Brenda's Blog

Student Journal

Brenda Spears – 2201841702

In the third week of the Intelligent Systems course, we discussed on informed search (BFS, A* search, and heuristic) and local search. Informed search uses domain knowledge for it to be able to lead selection of the best path to continue its search.

Heuristic is a technique used to quickly solve a problem using the rule of thumb, intuitive judgement, or common sense. For BFS, sort nodes in the nodes list by increasing values of an evaluation function f(n) that incorporates domain-specific information.

The difference between A* algorithm and greedy BFS is A* can be morphed into another path-finding algorithm by simply playing with the heuristics it uses and how it evaluates each node, but BFS is very useful when you want to find the shortest and most optimal path by traversing as few edges as possible. Both A* and greedy BFS is complete, which means that it will always find a solution if it exists.

For our group project, my team and I had not progress this week as we were still deciding on the ideas and still researching on the algorithms.

Read more

Student Journal

Brenda Spears – 2201841702

In the second week of the Intelligent Systems course, we were familiarized with uninformed search (DFS, BFS, etc.) .

In this lecture, we learned about the problem-solving agent, the goal or problem the agent tried to achieve or solve, the actions needed to be taken, and the knowledge the agent needs. The agent needs sufficient and adequate informations in order to reach the goal and able to deliver the descriptions of the situation.

Then, we learned the uniform search strategies. BFS using queue, DFS using stack, Depth-limited search (DFS with limited depth), Uniform-cost search using priority queue to order nodes, sorted by their path costs, IDS which requires modification to the tree search algorithm.

BFS properties:

– A finite branching factor makes it complete.

– All edges having the same cost makes it optimal.

– Take a lot of time and memory to find solutions with large number of steps.

DFS properties:

– May not terminate without a loop detection or depth bound.

– It will not be complete without a cycle detection.

– Does a chronological back tracking.

UCS properties:

– Branching factor is finite makes it complete.

For our group project, my team and I had come up with a decision to create a program that could detect facial emotions using tensorflow. We might integrate it, but still in the process of researching and finding other ideas.

Read more