NEWS - What one should consume?

Everyday we end up reading something as NEWS and with mobile many topics are thrown over us in the name of notifications.

While reading NEWS today, I realized that how much the NEWS that we had a decade before than the NEWS that we see had changed.

"Doordharsan" was the only channel that there to watch when TELEVISION came to every home. Just like RADIO, Entertainment and NEWS were the only focus of the TELEVISION MEDIA. NEWS was not about ACTORS and Cinema. Nowadays, the first few topics in NEWS are mostly related to Cinema and Entertainment very rarely i could find the real NEWS.

Those old days, GOVEMENT elections, policies, regulations, disaster management were the topics that predominantly occupied NEWS. It was very much helpful for COMMON man to prepare oneself for any drought & famine. Even today the value of such information are very valuable. The velocity with which changes happening in and around us have changed, but there are few things which has not changed or going to be changing for another decade. For example Farming, Weather, Famine, Drought, Government elections, taxes, transport, petrol prize, gold prize are not going to change drastically. Industries, technologies, economics, trade and different kinds of entertainments are going to change slower or faster but they are going to affect us differently than prior said items.

It is better to stabilize oneself with NEWS which are slowly changing and have huge impact on daily life as given the example given above which are slowly changing or very important with not much drastic changes.

Entropy, Cross Entropy and KL Divergence

In the below video, following were my learning.

I learned entropy in information theory to check and encode bits to improve entropy and in decision tree formation based on test data (supervised learning).

Never combined entropy with probability distribution. To some extend I could remember Huffman encoding relation with probability and never with change in distribution.

Mostly I have considered probability distribution to be static, but in reality it is definitely it is dynamic except for rare cases.

Cross Entropy is an entropy with respect to actual vs predicted distribution and KL divergence is the change in entropy with predicted distribution from actual distribution.

Entropy = Cross Entropy + KL Divergence

https://www.youtube.com/watch?v=ErfnhcEV1O8

GATE Preparation Reference

Jenny Mam

https://www.youtube.com/watch?v=jj2HAvp51sU (one important point, everyone has their own strength and weakness, we should focus based on it)

why and opportunities?

https://www.youtube.com/watch?v=anssDGhoFwo

Ravindra Babu Ravula

Changes in GATE Syllabus? - https://www.youtube.com/watch?v=nJERS0RgKVk&feature=youtu.be

COMPILER DESIGN - Code Optimization & Liveness Analysis 

OS - system calls and i/o scheduling

COMPUTER N/W

NAT (private address) and Sockets added

IPv6, Firewall, wifi, n/w security removed.

(Old) MCQ - Multi Choice Question - make only one choice

1 question, -ve marking is there (0.33). 2 question, -ve marking is there (0.66).

(New) MSQ - Multi Select Question - No Negative Marking & No Partial mark - MORE than 1 Option. 

(Old) NAT - Numerical Answer Type - No Negative Mark 

Subject Wise:

1. Discrete Mathematics & Probability

2. TOC & Compiler Design

https://www.youtube.com/watch?v=eqCkkC9A0Q4&list=PLEbnTDJUr_IdM___FmDFBJBz0zCsOFxfK

3. Computer Architecture

4. Operating System

5. Computer Networks

6. Data Structure and Algorithms

7. Digital Logic

8. Aptitude

9. C Programming

https://questions.examside.com/

https://www.youtube.com/watch?v=-aQH0ybMd3U

https://www.youtube.com/watch?v=HxAhOETcvr4

https://www.tutorialspoint.com/digital_communication/

தெளிந்த நல்லறிவை தினம்தோறும் வேண்டுங்கள்

இவ்விரு பாடல்களும் தினம் பாடி வேண்ட இறையருள் பெருகும் 

எண்ணிய முடிதல் வேண்டும், நல்லவே எண்ணல் வேண்டும்;

திண்ணிய நெஞ்சம் வேண்டும், தெளிந்த நல்லறிவு வேண்டும்;

பண்ணிய பாவம் எல்லாம், பரிதி முன் பனியே போல,

நண்ணிய நின் முன் இங்கு நசித்திடல் வேண்டும் அன்னாய்!

ENNiya mudidhal vaeNdum, Nallave eNNal vaeNdum;

[Let all thoughts be accomplished, And only goodness be thought of;]

ThiNNiya Nenjam vaeNdum, TheLindha NallaRivu vaeNdum;

[Let my heart be strong, And my mind be clear;]

PaNNiya paavam ellaam, Paridhi mun paniye pola,

[Let all sins that I've committed, Become like snow before sun,]

NaNNiya Nin mun ingu Nasithidal vaeNdum Annaai!

[And before your righteous form, Get crushed Oh Mother!]

யாதும் ஊரே யாவரும் கேளிர்
தீதும் நன்றும் பிறர்தர வாரா
நோதலும் தணிதலும் அவற்றோ ரன்ன
சாதலும் புதுவது அன்றே, வாழ்தல்
இனிதென மகிழ்ந்தன்றும் இலமே முனிவின்
இன்னா தென்றலும் இலமே, மின்னொடு
வானம் தண்துளி தலைஇ யானாது
கல் பொருது மிரங்கு மல்லல் பேரியாற்று
நீர்வழிப் படூஉம் புணைபோல் ஆருயிர்
முறை வழிப் படூஉம் என்பது திறவோர்
காட்சியில் தெளிந்தனம் ஆகலின், மாட்சியின்
பெரியோரை வியத்தலும் இலமே,
சிறியோரை இகழ்தல் அதனினும் இலமே. (புறம்: 192)

To us all towns are our own, everyone our kin,
Life's good comes not from others' gifts, nor ill,
Pains and pain's relief are from within,
Death's no new thing, nor do our bosoms thrill
When joyous life seems like a luscious draught.
When grieved, we patient suffer; for, we deem
This much-praised life of ours a fragile raft
Borne down the waters of some mountain stream
That o'er huge boulders roaring seeks the plain
Tho' storms with lightning's flash from darkened skies.
Descend, the raft goes on as fates ordain.
Thus have we seen in visions of the wise!
We marvel not at the greatness of the great;
Still less despise we men of low estate.

Kaniyan Poongundran, Purananuru - 192

Learning Upside Down - Cone of Learning

My preferred style of learning so far has been reading. In order to save time I usually used to take middle ground reading a lot of Slides instead of spending lot of time on videos to get a bird eye view on new topic which i have been never exposed to. Cone of Learning brings in a new idea of what gets into my brain. It is vital and very important to know it. It is also important to prioritize the activities and give weight and time to activities which brings in more learning.

For sometime, at least with my master course, i have been giving preference to books as the course of action and knowledge flow in books will be far better than google random stuff. Now at the same time I see taking random topic and experiencing helps to program my brain better more than simple course flow. so, i let myself to look at the course flow for completeness without a serious study at the same time experience random topic time to time for a serious study.

Being aware that learning is expensive in terms of time when we want to make 90% with respect to cone of learning dictation. 

Thought Programming with sama, dhana, bedha, dhanda and timeout

Small Note: In the diagram i see, too much of information could bias us (overfitting, low variance), too less information will result in high variance (underfitting), bias and variance are weighted and stored with our experiences to act. Our Action is driven by biases, variance and weight & memory given the situation.

Priming is nothing but suggestion providing, it is a way to feed input to our brains hassle free beyond saying, grabbing attention and making yourself or others to listen. They help us to learn a lot quickly via biasing our thought process. Biases can be seen both positive and negative. Without bias there is no learning and without variance there is no new learning. Dynamic weighing of the biases & variance based on situation helps to remain sane and rational.

Timing out on few thoughts seems to be another very important tool our brain employs when it runs into long duration loops. If not the person goes berserk and sometimes mad. Time sense of our brain is very vague, it is completely relative to our thought world.

Sama, dhana, bedha, dhanda:

Our brain engine which keeps running in loop with or without timing sense is difficult to Control. sama, dhana, bedha and dhanda can be employed by us for yourself. The order of employment is very important here.

Sama - Making peace with what our brain says.

Dhana - Enticing the brain for the situation

Bedha - Logical Argument

Dhanda - Penalizing or Under weighing the thought, here is where we have to put a timeout and end the thought.

Sama, dhana, bedha & dhanda is a straight forward tool, sometimes we may feeling boring to apply it mostly depending on the prevailing situation. It is not mandatory to apply, I felt that beyond applying sama, dhana, bedha and dhanda over our children or others, it is applicable to yourself and to our brain being an incredible idea.

Beyond becoming victim to the situation. Creation of situation helps us to feel the mirage control. Sometimes we do so :) and while doing so, we don't compromise our biases, variance and weights. It is disheartening act but one has to fly to higher dimension to double check the logic and rationality.

Datastructure ADT

Reason for listing them down is for every methods in ADT as it difference from structure to structure and we have check runtime performance.

The way performance is calculated is based on inputs or sometime based on internal properties like height, depth, degree etc.,  But all the performance dependent on the Representation or implementation details (like graph / tree using vector, array, list etc.,)

Data Structures like Stack, Queues, Linked Lists, Tree, Graph are called as Abstract Data Types with few behaviors or operations or methods.

They are very special because, one has to know them before implementing any algorithm with them in advanced algorihtms where we don't make use of exact data structures provided out of box from programming languages but we will design them for our custom purposes with tweaks.

1. Stack  - push, pop (INSERT and REMOVE, no traversal)

REPRESENTATION - ARRAY (BEST), LINKED LIST

2. Queue - enqueue, dequeue  (INSERT and REMOVE, no traversal)

REPRESENTATION - ARRAY (BEST), LINKED LIST

3. Linked list - first, last, before, after (Traversal is like a person climbing a rope)

        Insert for Singly linked list alone is a special case, it take much time when compared to doubly linked list. (It is not a chain where we could de-link and link anywhere, every time one has to traverse from first)

4. vector - rank based. rank 

    sequence - mixed of vector and list.

REPRESENTATION - ARRAY(BEST), LINKED LIST

5. Tree - parent, children, root (like first), leaf (like last), internal, external. 

degree - count of children.

REPRESENTATION - VECTOR AND LIST

Traversal is with iterator - hasNext, next at each level.

Sub Structures

Spanning tree

Special Traversal

PreOrder, InOrder, PostOrder

Specialized Structures with special properties

Binary Tree - leftchild, rightchild

Heap - binary tree with Total order reflection

6. Graphs (vertices and edges) -  (there is no first, next, last, nor parent, children)

incident (incoming/outgoing) edges, adjacent(nearby) vertices, parellel edges.

REPRESENTATION - 

LIST, MATRIX.

EDGE LIST, ADJACENCY LIST and ADJACENCY MATRIX

Paths, Cycles - Simple and complex.

Sub Structures

Sub graph

Spanning Tree, Forest

Connect Graph

Special Traversals

BFS,

DFS

Shortest Path

Specialized Structures with special properties

UNDIRECTED graph

DIRECTED graph

Weighted Graph 

Non Weighted Graph

degree - count of incoming/outgoing edges.