Sunteți pe pagina 1din 5

Artificial Intelligence Midterm Exam

Sibgha Ishaq
17591556_006 UOG Lhr Campus

Q#1:
Heuristic is a good guide search. Heuristic is a method that might not always find the best
solution but It guaranteed to find a good solution in reasonable time. But sacrificing the
completeness and it increases efficiency
Heuristic function:
• Estimate cost of cheapest path from node (n) to a goal node.
• Heuristic function is a way to inform the search about the direction of a goal.
• It provide a informed way to guess which neighbor of a node will lead to goal
Q#2:
Convolutional Neural Networks (CNNs) is the most popular neural network model being used
for image classification problem. The big idea behind CNNs is that a local understanding of an
image is good enough. The big idea behind CNNs is that a local understanding of an image is
good enough. The practical benefit is that having fewer parameters greatly improves the time it
takes to learn as well as reduces the amount of data required to train the model.
Q#3:
Online search agents operate by interleaving computation and action. First it takes an action, and
then it observes the environment and computes the next action. Online search is a good idea in
dynamic or semi dynamic domains and stochastic domains. Online search is a necessary idea for
an exploration problem, where the states and actions are unknown to the agent.
Q#4:
• Artificial Intelligence is the simulation of the human process by machines (computer
systems). These processes include the learning, reasoning, and self-correction. We need
Artificial Intelligence (AI) because the work that we need to do is increasing day-to-day.
So it's a good idea to automate the routine work.
• Artificial intelligence (AI) technology and machine learning are increasingly being used
to analyze human behavior so that apps can predict what users want and when they want
it. These apps make activities like ordering groceries, watching movies, and listening to
music easier.
• Artificial Intelligence can provide humans a great relief from doing various repetitive
tasks. The technology can learn the work once and repeat it as many as desired by its
human programmer. Automation of different tasks reduces the workload from dull and
repetitive tasks.
Q#5:
Breadth-first search is just keeps expanding all the leaves and checks if one of them is the goal.
Breadth first search are uninformed searches where as Best first search is informed search. Best-
first search is depends on the heuristic function and the structure of the problem. If the heuristic
function is not good enough it can mislead the algorithm to expand nodes that look promising,
but are far from the goal.
Breadth first search:
• informed search
• Fifo ( queue)
• Shollowest node
• Complete
• Optimal
• Time complexity
O(v+E)
V = no of node
E = no of edges
• In AI complexity o(bd)
B = branch factor
d=no of nodes
Depth first search:
• Blind search
• Broad force search
• Uninformed technique
• Stock (lifo)
• Incomplete
• Deepest node
• Non optimal
• Time complexity o(v+E)
• In AI Complexity o(bd)
Q#6:
If an algorithm is complete, it means that if at least one solution exists then the algorithm is
guaranteed find a solution in a finite amount of time and If a search algorithm is optimal, then
when it finds a solution it finds the best solution.
A complete algorithm is one that guarantees that if a path to the goal exists, the algorithm will
reach the goal. Note that completeness does not imply optimality of the found path
For example, breadth-first search (BFS) is complete (and in fact optimal if step costs are
identical at a given level), while depth-first search (DFS) on trees is incomplete (consider infinite
or repeated states)
Q#7:
Followings are the future of neural networks:
Integration:
The weaknesses of neural nets could easily be compensated if we could integrate them with a
complementary technology, like symbolic functions. The hard part would be finding a way to
have these systems work together to produce a common result and engineers are already working
on it.
Sheer complexity:
Everything has the potential to be scaled up in terms of power and complexity. With
technological advancements, we can make CPUs and GPUs cheaper and/or faster, enabling the
production of bigger, more efficient algorithms. We can also design neural nets capable of
processing more data, or processing data faster, so it may learn to recognize patterns with just
1,000 examples, instead of 10,000. Unfortunately, there may be an upper limit to how advanced
we can get in these areas but we haven’t reached that limit yet, so we’ll likely strive for it in the
near future.
New applications:
Rather than advancing vertically, in terms of faster processing power and more sheer complexity,
neural nets could (and likely will) also expand horizontally, being applied to more diverse
applications. Hundreds of industries could feasibly use neural nets to operate more efficiently,
target new audiences, develop new products, or improve consumer safety yet it’s criminally
underutilized. Wider acceptance, wider availability, and more creativity from engineers and
marketers have the potential to apply neural nets to more applications.
Q#8:
In standard backprop, too low a learning rate makes the network learn very slowly. Too high a
learning rate makes the weights and objective function diverge, so there is no learning at all. If
the objective function is quadratic, as in linear models, good learning rates can be computed
from the Hessian matrix (Bertsekas and Tsitsiklis, 1996). If the objective function has many local
and global optima, as in typical feedforward NNs with hidden units, the optimal learning rate
often changes dramatically during the training process, since the Hessian also changes
dramatically. Trying to train a NN using a constant learning rate is usually a tedious process
requiring much trial and error. the learning rate is a configurable hyperparameter used in the
training of neural networks that has a small positive value, often in the range between 0.0 and
1.0. During training, the backpropagation of error estimates the amount of error for which the
weights of a node in the network are responsible.
Q#9:
It seems that you will only start with game AI programming once you know what the job
opportunities are . I am going to ignore the jobs part for now and let you know about some very
basic stuff as to how you can start.

AI in games like AoE , Dota or even FIFA is very very advanced . Tons of research and decades
of work goes into tweaking the AI in these games.

The best way to see how AI and games work in general is to look at old games. Games like
PAC-MAN show us a perfect example of how "AI" can be implemented , without actually
implementing it.
Q#10:
Systems that thinks like human and systems that act like human:
In artificial intelligence, an expert system is a computer system that emulates the decision-
making ability of a human expert. Expert systems are designed to solve complex problems by
reasoning through bodies of knowledge, represented mainly as if, then rules rather than through
conventional procedural code.
Systems that thinks like human and systems that thinks rationally:
Thinking humanly cognitive modeling. Systems should solve problems the same way humans
do. Thinking rationally the use of logic. Need to worry about modeling uncertainty and dealing
with complexity.
Q#11:
• Evaluate the initial state
• Loop until a solution is formed or there are no operation left
• Select and apply new operation
• Evaluate new state if good then current state then it is new current state
• Contain local area knowledge it does not have global area knowledge
• Best move its work its work done otherwise it will stopped
• It done not meet up nest move then it have does not back track
Problems in hill climbing search algorithm:
• Local maxima
• Ridges and alleys
• Plateau
Q#12:
Knowledge acquisition is the process of extracting, structuring and organizing knowledge from
one source, usually human experts, so it can be used in software such as an ES. This is often the
major obstacle in building an ES. There are three main topic areas central to knowledge
acquisition that require consideration in all ES projects. First, the domain must be evaluated to
determine if the type of knowledge in the domain is suitable for an ES. Second, the source of
expertise must be identified and evaluated to ensure that the specific level of knowledge required
by the project is provided. Third, if the major source of expertise is a person, the specific
knowledge acquisition techniques and participants need to be identified.
Knowledge Acquisition Technique:
At the heart of the process is the interview. The heuristic model of the domain is usually
extracted through a series of intense, systematic interviews, usually extending over a period of
many months. Note that this assumes the expert and the knowledge engineer are not the same
person. It is generally best that the expert and the knowledge engineer not be the same person
since the deeper the experts' knowledge, the less able they are in describing their logic.
Furthermore, in their efforts to describe their procedures, experts tend to rationalize their
knowledge and this can be misleading.
General suggestions about the knowledge acquisition process are summarized in rough
chronological order below:

• Observe the person solving real problems.


• Through discussions, identify the kinds of data, knowledge and procedures required to
solve different types of problems.
• Build scenarios with the expert that can be associated with different problem types.
• Have the expert solve a series of problems verbally and ask the rationale behind each
step.
• Develop rules based on the interviews and solve the problems with them.
• Have the expert review the rules and the general problem solving procedure.
• Compare the responses of outside experts to a set of scenarios obtained from the project's
expert and the ES.

S-ar putea să vă placă și