Prasad Talasila prasadtalasila

(Based on explanation by Dr. Jacob Høxbroe Jeppesen)

Introduction

Machine learning is the field of research and applications that has branched off from artificial intelligence in 1980s. Now, machine learning researchers try to develop engineering techniques that can help applications learn from existing data to perform better.

Since, machine learning is performing approximation, we can call it a higher dimensional regression technique based on statistics.

Techniques

The following installation methods were tried out.

Working instances of Ceph OSD service

Vagrant-based solution

Container-based solution (partially working)

With some trial and error, I managed to get a stand alone Ceph Object Storage Device (OSD) service working. The instructions to replicate the installation are as follows.

	import matplotlib.pyplot as plt
	import numpy as np
	from paho.mqtt.client import Client as MQTTClient
	from paho.mqtt.client import CallbackAPIVersion
	from paho.mqtt.client import MQTTv311
	import queue
	import struct
	import time

	HOST = "hostname"

	---
	# Variables here are applicable to all host groups NOT roles

	# This sample file generated by generate_group_vars_sample.sh

	# Dummy variable to avoid error because ansible does not recognize the
	# file as a good configuration file when no variable in it.
	dummy:

	# You can override vars by using host or group vars

	Assume that a node project exists in /home/alan/autolabjs with the corresponding dependencies in
	/home/alan/autolabjs/node_modules. Let's also say we have mocha dependency installed in node_modules.

	Now, a regular way to use the installed dependencies is to execute mocha is as follows.
	$cd /home/alan/autolabjs/
	$./node_modules/.bin/mocha <rest of the command>

	Since the soft links to all the installed node modules are available in
	/home/alan/autolabjs/node_modules/.bin directory, we can add this directory to the PATH variable
	to utilise all the installed dependencies like regular shell utilities.

	create a random arrangement of an input sequence (good for exam seating arrangements)
	$seq <first_no> <last_no> \| shuf

	If we have pre-determined list of students, we can use "shuf" command to generate the random permutation of the students
	and then give them a sequential seating order.
	$shuf student_list.txt > random_order_list.txt


	Smalltalk has
	Class - defines behaviors of instances
	Object - an instance of some class
	Selector - Name of a message
	Message - selector with parameters
	Method - code / procedure found in class for responding to a message


	It is not essential to have classes, we can do OOP with just objects


	Interasting courses
	OS using Rust, Prof David Evans rust-class.org/


	Jude Nelson, student of Prof Larry Peterson
	http://www.cs.princeton.edu/~jcnelson/index.html amazing work

	From paper
	An Early History of Smalltalk, Alan Kay
	(Ref: http://www.vpri.org/pdf/hc_smalltalk_history.pdf)

	Essential principles
	1) Everything is an object
	2) Objects communicate by sending and receiving messages (in terms of objects)
	3) Objects have their own memory (in terms of objects)

	Principles that kept on changing between versions of smalltalk

	As the complexity increases, architecture dominates materials.

	each object should be thought of as having a virtual computer inside.

	In networked environment, each object must have a URL and an IP address.

	Object solutions need a universal interface language using which an object can interact with another object. HTTP-like REST protocols seem like good candidates for this kind of work. SOAP was actually written for this kind of work, but never took off.
	It seems Alan Kay's idea of universal interface language was much broader than HTTP/SOAP.

	programming in the large and programming in the small are very very different.