Are algorithms evil?

Google AI + US military . Is it good, is it bad or is it evil

https://www.theguardian.com/technology/2018/mar/07/google-ai-us-department-of-defense-military-drone-project-maven-tensorflow

DoD’s Project Maven uses tech firm’s TensorFlow artificial intelligence systems, prompting debate both inside and outside company

Project Maven forms part of the $7.4bn spent on AI and data processing by the DoD, and has seen the Pentagon partner with various academics and experts in the field of AI and data processing. It has reportedly already been put into use against Islamic State.

Google spokesperson said: “This specific project is a pilot with the Department of Defense, to provide open source TensorFlow APIs that can assist in object recognition on unclassified data. The technology flags images for human review, and is for non-offensive uses only.”

  1. Terrorists deserve to die
  2. Civilians should be spared from attacks
  3. Better image classification in drones can lead to better decisions on killing terrorists versus collateral deaths.
  4. Civilian deaths further the cause of terrorism
  5. Terrorists often embed them with civilians as shields to deter airstrikes
  6. What is the definition of evil?

WHO IS EVIL

  • A civilian turns to a terrorist because his family was killed in a bad drone strike?
  • A government diverts terrorism funds from terror funds to corruption?
  • A government that keeps killing smaller number of terrorists based on events a decade ago to keep the military industrial complex in good shape for the next war?
  • The computer scientist who created a classification algorithm to detect cancer tumors in MRI?
  • The company that employs him and owns the algorithm , hands it to the government for killing people or any usage in return for money as well as help in anti -trust matters.

Movie Review: Sonu Ke Titu Ki Sweety

We really enjoyed laughing at this movie despite the R rated jokes and bleeps. Its a bro movie about two guys who are friends since childhood and the girl who gets engaged to one and almost kills the friendship. With a lot of situational comic gimmicks  and fine cast it is easily one of best bollywood comedies in some time.

Some things which almost derail the movie are a misogynistic plot with bros before #hos as the theme.

Amazing Binder turns github repositories to a collection of interactive notebooks

 

Just go to https://mybinder.org/

 

 

Turn a GitHub repo into a collection of interactive notebooks

Have a repository full of Jupyter notebooks? With Binder, open those notebooks in an executable environment, making your code immediately reproducible by anyone, anywhere.

I turned my repo from  to https://github.com/decisionstats/pythonfordatascience

to

https://hub.mybinder.org/user/decisionstats-p-nfordatascience-vhxwj6n2/tree

How it works

1
Enter your repository information
Provide in the above form a URL or a GitHub repository that contains Jupyter notebooks, as well as a branch, tag, or commit hash. Launch will build your Binder repository.
2
It builds a Docker image of your repository
Binder will search for a dependency file, such as requirements.txt or environment.yml, in the repository’s root directory (more details on more complex dependencies in documentation). The dependency files will be used to build a Docker image. If an image has already been built for the given repository, it will not be rebuilt. If a new commit has been made, the image will automatically be rebuilt.
3
Interact with your notebooks in a live environment!
JupyterHub server will host your repository’s contents. We offer you a reusable link and badge to your live repository that you can easily share with others

https://jupyterhub.readthedocs.io/en/latest/

JupyterHub

JupyterHub, a multi-user Hub, spawns, manages, and proxies multiple instances of the single-user Jupyter notebookserver. JupyterHub can be used to serve notebooks to a class of students, a corporate data science group, or a scientific research group.

jupyterlab is here!

Get your copy at http://jupyterlab.readthedocs.io/en/stable/index.html

From command line do the following

  • Upgrade Jupyter notebook

pip install notebook —upgrade 

  • Install  Jupyterlab

pip install jupyterlab

  • Launch Jupyter Lab

jupyter lab


A very nice design and excellent for working- lets see how fast people adapt to it along with other kernels in Julia, R, SAS et al