Week 3

This week we started off strong, learning about AI and had some really fun games in class (which I took part in :)). We began with a discussion on what exactly is intelligence in general where most of the class came into a similar consensus that essentially intelligence is something that appears to possess sentience, capable of making decisions on its own, solve problems and has the ability to learn.

Then when it came to Artificial Intelligence, whose dictionary definition states that it is "the capability of a machine to imitate intelligent human behavior", it has advanced to the pint where it is capable of solving near impossible math problems that could take humans years, searching for items in long and complicated lists, sorting items, based on a particular attribute and many more functions that humans would find a lot more difficult.

We've come to learn that the basic concept of AI took off in the 50's which was then propelled into machine learning in the 80's and finally how we are now living in the age of deep learning.

We then had an activity based on the Turing test, which was coined by Alan Turing in the 50's in which it tests a machine’s ability to exhibit ‘intelligent’ behavior by asking a question and getting two responses between a human and a computer, and if the evaluator is unable to

guess which one is which, the machine passes. Me and my friend Saim volunteered to partake in a live Turing test where I was given predetermined answers as the computer, and Saim was tasked with writing his own and it was up to our class to figure out who was who (or... what?). Our class ended up guessing wrong (by a landslide) which lead to the computer clearing the Turing test.

Up next we were introduced to the concept of Machine Learning which is an application of Artificial Intelligence (AI) that provides systems the ability to automatically learn and

improve from experience without being explicitly programmed and it's subset Deep Learning that has networks capable of learning based on complex learning algorithms especially useful for large data. We then started our first activity:

Activity #1

examples of machine learning and deep learning

Machine Learning:

• Game development, used for NPC's (non playable characters) in many games.

• Translations, speech recognition used to translate different languages seamlessly.

• Spam Email recognition.

Deep Learning:

• Vision for driverless vehicles, understands the realities of the road and how to respond to them.

• Image colorization, deep learning algorithms are able to use the context and objects in the images to colour them to basically recreate the black-and-white image in colour.

• Personalized entertainment, suggestions on streaming services using algorithms based on your interests.

Moving on to our next topic, State Space Search which is a tool that takes a problem and break it up into a problem space and depending on how the space looks, we can use particular search algorithms to find the optimal answer by moving through the space. Optimal can basically mean anything depending on the nature of the domain being

searched, for e.g. Google Maps – shortest path from A to B and tic-tac-toe – The best move we can make, looking ahead to see the possible effects, which lead to a discussion on 'The intelligent paper' and question of which it was truly intelligent or not. After playing a game of tic tac toe against the paper's henchman (Milan) it was pretty clear that the outcomes were foreseen in such a way that we humans could only hope to tie against the paper and there was no possible way in which we could beat it.

Activity #2

depth-first search

Depth-first search: The Depth First Search ( DFS ) is an algorithm for traversing or searching tree or graph data structures which uses the idea of backtracking.

Advantages

• DFS consumes very less memory space.

• It will reach at the goal node in a less time period.

Disadvantages

• It is possible that may states keep reoccurring. There is no guarantee of finding the goal node.

• Sometimes the states may also enter into infinite loops.

Circling back to Machine Learning, we learnt about it's main two algorithms, Supervised Learning aka Classification and Unsupervised Learning aka Clustering.

Classification is the predicting category of input data based on identified set of categories that are grouped together based on predetermined characteristics includes:

• Object classification

• Handwriting recognition

• Image segmentation

• Speech recognition

Clustering is the grouping of data. Includes:

• Search result grouping

• Grouping of shopping items

• Recommender Systems

We then came to everyone's favorite assignment, where we were given a point and tasked to figure out which position it belongs to, based on nearest neighbor classification, using the distance formula.

Activity #3

The new point, (5,2), belongs to O (4, 1)