IA Test

1. What is "artificial intelligence"?
Give examples to illustrate its specific application areas and

scenarios.
It is the area of science in charge of developing elements that have intelligent behavior, that is,
as if they were carried out by human beings. It covers sciences such as computer science,
psychology, mathematics, physics, among others.
Some examples include: household devices such as vacuum cleaners, creation of chatbots
through NLP (Natural Language Processing), video games that seek to make the "machine" act
coherently with reality, autonomous vehicles, a more progressive form of chatbots are virtual
assistants (Cortana, Siri, Alexa, etc.), other examples are online translators, services that use
Machine Learning identifying search and consumption habits.
2. What is "machine learning"? Please elaborate on the two machine learning methods and
outline their flow process and cognitive mechanism.
It is a method that allows us to extract patterns and relationships that exist in the data through
the use of algorithms executed in computers. These statistical elements allow us to predict
behaviors and improve decision-making.
Cognitive machine learning refers to the combination of machine learning and brain
cognitive mechanism, specifically, combining the achievements of machine learning
we have studied for many years with the mind model CAM
3. Explain the definition of "deep neural network" based on your own understanding and point
out its basic cognition feature.
It is an artificial neural network with several layers hidden between the input and output layers,
these networks can model complex non-linear relationships.
The purpose of it is to receive a set of inputs, perform progressively complex calculations and
give solution (output) to real world problems such as classification
4. Please analyze the advantages and disadvantages of "perceived intelligence".
Advantage:
We have automated systems that make our video more comfortable like GPS, Siri or Cortana.
They do not require breaks or refreshments, that is, they can work hours even weeks.
Saving important resources using digital assistants
Disadvantages:
Its implementation is expensive since it requires the use of equipment and complex algorithms.
They do not have emotions or values, which implies more an ethical question.
5. Please elaborate on the main research content and future development direction of "brain-like
intelligence".
Today's computing power has enabled the development of artificial intelligence and even more
so, the hope that this level of development will continue for years to envision better algorithms
and AI equipment.
Although there are always reservations when this science is linked to neuroscience, not many
articles related to this area
What is the reason? Of course, we can only speculate. However, on the one hand, neuroscience
too much (from the point of view of brain intelligence) focused on the details of neural
processing rather than on the principles of
large-scale processing. On the other hand, brain-like intelligence research must be careful to
emancipate itself, to a certain degree - from the reach of technology.
For 500 years the development of AI has allowed an important development, which perhaps in
the future will allow us to simulate behaviors that today seem quite complex to us, but which
will cognitively challenge us seems unlikely.
6. Please elaborate on Hierarchical Temporal Memory (HTM) The core principles and basic
processes of.
HTM is a machine learning technology that allows capturing the characteristic and algorithmic
properties of the neocortex.
The neocortex is essential in the thinking of mammals, it allows functions such as vision,
listening, touching, moving, language and allows planning functions, for this it uses a large
number of neural algorithms, in this case it is not the case, biology suggests that the neocortex
uses a common set of algorithms.
On the principle of differential workings between the human mind and computers, this concept
would allow computers to learn and make progressively more sophisticated models.
For Fuster, the keys of the memory networks would be the following:
• The memory code is relational. Memory networks, cognits, are associations between
groups of neurons more or less dispersed in the neocortex. The concept of a neural
column is not used.
• Perceptual memory networks are hierarchically organized in the posterior cortex. The
executive memory networks, in the frontal cortex.
• The prefrontal cortex, at the peak of the action-perception cycle, mediates temporal
events with its executive integration of working memory.
• This working memory is maintained by recurrent activity between the prefrontal cortex
and the associative areas of the posterior cortex.
HTM principles
Hierarchy: An HTM network consists of regions organized in a hierarchy. The region is the main
unit of memory and prediction in an HTM. Typically, each HTM region represents one level in
the hierarchy. As the hierarchy is promoted, there is always convergence, that is, multiple
elements in a child region converge on one element in a parent region
Regions: linked within a hierarchy comes from biology. The neocortex is a large layer of
neuronal tissue about 2mm thick. Biologists divide the neocortex into different areas or
"regions." Some regions receive input directly from the senses and other regions only do so
after it has passed through several other regions before. It is the connectivity between the
regions that defines the hierarchy.
Sparse Distributed Representations: It is because only a few neurons are activated when they
perform a specific function,
Time Role: Because biological functions differ in time, vision, hearing, among others, the HTM
inputs must vary in time for their training.
Please combine the artificial intelligence related theories and methods learned in the course to
design a software
Advanced artificial intelligence products combining hardware, try to draw a system structure
diagram, and explain its application
Use scenarios and values, basic functions of each part, and specific workflows.
Facial recognition drone for friendly and enemy troops
Train Data Val Data

Deep Learning
(DATA) (DATA)
Audio
Image
Video
Data
Design of a system for the recognition of the facial characteristics and

armament of friendly and enemy troops.
For this, the use of deep learning is proposed, by means of learning the
presence of enemy troops can be detected to carry out intelligence,
which will allow decision-making such as avoiding combat or carrying out
maneuvers that can avoid it.
This implementation could be used in the same way to make the best
decisions in the event that a soldier has been injured, informing about
the medicines that they must take when rescued.

IA Test

Uploaded by

Copyright:

Available Formats

IA Test

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IA Test

Uploaded by

Copyright:

Available Formats

1. What is "artificial intelligence"?

Give examples to illustrate its specific application areas and

4. Please analyze the advantages and disadvantages of "perceived intelligence".

Facial recognition drone for friendly and enemy troops

Train Data Val Data

Design of a system for the recognition of the facial characteristics and

You might also like