m.phys.org/news/2019-05-politicians.html

Joseph A. Hamm, Corwin Smidt, Roger C. Mayer. Understanding the psychological nature and mechanisms of political trust. PLoS ONE 14(5): e0215835. doi:10.1371/journal.pone.0215835.  May 15, 2019.

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0215835

# sec: adversarial audio attacks; small sound perturbations to hack a Machine Learning model and remedies

<< Adversarial audio attacks can be considered as a small perturbation unperceptive to human ears that is intentionally added to the audio signal and causes a machine learning model to make mistakes. >>

Mohammad Esmaeilpour, Patrick Cardinal, Alessandro Lameiras Koerich. A Robust Approach for Securing Audio Classification Against Adversarial Attacks. arXiv:1904.10990 [cs.LG] Apr 24, 2019.

https://arxiv.org/abs/1904.10990

Ingrid Fadelli. An approach for securing audio classification against adversarial attacks. May 7, 2019.

https://m.techxplore.com/news/2019-05-approach-audio-classification-adversarial.html