What did I do :

• We did an exercise on the train of single layer preceptor, which is not as hard as I expected.
• Secondly, we discuss the deep learning on sample node layers.
• Lastly, we did a group discussion on specific topic inside this session, I remember I was the first group.

Obstacles/Difficulties :
The exercise today is not that confusing, the only difficulties that I encounter today are after class final project selection, we had one week left and my group was not ready for a specific topic to do for final project.

What did I learn:
Today we learn more biological theory of Intelligent System, the structure of neurons, preceptors and a bit of deep learning theory.

What did I do :

• Today I watch youtube regarding the topic of the class, which is the neural network.
• Second, we discussed entropy, which is the uncertainty or randomness with a particular variable.
• Third, we discussed information gain, decision trees and so on, then do the exercise of one sample table.

Obstacles/Difficulties :

The difficulties of today to find out the entropy and information gain of one’s graph, there were many scattered data, and I was still confused about the meaning of the formula of calculation entropy amount.

What did I learn:
Today I learn about more about supervised learning, consists of neural network, entropy, information gain and decision tree construction.

What did I do :

• Today we did an online based quiz in a website called quiziz, its the first quiz so the score will be taken to the report.
• Discussing topic for today which is rule mining.
• Find our confidence for data based on the association of rule mining, mostly with the theory of probability.
  

Obstacles/Difficulties :
The difficulties of today are to catch up the exercise since we had the first quiz, the quiz was also difficult because each question has a time limit, and all of the questions were multiple choice. so I need to think fast for answering the question.

What did I learn:
Today we learn about Associate rule mining, in this topic we learn to find confidence based on the probability of data that appears inside a set. it is one of the different techniques approaches to Artificial Intelligence.

What did I do :

• Today we did the exercise for the current topic with friends for attendance.
• Discussing Cluster techniques with Simpson’s family as an example from the slide.
• Discussing all kinds of calculation with some example of simple data sets （Manhattan).

Obstacles/Difficulties :
The difficulty of today is doing the exercise because it takes a lot of time if the answer did not match what we expected. since the task needs several repetitions of calculation if we did not reach the answer.

What did I learn:
Today we learn about a topic called “learning by observation”. This topic consists of types of matching learning, clustering technique, manhattan, hierarchal and Euclidean calculation on data, K-means algorithm

What did I do :
1. Today we did 2 assignments about Naive Bayes Classifier together
2. Reviewing about week 1 modules regarding to the definition of A.I and Machine Learning.
3. Discussing with friends regarding to final projects.

Obstacles/Difficulties :
1. The hardest part of today is doing the tasks actually since the time limit was pretty fast and it was for the class attendance, so we needed to do it together in cisco and discuss it too with others.
2. Since we were learning Unsupervised Learning, Probabilities (conditional) needed to be review since I almost forget some of the Probability knowledge.

What did I learn:
1. Unsupervised Learning, in that topic we learn about conditional probability and naive baiyes to find possibles in order to predict the outcome results, also we compare both Unsupervised and Supervised learning before we start the main topic.

In this week we learn about adversarial search, there are many topics today such as :

– Minimax algorithm where there it takes min/max value based on the level of the tree.
-Small Introduction of AI Games such as backgammon and chess.
-Alpha-Beta pruning which aims to decrease the number of nodes when a search occurs towards the tree.
– Game playing as Search which has an initial state, the legal movement and the result.
-Non-deterministic Games
etc.

The most interesting part of the day is a sudden light off in the FX mall, we were all ready to go home but unfortunately, the light went on again after 3 minutes, I was super disappointed but at least we can listen to the class since IS is very hard.

The most challenging part of the day are the tasks because to analyze the whole topics again its very hard for me, I need to work with my friends in order to understand the whole things.

What do we learn in class :

In this week we learned about two main topics, first is informed search and second is Local Search.

Informed Search:

First, we talked about each data contains heuristic h(n), with that, we discuss the main section which are Best first search, Greedy search, and A* search.

To sum up briefly:

Best first search tries to expand most desirable
unexpanded nodes.

Greedy search is the search that tries to prioritize the smallest heuristics among all the paths to find the destination of a node, it is very fast but not optimize

A* search is a search that tried to find all the possibility of paths and choose the smallest edge to continue to find another possibility of paths, it is not as fast as Greedy search but its more optimize.

Local Search :

Hill-climbing, Complete Search, TSP and Genetic Algorithm.

Hill-climbing is basically tried to find a good solution for one algorithm by seeing the complexity analysis.

The complete search indicated that every node is a solution, the operators go from one solution to another, it can stop any time and have a valid solution and same as hill-climbing search is now about finding a better solution.

TSP is basically to find out the shortest path from starting at point n to the destination ( the starting point) by travelling each of nodes at least ones.

And Genetic algorithm is kinda related to Biology, it’s a metaheuristic that related to natural selection. I think the Genetic algorithm was the hardest for me since the lecture said it was related to Biology and I am not good at Biology.

My obstacles and challenges:

At that day I did not feel very well so I could not concentrate on the class really well so it is very hard for me to understand all the types of search at once especially local searches. The second exercise is also very hard since you need to go over the slides, again and again, to look for solutions, but for now, I only see what my friends did just in case I could get some ideas from them.

What do we learn in class :

Today we were learning about Uninformed Search, Problem-solving agents and reviewing topics about state spaces since this course has a very strong connection with Analysis of Algorithm, therefore we tried to get over those once again, unfortunately, most of us forget about the topics so our lecture needs to repeat few things before we continue.

In public solving agents, there are 6 steps.

• States 
• Starting  
• Actions 
• Transition Model 
• Goal test 
• Path cost

From the wording itself, it is clear that each state represents the steps sequentially before any algorithm started to run.

Before searching, we need to know the nodes first and each node consists of these following :

• State 
• Parent node 
• Action 
• Path cost 
• depth 

Now the different types of uninformed searching are BFS, DFS, UCS, DLS and IDS, different types of search have different time complexity and different ways to implement depends on different situations.

Lastly, we were given a small task for us to finish whenever we want since it will not be marked, but we manage to finish it before class, therefore, we could discuss the results afterwards.

My obstacles and challenges:

It is very difficult for me to understand the whole search strategies with their complexity in one day. So it might need more time in order to understand the whole topics on day 2.

Today is the first day of the class, seeing the materials of the whole course made me nervous, today I knew a lot of new knowledge since the slide itself contains more than 60 pages.

What I learn :
AI, Artificial intelligence, has 4 parts, thinking both humanly, rationally, and acting both humanly, rationally. there is 8 interdisciplinary nature of AI. which are Psychology, Biology, Computer Science & Engineering, Philosophy, Biology, Cognitive etc. We also talked about the brief history of AI and Machine learning which started in 1956 till now. as well as AI’s application domains.

These are all still the introduction of the whole course, I am looking forward to knowing more pieces of information regarding to this course next couple of weeks.