Professional Documents
Culture Documents
Figure 1.................................................................................................................................................3
Figure 2.................................................................................................................................................4
Figure 3.................................................................................................................................................6
Figure 4...............................................................................................................................................10
Figure 5...............................................................................................................................................11
Figure 6...............................................................................................................................................12
Figure 8: Poster..................................................................................................................................23
Activity 1
Introduction
Intelligent systems and Artificial intelligence can be very important aspects of the
application in several industries especially those involving automation of various services which
they offer. Intelligent systems can be defined as machines that are technologically advanced such
that they can perceive and respond to the environment in which they exist. Artificial intelligence
refers to the ability of machines to demonstrate intelligence that is naturally displayed by humans.
Artificial intelligence can be in the form of speech recognition and problem-solving while
intelligent systems can be in the form of neural networks or fuzzy systems and expert systems. The
systems that were applied during the session include expert systems and smart driving cars
respectively (Wahl, et. al, 2018). These systems have demonstrated solutions to various problems
Figure 1
Background
Computers can store a lot of information and they can process the information fast proofing
to be more accessible whenever they are required. The introduction of expert systems that
mimicked the process of decision making by human expertise acted as a breakthrough for the
introduction of artificial intelligence. The program was able to inquire about various situations and
how they can be able to respond to them. The expert systems can be used to advise non-experts
while using a program leading to the systems becoming widely used in industries. Logic
programming has been essential in improving artificial intelligence (Jarvie, 2012). The computing
power of a computer surpasses all the needs as time passes thus the capabilities of artificial
intelligence. The existence of big data in the current age can be cumbersome to be processed by a
person hence artificial intelligent can in handy in banking, entertainment, and marketing. Massive
artificial intelligence has enabled brute force learning through the use of big data.
in decision making thus solving complex problems through the representation of knowledge by the
use of mainly if-the rules instead of the use of conventional procedural code. The inference engine
and knowledge base which exist in artificial intelligence are very important in the representation of
Figure 2
Intelligent Systems can be used in combination with a thermal-based camera system which
contains a black body device used for calibrating the sensors of the image against the background.
This system was being used during group work. It applied technology in measurement of the body
temperature of the malaria patients through the use of a two-piece thermal scanner. The image
screen can be used to analyze up to five thousand people who come through the image screen. The
system could understand the geometry of the face and which is a representation of the body party
that is flat. The system had the capability of sending an email once the person who is being
measured has a high temperature. Hospitals were the potential clients who can put the system into
The future of artificial intelligence has been noted to be underway in different fields and
therefore expected to be better in the future. Companies can answer emails of their customers
through the use of machines that have been programmed to answer specific questions. The
machines can also receive calls and provide solutions for various inquiries from the customers. The
driverless cars ave has shown the ability of the machines to surpass human cognitive abilities in all
Figure 3
Artificial intelligence involves deep learning and machine learning. Artificial neural
networks can be created through the use of deep learning. This is an entire depiction of the human
brain through the neurons which are connected to form a network. AI uses basic algorithms in the
performance of its functions which can be regarded as traditional software but being applied in a
different manner (Mao, et. al, 2016). The intelligence can be noticed in the different warehouse
management systems in which shortage can be identified and its cause through the use of artificial
intelligence. Specific processes have been made possible through the replacement of decision
Insights
Artificial intelligence and intelligent systems have improved productivity, efficiency, and
accuracy in the organizations which are being applied. The incredible capabilities which artificial
intelligence has shown have enabled organizations to use the advantages that it comes with thus
enhancing human intelligence leading to gaining better value from the data that exists (Perez-Neira,
A.I. and Campalans, 2010). The benefits of intelligent systems have enhanced decision making
Competitive advantage and reduced cost are being gained by the companies which have
employed AI systems. The change of the way through which operations are run can be an issue to
The future of Artificial intelligence is positive as it is likely to create more jobs that they will be
lost. AI is likely to impact the economy as the creation of jobs will require certain specific skills to
be implemented in the new systems. Basic problem tasks are likely to be replaced as AI systems can
There is a need for businesses to take advantage of AI by achieving its effectiveness through
successful implementation which will enable everyone to benefit. The power of AI can be gained
through addressing the issues which exist at education level and business level. The focus is
required to be on training students on roles that are directly associated with working with artificial
intelligence. There is a need for emphasis on the subjects that act as a stem (Amershi, et. al, 2019).
Conclusion
Artificial intelligence and intelligent systems are very essential in the daily lives of humans.
This is through their application in most of the sectors leading to the achievement of efficiency.
Expert systems can compete with the specialist level in narrow ways but they may not be able to
function according to the daily situations. Artificial intelligence trends consist of many different
References
Amershi, S., Weld, D., Vorvoreanu, M., Fourney, A., Nushi, B., Collisson, P., Suh, J., Iqbal, S.,
Bennett, P.N., Inkpen, K. and Teevan, J., 2019, May. Guidelines for human-ai interaction. In
Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (pp. 1-13).
Zhang, J.S., and Chen, A.X., 2012. Review of quantum discord in bipartite and multipartite
Perez-Neira, A.I., and Campaigns, M.R., 2010. Cross-Layer Resource Allocation in Wireless
Communications: Techniques and Models from PHY and MAC Layer Interaction. Academic Press.
Mao, H., Alizadeh, M., Menache, I., and Kandula, S., 2016, November. Resource management with
deep reinforcement learning. In Proceedings of the 15th ACM Workshop on Hot Topics in Networks
(pp. 50-56).
Cicalo, S., Tralli, V. and Perez-Neira, A.I., 2011, May. Centralized vs distributed resource allocation
in multi-cell OFDMA systems. In 2011 IEEE 73rd Vehicular Technology Conference (VTC Spring)
Li, M. and Mourikis, A.I., 2012, October. Vision-aided inertial navigation for resource-constrained
systems. In 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems (pp. 1057-
1063). IEEE.
Bresina, J., Dearden, R., Meuleau, N., Ramkrishnan, S., Smith, D. and Washington, R., 2012.
Planning under continuous time and resource uncertainty: A challenge for AI. arXiv preprint
arXiv:1301.0559.
Jarvie, D.M., 2012. Shale resource systems for oil and gas: Part 2—Shale-oil resource systems.
Wahl, B., Cossy-Gantner, A., Germann, S. and Schwalbe, N.R., 2018. Artificial intelligence (AI)
and global health: how can AI contribute to health in resource-poor settings?. BMJ global health,
3(4), p.e000798.
Skobelev, P., 2018, June. Towards autonomous AI systems for resource management: applications
Activity 2
Introduction
Deep learning refers to the part of machine learning which uses multiple layers to
progressively extract higher-level features from the raw input. It makes the use of neural networks
to learn a representation of features directly from the data that has been presented. A trained neural
network can be used in the identification and removal of various artifacts such as noise. Therefore
deep learning acts as a datastore for image preprocessing and it is used for image augmentation. The
computer has to be trained to recognize various images about their real appearance. The more the
data that is fed to the computer, the better it becomes of recognizing the images. The images were
labeled for the computer to recognize the patterns which were present in the image. Therefore, the
computer will have the capability to build its own cognition. This is possible through the use of
python and TensorFlow in writing the program (Van Ginneken and Sánchez, 2017).
Background
Image processing using a dataset of the image involves classifying the images. An image
classifier is used through tf.keras.Sequential model and being able to load the data through the use
use of a class that will be able to work on disk. Overfitting should be identified and prevented.
Image processing can be done by either using greyscale or RGB values (Janowczyk and
Madabhushi, 2016).
Figure 4
A dataset that consists of several images is used in the process and the different classes of
the images are labeled to make the dataset ideal. The existence of massive labeled data sets and
GPU computing has led to the improvement of the concept of deep learning as it acts as a fuel to the
improvement which has been noticed. The availability of improved architecture such as inception
modules and highway networks has enabled the depth and flexibility of the network. The software
platforms such as chainer and TensorFlow have provided automatic differentiation leading to faster
prototyping and fewer errors being experienced. Different types of data should be used while
Description
Deep learning is very essential in daily lives as it has many applications such as image
classification. Image classifications can be seen in apps such as Facebook which provides an auto-
tagging suggestion. This is through the use of deep nets and the ability to take pictures at different
angles then labeling the name which belongs to the picture. It is now possible to be able to
recognize different objects in a certain picture and even relate it to a given occasion. For instance, if
a picture has been taken with chairs, a place where furniture is sold can be recommended for the
Figure 5 Speech
recognition. Speech is a form of communication in the society of humans. A model can behave the
same way as humans by recognizing a speech and responding to it. For instance, Siri of Apple.
There has to be a consideration of many factors such as age, gender, and quality of sound.
Market prediction. Deep learning models can be used to predict the buying and selling
trends of traders depending on the dataset which has been used in training the models. Therefore, it
is easy for a trader to make a decision depending on how they desire to invest (Litjens, et. al, 2017).
Digital advertising. It is easy to categorize users depending on their purchase history as well
as browsing history and recommend relevant advertisements that are personalized and real-time.
Fraud detection. Multiple data sources of deep learning can be used to determine if a decision is a
Deep learning has made tasks that could be tough to be easy. The applications of deep
learning range in different sectors such as health and education. There will be more applications in
the future through the advancement that is being witnessed with the application of deep learning
of limited data points. In this way, the performance of the model can be impacted negatively.
Overfitting can be a problem because its evaluation of machine learning through the data that it is
trained upon can be different from the evaluation which will be performed on the unseen data.
Overfitting can be limited through the use of various techniques such as the use of a resampling
technique to do the estimation of the accuracy of a model and holding back a validation dataset
Preventing overfitting
It is very useful to detect overfitting although it may not solve the entire problem but it can
Remove feature. The inbuilt features which exist for algorithms help in training a model.
However, those that do not have can be removed manually and improve them through
generalization. Sometimes it can be difficult to justify certain features but it can be done through
Regularization. It refers to a broad range of techniques that are used for forcing a model to
be simpler. It depends on the type of learner being used. It can be tuned through cross-validation.
Ensembling. It refers to a machine learning method that is used for combining predictions being
obtained from multiple separate models. It can be through bagging which attempts to reduce the
chance overfitting complex models or boosting which attempts to improve the predictive flexibility
Conclusion
Deep learning is very essential in image processing through its application in image
processing by involving machine learning in the process. It involves modern technology in its
application such as speech recognition and fraud detection. Machine learning can be applied in
problem-solving through the method of induction. Generalization can be used in the application of
new data to a model. Overfitting can perform well on the training data while poorly on other data.
References
Kussul, N., Lavreniuk, M., Skakun, S. and Shelestov, A., 2017. Deep learning classification of land
cover and crop types using remote sensing data. IEEE Geoscience and Remote Sensing Letters,
14(5), pp.778-782.
Janowczyk, A. and Madabhushi, A., 2016. Deep learning for digital pathology image analysis: A
Guo, Y., Liu, Y., Oerlemans, A., Lao, S., Wu, S. and Lew, M.S., 2016. Deep learning for visual
Cha, Y.J., Choi, W. and Büyüköztürk, O., 2017. Deep learning‐based crack damage detection using
pp.361-378.
Bayar, B. and Stamm, M.C., 2016, June. A deep learning approach to universal image manipulation
detection using a new convolutional layer. In Proceedings of the 4th ACM Workshop on Information
Wan, J., Wang, D., Hoi, S.C.H., Wu, P., Zhu, J., Zhang, Y. and Li, J., 2014, November. Deep
learning for content-based image retrieval: A comprehensive study. In Proceedings of the 22nd
Affonso, C., Rossi, A.L.D., Vieira, F.H.A., and de Leon Ferreira, A.C.P., 2017. Deep learning for
Litjens, G., Kooi, T., Bejnordi, B.E., Setio, A.A.A., Ciompi, F., Ghafoorian, M., Van Der Laak, J.A.,
Van Ginneken, B. and Sánchez, C.I., 2017. A survey on deep learning in medical image analysis.
Activity 3
Introduction
Prolog makes the use of logic clauses to solve problems. The clauses contain two parts that
are the right-hand side predicate and the left-hand side predicates. It consists of facts and rules
KB1, KB2, KB3. It is classified according to the number of facts which are contained in a clause.
For instance, KB1 contains a single fact which is true while KB2 contains two facts and KB3
contains more than two facts. All these facts are unconditionally true. The facts and rules will state
how a given problem will be solved to attain a solution. In the facts which have been given on the
question, the determination is the animal is a cheetah depends on the facts and the rules which have
been given. The facts are checked if they ascertain rules for an animal to be a cheetah for the
program to indicate if it is true or false. The output of the program will depend on the facts and
rules which have been input. Therefore, the prolog does not reason the same way humans reason
having a predetermined mind through the different features which they know exist for given animal
Background
Computing questions can best be answered using the knowledge base. Its origin can be
traced back to the capability of the employees to answer questions on emails and telephone calls.
Later they realized that customers often asked similar questions. There was a need for them to
develop a system that could help them answer the frequently asked questions and even solve the
more complex questions more easily. The enhancement of the knowledge base was to help in
answering such questions which required lengthy search. The web features of searching emerged
The different symbols which were used in the knowledge base needed to be understood to
solve the different problems which were faced. The interpretation of the problems for answers
requires that the inbuilt predicates interpret the problem which is being presented to apply the facts
and rules which exist in the program. Lack of specific ways through which files could be formatted
led to the use of knowledge base facts and rules to be able to solve the different problems which
existed. Faster search options were provided by the knowledge base through their easy maintenance
eats(meat).
spots(yes).
colour(tawny).
stripes(no).
feathers(yes).
mammal(yes):- gives_milk(yes).
carnivore(yes):-mammal(yes),eats(meat).
animal(cheetah):-
carnivore(yes),colour(tawny),spots(yes),feathers(no).
animal(lion):-carnivore(yes),colour(tawny),spots(no),feathers(no).
animal(tiger):-
carnivore(yes),colour(tawny),stripes(yes),feathers(no).
animal(tawnyowl):-
carnivore(yes),colour(tawny),spots(yes),feathers(yes).
Animal (cheetah) will be the keyword being searched by prolog. It will run through the facts
and the rules and when an animal (cheetah) is found, it will go back and search the facts which th e
cheetah has. It will proceed to check what is on the left-hand side of the rule. It will run through all
the rules and find different kinds of animals. It will go back to the beginning of the program and
start again. It will proceed to indicate that a cheetah provides milk and it eats meat. The color of the
cheetah is also tawny but it does not have feathers. It will skip the strips and feathers. On reaching
the mammal, on the left-hand side of the rule, it is a fact that it is a mammal that gives milk. It
continues down the rules and at the animal (cheetah) it will be able to find that the animal is not a
eats(meat).
spots(no).
colour(tawny).
stripes(no).
feathers(no).
mammal(yes):- gives_milk(yes).
carnivore(yes):-mammal(yes),eats(meat).
animal(cheetah):-
carnivore(yes),colour(tawny),spots(yes),feathers(no).
animal(lion):-carnivore(yes),colour(tawny),spots(no),feathers(no).
At first, the program will run until it finds an animal (lion) on the left-hand side. It will look
at the right-hand side rule it will find carnivore (yes) then it will go back to check the facts for the
lion being a carnivore. The facts will be true making it to proceed to the next fact of the animal
(lion). The color of the lion and if the lion has spots as well as feathers and it will realize that they
do not exist. All these facts will be true about the characteristics of a lion which have been input
recognize the facts in order to identify the animal as being tiger (Costa, Rocha and Damas, 2012).
gives_milk(yes).
eats(meat).
spots(yes).
colour(tawny).
stripes(yes).
feathers(no).
mammal(yes):- gives_milk(yes).
carnivore(yes):-mammal(yes),eats(meat).
animal(cheetah):-
carnivore(yes),colour(tawny),spots(yes),feathers(no).
animal(lion):-carnivore(yes),colour(tawny),spots(no),feathers(no).
animal(tiger):-
carnivore(yes),colour(tawny),stripes(yes),feathers(no).
Task 5
gives_milk(no).
eats(insects).
spots(yes).
colour(tawny).
stripes(no).
has_feathers(yes).
omnivorous(yes):-spots(yes),eats(insects).
Task 6
gives_milk(no).
eats(insects).
colour(black).
stripes(no).
feathers(no).
amphibian(yes).
mammal(no):- gives_milk(no).
animal(frog):-
amphibian(yes),colour(black),stripes(no),feathers(no).
Insights
Prolog does not know the facts as humans do but it makes use of backtracking in which it
has to go back to the previous facts to determine if they satisfy the requirements of a certain rule.
The query that is made has to be compared with the data that is stored in the database to realize if it
is true or false according to the facts which have been stored. The trace for a given predicate can be
made after checking all the facts and the rules which exist for a given predicate. The relationship
between the facts and rules has to be logical to realize their applicability in various questions
Conclusion
The applicability of the rules and facts written into a program depends on the solution
which they are meant to solve. The different facts represent the knowledge that is fed into a
program to solve a specific problem. The built-in cannot appear as the principal functor in a fact or
the head of a rule. Thus this has to remain the same as changing it can lead to changing the meaning
of the definition. Matching between two terms can say to exist if they are identical or they have
been made to be identical by the look of a variable. The use of variables acts as a guide to the facts
References
Wielemaker, J., Schrijvers, T., Triska, M. and Lager, T., 2012. Swi-prolog. Theory and Practice of
Costa, V.S., Rocha, R., and Damas, L., 2012. The yap prolog system. Theory and Practice of Logic
Wielemaker, J., Lager, T. and Riguzzi, F., 2015. SWISH: SWI-Prolog for sharing. arXiv preprint
arXiv:1511.00915.
Schneider-Kamp, P., Giesl, J., Ströder, T., Serebrenik, A. and Thiemann, R., 2010. Automated
termination analysis for logic programs with cut. Theory and Practice of Logic Programming, 10(4-
6), pp.365-381.
Amaral, C., Florido, M. and Costa, V.S., 2014, June. PrologCheck–property-based testing in prolog.
In International Symposium on Functional and Logic Programming (pp. 1-17). Springer, Cham.
Efremidis, A., Schmidt, J., Krings, S. and Körner, P., 2018, September. Measuring coverage of
prolog programs using mutation testing. In International Workshop on Functional and Constraint
$\lambda $$ Prolog Interpreter. In Logic for Programming, Artificial Intelligence, and Reasoning
Wielemaker, J. and Costa, V.S., 2010. Portability of Prolog programs: theory and case-studies.
Mera, E. and Wielemaker, J., 2013. Porting and refactoring Prolog programs: the PROSYN case
Swift, T. and Warren, D.S., 2012. XSB: Extending Prolog with tabled logic programming. Theory
Figure 7: Poster