2020 Chaudhari A Review Article On Artificial Intelligence Change in Farmaceutical Formulation and Development

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.

org (ISSN-2349-5162)

A REVIEW ARTICLE ON ARTIFICIAL


INTELLIGENCE; Change in Modern Techniques
of pharmaceutical Formulation and Development.
Manoj K Chaudhari1* and Vipul P Patel2
1
Resarch Scholar, Department of Pharmaceutics (PG), Sanjivani College of Pharmaceutical Education and Research Kopergoan
.Maharashtra, India.
2
Proffeser and Head of Department, Department of pharmaceutics (UG), Sanjivani College of Pharmaceutical Education and Research
Kopergoan .Maharashtra, India.

Abstract
Artificial intelligence the branch comes under computer science that deals with having problem solving ability by aid of symbolic
programing. Health care related problems can be solving with huge applications by artificial intelligence. Development of expert
system is an important and useful application of artificial intelligence. Doctors can now work accuately and trustfully, with advantage
of Artificial intelligence. Under the human supervision lot of institutions are employing robots for performing certain activities.
Reduction in time of drug development which results in reduction of cost and time is major advantage of the artificial intelligence. To
make pharmacy profession more efficient large number of researches is being carried out. In this article we can understand Artificial
intelligence application in drug development along with various tools used to do it.

Key words: Artificial intelligence, Drug discovery, ANN in F & D Tools of A.I, Applications of A.I.

1. Introduction:
Artificial intelligence can define as branch of computer science which helps us to problem solving with aid of symbolic programing.
Even you are familiar or not but in the 20th century, we are all addicted by AI. At present live examples of A.I which we are using in
our routine life. For example, social media like what’s app, Facebook, Telegram. Search engines such as Google and yahoo.
[1]Product recommendations, Email filters etc. Development of expert system is a main application of A.I. (Artificial intelligence).
A.I was come into existence in year 1956. People who developed A.I are Allen Newell, Herbert A. Simon. Logic Theorist was first
Artificial Intelligence system saw in 1956. In present market following are the top 10 Pharmaceutical companies which are using A.I
Pfizer, Roche, Novartis, Johnson & Johnson, MSD, Sanfoi, Abbive, Glaxosmithkline, Amgen, Gilead Sciences.[2]following are some
of the tools used by above industries and healthcare institutes IBM Watson for oncology , Robot pharmacy , MEDi robot , Erica robot
, TUG robots , Berg . Machine learning, neural networks, Artificial intelligence (AI) and deep learning are useful in solving problems
which are most critical and take lot of time to solve by humans, but same human like commitment making, deductive reasoning and
conclusion by a computer is yet to be developed; link between artificial intelligence and algorithm have been grown stronger. Voice
recognition, planning, learning, solving of problem are some of the artificial functions fit into the computers. [3]

2. Basics of Artificial intelligence:


Machines can only act like humans and response like humans only if that information about activity is store in that machine. Artificial
intelligence has ability to judge objects, properties, categories and relations between all with the help of information coded in it. Begin
common sense; logically thinking and ability of problem-solving functions in machines is a very difficult work to develop in machine.
A.I. Learning involves Machine learning as a key part with not using or giving any kind of command given need a skill to recognize
patterns in run or flow of inputs, and ability of learning alongside assign control includes numerical regressions and Classification.
Concludes of the category is done by classification an object belongs to and regression compact alongside acquiring a set of numerical
output or input examples, thereby finding of new functions allow the creation of acceptable and satisfactory outputs from the specific
inputs. [4] Theoretical computer science compact alongside mathematical analysis of machine learning algorithms and their
demonstration. Machine recognition deals alongside the ability to employ sensory inputs to gather the different characteristics of the
world, meanwhile computer vision is the potential to analyze visual inputs along with some sub- problems such as facial, object and
motion recognition. Robotics is also a crucial field connected to AI. Robots need smart intelligence and reconnaissance to navigate
tasks such as object handling and navigation, along with sub-problems of mapping, localization and motion planning. Deep learning is
really just a term to express definite types of neural networks and related algorithms that absorb raw input data. Data operation is done
through many layers of nonlinear transformations of the input data in relation to calculate a target output. Unsupervised characteristic
removal is also an zone where deep learning shine.[5] Feature extraction is when an algorithm is able to automatically derive or
construct meaningful features of the data to be used for additional learning, generalization, and understanding. The load is
traditionally on the data scientist or programmer to carry out the characteristic selection in most other machine learning approaches,
along with characteristic selection and engineering. Characteristic selection usually involves some amount dimensionality reduction as
well, which is reducing the amount of input features and data required to create meaningful results. This has many uses, which include
simplification, computational and memory power reduction, and so on. Programmers would teach a neural network to observe an
object or phoneme by wipe out the network with digitized versions of images comprises those objects or sound waves containing
those phonemes. If the network is unable to correctly recognize a particular pattern, an algorithm would adjust the weights. The final
JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1466
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
goal of this training was to get the network to consistently recognize the patterns in speech or sets of images that we humans know as,
say, the phoneme “d” or the image of a dog. [6] This is much the same way a child learns what a dog is by noticing the details of head
shape, behavior, and the like in furry, barking animals that other people call dogs. Evolution of A.I began from early mob who took an
effort to develop it, and the algorithmic approaches over the years included decision tree learning, inductive logic programming. [7]
Clustering, reinforcement learning, and Bayesian networks among others. As we all know, none achieved the ideal goal of common
AI, and even limited AI was mostly out of reach with early machine learning approaches. As it turned out, one of the very best
application premises for machine learning for many years was computer vision, however it still need a great deal of hand-coding to
get the job done. Individuals would go in and write hand-coded classifiers like edge detection filters so the program could point out
where an object started and stopped; shape discernment to decide if it had eight sides; a classifier to acknowledge the letters “STOP.”
From all those hand-coded classifiers they would promote algorithms to make sense of the image and “learn” to determine whether it
was a stop sign. Good, but not mind-bendingly great. Mainly on a foggy day when the sign isn’t ideal visible or a tree obscures part of
it. There’s rise in computer vision and image detection didn’t come close to match humans until very recently, it was too fragile and
too prone to error. Time and the right learning algorithms made all the difference. [8]

3. Terminologies used in Artificial Intelligence.

a. Neural network
In neural network algorithms are arrange in a sequence that line up prior to acknowledge basic sequence in a set of data from a
ingoing process that copy the way the human brain operates and work. Above process, neural networks work on neurons system; it
can artificial and may be natural in nature. Neural networks have ability of modifying to vary input; therefore best possible result is
created by network without needing of output criteria to design again. The concept of neural networks, which has its roots in artificial
intelligence, is quickly gaining favor in trading systems development. [9]

b. Basics of Neural Networks


Neural networks, in industrial evolution money/financial conditions matters, evolution help the existence of process such as
algorithmic trading, time-series forecasting and classification of securities. [10] A neural network has ability works similar and smart
like the neural network of human brains. Mathematical function acting is done by neuron in a neural network that gathers and also
classifies information according to a particular construction. [11]
Interconnected nodes contain by a neural network in the form of layers. Each node is a perceptron, perceptron is an algorithm for
supervised learning of binary classifiers and is same like to a multiple linear regression. [12], Perceptron’s are organized in
interconnected layers in a multi-layered perceptron (MLP). Collection of input patterns is done by input layers. The output layer has
classifications or output signals to which input patterns can map. [13] Input weightings are fine-tune by hidden layers up to the neural
network’s edge of error is least. With regard to the outputs assumption is done that hidden layers gather salient features in the input
data that consist of predictive power. Realization of a utility similar to statistical techniques such as principal component analysis is
done by feature extraction of outline. [14]

c. Fuzzy Logic
Fuzzy logic is highly used in expert system it is a modern control systems. Fuzzy logic has ability to solve problem same like human
being but much more faster than human brain and same related with decisions making. It workout by gathering of information and
converting into more significant information by line up of limited truths as Fuzzy sets. In the large company business it is used for
decision making support systems and personal development. It has benefit in chemicals manufacturing industry for maintaining the
pH of chemicals, drying, and chemical distillation process. [15] Natural language processing and various applications in Artificial
Intelligence are grabbing by fuzzy logic. Fuzzy Logic is used in company with benefit of Neural Networks as it imitates make
decisions rapid and easier as compare to human brain. This is made successful by collection of data and converting that data into more
meaningful data with the help of by forming partial truths as Fuzzy sets. [16]

d. NeuroFuzzy logic
Learning algorithm are used by fuzzy system and this algorithms are obtains from neural network theory to govern its parameters by
data samples processing. Representation of Modern neuro-fuzzy systems are usually done as special multilayer feed forward neural
networks e.g. ANFIS, GARIC, NEFCLASS etc. [17] Following is an considered example of fuzzification of neural network
architecture; Self-organizing feature maps. Propagation, connection weights and activation functions vary from common neural
network in neuro fuzzy networks. According to data collected different approaches found but most widely use term in neuro fuzzy
system for approached which shows following different properties: Neural network theory helps fuzzy system by training it with
algorithm provides by theory of neural network. Local information, procedure of learniing runs on it and also makes local moderation
in basic fuzzy system. A 3-layer feed forward neural network is a outlook of neurofuzzy system. Input variables represented by first
layer, fuzzy rules represented by middle which is also known as hidden layer and third and last layer output variables. Encoding as
(fuzzy) connection weights is done of fuzzy sets. It’s not mandatory to do representation of a fuzzy system like this to apply a learning
algorithm to it. [18]

e. EVOLUTIONARY COMPUTING:
It is common term which elaborates computational process in which evolution of solutions done with help of rules of inheritance;
recombination, mutation and selection. In formulation research one particular subset of this, evolutionary algorithms play an
important role as its application to formulation research. [19]

f. GENETIC ALGORITHMS
John Holland in 1970 made use of Genetic algorithms. Search technique especially suited to optimization; assumption of trial
population, and this evolves an iterative process is provided by genetic algorithms. Initial population of solution is generated and
fitness of every member of the population is evaluated during this process.
JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1467
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
g. MACHINE LEARNING(ML)
ML is a method which is used for doing analysis of data in which building of analytical model is done automatically using algorithms
that frequently learn from data or important information. Machine learning classified into three classes that are supervised learning,
unsupervised learning, reinforcement learning. And these three classes again sub classified into supervised learning into classification
which result in disease diagnosis and regression which result into drug reinforcement/ADMET prediction. Unsupervised into
clustering which result in disease sub type discovery and feature finding which result into disease target discovery. Reinforcement
learning into decision making which result de novo drug design and excision into experiment design.

Figure 1; Machine learning Classification

4. PROGRAMING LANGUAGES USED IN A.I

SR NO LANGUAGE USES
1 Python Most Effective language for A.I,Machine learning and A.I algorthms
can easily implimented,
2 Prolog Used in knowledge based and expert system,Pttern mattiching,Prebase
data structuring,Automatic Tracking.Use in medical project.
3 Java Simplified work with large scale project, good user interaction,use for
making graph and interfaces.
4 Lips Oldest and most suited for A.I,Develop by father of A.I,Capable of
processing symbolic information,Create Dynamic object easily.
5 R Statistical Programming language,analysing and manupulating data
for statistical purpose,Produce methamathical symbol.
Table no 1; PROGRAMING LANGUAGES USED IN A.I
JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1468
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
5. Introduction to Artificial neural network [ANN]:
With the help of lot of experience in algorithms learning can lead to simulation of some functions of human brain by ANN in
computer programing. ANN have sensational information processing features like, parallelism, unperfected and non-success
tolerance, learning, non-linearity, robustness, skill to drive imprecise and fuzzy information and their ability to discover.
Optimization, pattern classification, clustering, pattern classification such real life complicated problems can be solve with help of
ANN.ANN models have developed many applications.
On basis of ground root learning of algorithm, coaching of ANN model consists of supervised and non-supervised.
In supervised training data sets introduced as an input/output data sets; Then unsupervised training, introduction of input data is done
individually, and identification of patterns in data is learn by model. On topology basis the connection of ANN could be feed forward
or it can be feedback. The connection between the nodes never forms cycle in feed forward ANN. The model of ANN must
recapitulate long time with potential before production of a response is done in some of the feedback ANN model when presentation
of input is done each time. Feed forward model of ANN are easy to train as compare with feedback ANN models.
Some of types of ANN models are categories into three types are associating networks, features extracting network, non-adaptive
network. For classification of data and its prediction associating networks are applied here; for organization of supervised learning
need correlated output (dependent variable) and input (independent variable). For data dimension reduction feature extracting network
are mostly used and to do unsupervised or competitive learning only input values are needed. [36]

a. Artificial Neuronal Network in Design of controlled release drug delivery systems


Huge advantage obtained from controlled released drug delivery system over the conventional drug delivery system.[34] However,
due to the reason of present of complexity of this drug delivery system there are prominent dares to controlled release drug delivery
systems development.[35] For formulation, development and evaluation of controlled release drug delivery system with help of
Traditional statistic RCM(response surface methodology) is developed techniques, but there is presence of lot of limitations and
restrictions.[36] Due to the restrictions of RCM, another evolutionary and advanced technique called artificial neural network (ANN)
have left behind RCM and gain pick of popularity for control release drug delivery system dosage forms development.[37]
In formulation of CRDDS to develop it associating networks are applied. The nonlinear relationship is seen between drug release
profiles of controlled drug release system and formulation and process factors, and also understanding is not well. Associating
networks have ability to map or script the link between the formulation and process variable and response like in vitro drug release
profiles training and learning process. [38] Associating networks used to generate the response such as in vitro drug release profile of
newly discovered drug or formulation which is prepared with varying composition and manufacturing process. Optimization can also
be done with help of associating network model in which trained and learned ANN play important role in selecting optimal
formulation and process parameters. [52]
Use of ANN can be used in formulation by some of the steps which are, first of all data should be extracted from the research papers.
Data can be encoded into software for formulation and development. Then data is set and training of that data is done, validation is
done and testing of that data is done. Then with the help of neural network decision making is done and prediction model is get
created and with help of prediction accuracy outcomes of result related to formulation and development of dosage form or product is
done.

Figure 2; the flowchart of establishing model

b. ANN in Formulating Microsphere of Poly lactide co glycoside (PLGA) .


Glycolic acid and lactic acid, PLGA is a co polymer of these two polymers. Type of drug, Size of Particle, microspheres morphology,
conditions of release, etc. are some of the properties on which drug release of PLGA microspheres depend. Batch of applicable
properties selection for PLGA is really a challenging task for machine learning; it is more demanding because there are more than
four hundred considered characteristic. For this work, we construct the selection of analytic characteristics for PLGA as a multi-
objective optimization issue with the objective of reducing the mistakes of forcasting the dissolution profile while minimizing the
quantity of attributes chosen.[39] For selecting the ideal property set for forecasting the dissolution profile of PLGA Four bio-inspired
optimization algorithms: antlion optimization, binary version of antlion optimization, grey wolf optimization, and social spider
optimization are selected. Besides these, for comparisons LASSO algorithm is also used. Choosing of important and crucible
variables is carry out under the consideration that both predictability and model simplicity are of twin importance to the decider result.
Input variables set are applied to search least generalization error over different predictive models and their construction during the
feature selection process. The methodology assess making use of forecasting modeling for this work different tools are selected, these
JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1469
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
tools are random forests, Cubist, artificial neural networks (monotonic MLP, deep learning MLP), multivariate adaptive regression
splines, regression tree and classification, and blend systems of fuzzy logic and evolutionary computations (fugeR). The comparison
of experimental result is done with the results pres by Szlȩk. We get a systematic and normalized root mean square error (NRMSE) of
15.97% vs 15.4%, and the number of selected input features is very smaller, 9 versus 11. [37]

c. ANN in Modified Release Solid Dosage Forms.


Application of QbD (Quality by design) approach in pharmaceutical product development has force researchers or impels researches
in the pharmaceutical industry to apply DoE that is Design of experiment as statistical tool, in development of product. RSM that is
Response surface methodology is most widely used in among all DoE techniques. Pharmaceutical development is more in lead due to
progress of computer science. In drug formulation machine learning tool take an important place along with implementation of
statistical tool. The first paper which was describing about application of ANN in modified drug release product was found twenty
years back. After that humeral work is done by researches towards application of new techniques, most preferably ANN in modeling
of production, drug release and drug stability of modified released solid dosage forms. [54]
d. Drug development process
Job of searching effective and post marketing surveillance new drugs is formidable and always the toughest and lot of effortful a part
of development of drug. it's origin by the massive size of what's called chemical space, it's approximate to be within the sequence of
1060 molecules. Possible uses of AI come up with the possibility to counter the disorganization and unreliability that appear within
the classical drug development methods while reducing bias and human interference within the process.
The other uses of AI in drug development involve the forecasting of feasible synthetic routes for drug-like molecules ,
pharmacological properties , protein characteristics in addition as efficacy, drug combination and drug–target association and drug
repurposing. Also, the spotting of recent pathways and targets using omics analysis becomes practicable via the creation of novel
biomarkers and therapeutic targets, personalized medicine supported omics markers and discovering the connections between diseases
and medicines. [20]
e. A.I and ML in Non-compartmental data analysis.
SimBiology and MATLAB are two software which plays important role in non-compartmental data analysis. There is presence of one
tool in MATLAB called SimBiology which is helpful for pharmacokinetic and pharmacodynamics model. For terminal slope and
linear interpolation method, parameter estimation work is done by SimBiology with the help of unweighted linear regression.
SimBiology is also helpful for clinical pharmacology. Fullform of MATLAB is (Matrix Labrotary), mostly used for simulation,
numerical analysis and modeling. AUC (Area Under Curve) is also find with help of SimBiology which uses linear trapezoid method
to calculate it. [39].

6. Tools of Artificial Intelligence


Different types of tools have been come into exist to solve the different problems of pharmaceutical industry. Outcomes of these tools
have been promising. Some of the famous tools that have achieved mind-blowing popularity are as follow.

a. IBM Watson for oncology


Supercomputers develop by IBM name as Watson.it is combination of sophisticated analytical software and A.I device design for
answering question. It is design for treatment of cancer, it assist oncologist in taking better decision. It works by analyzing the
medical data of a patient from a wide network of data and expertise and then allotting treatment options based on the information
obtained. It is effective in analyzing meaning and context data in clinical notes and reports, May be properly structured or
unstructured. It can simply gather analytical information in regard to the patient and write it in plain English which can turn out to be
a very critical step in providing the correct treatment plan for the patient. It merges critical characteristics from the file of a patient
with external research, clinical research, and data and then suggests or decides the most effective treatment plans that can be apply for
a patient. Watson has an enormous assembling of information from literature and rationales curated by MSK, over 200 textbooks, 12
million text pages, and over 290 medical journals. [51] At present, one of the Indian software engineer, aged 39 years, was diagnosed
with breast cancer of a rare form and it was developing across both her breasts very fast which was posing a threat of both breast
removals. Her medical records along with her genomics data were fed into Watson by Dr. Somashekhar, an oncologist in Bengaluru,
and Watson provided viable treatment options within 60 s. [44]

b. Robot pharmacy
With the Aim of increasing the safety of patients, UCSF Medical Center makes use of robotic technology for the preparation and
tracking of medications. According to their report, the technology has successfully prepared 3,50, 000 medication doses with
negligible error. The robot has proved to be far better than humans both in size as well as its ability to deliver accurate medications.
The abilities of the robotic technology include preparation of oral as well as injectable medicines which include chemotherapy drugs
that are toxic. This has given free space to the pharmacists and nurses of UCSF so that they can make use of their expertise by
focusing on direct patient care and working with the physicians. [53]
Within the automated system of the pharmacy, the computers first collect medication command electronically from the physicians and
pharmacists of UCSF. After this, individual doses of pills are collected, packaged, and dispensed by the robotics. This is followed by
machines assembling the doses onto a bar-coded plastic ring. The thin plastic ring contains all medications that have to take by a
patient within a period of 12 h. Adding to the capabilities of the automated system is their ability to prepare sterile preparations that
are meant for chemotherapy along with filling of intravascular syringes with the right medications.
The automated facility also consists of an inventory management system that keeps track of every product along with a refrigerated
and two non-refrigerated pharmacy warehouses for providing with storage and withdrawal of supplies and medications. All these
facilities are fully automated. [53]

JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1470
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
c. MEDi robot
MEDi is known for its medicine and engineering designing intelligence. One of the professor of Community Health Sciences at the
University of Calgary in Alberta led a project for pain management robot. Idea of developing this robot she got when she saw children
scream during medical procedure when she was working in a hospital. [38] The robot first builds friendly relation with the children
and then tells them what to expect during a medical procedure. During the medical procedure, it instructs them on what should be
done, how to breathe during the procedure, and how to cope. [55] As we know robot cannot think, plan, or make decision, it can be
programmed in such way that it shows to have AI.

Aldebaran Robotics manufacture MEDIi, it is consisting internally build facial recognition technology, it is having ability to speak 20
variety of languages and is highly adaptable to different situations. The retail price of the robot is $9000, however, the cost rises to
$15000–$30000 when the applications needed for the robot to help in medical procedures are installed. The robot was initially
developed for pain management, but with time its use has expanded to comfort between procedures, physical rehabilitation, and
fundraising. [33]
d. Erica robot
Hiroshi Ishiguro, a professor at Osaka University in japan developed Erica, a new care robot. Erica was developed in collaboration
with the Japan Science and Technology Agency, Kyoto University, and the Advanced Telecommunications Research Institute
International (ATR). It can speak Japanese and has a combination of European and Asian facial features. [34] Like any normal human
being, it likes animated films, desire to visit south-east Asia, and wants a life partner who would chat with it. The robot cannot walk
independently; however, it has been developed with the ability to understand and answer questions with human-like facial
expressions. Erica is the “most beautiful and intelligent” android as Ishiguro fixed up the features of 30 beautiful women and used the
average for designing the robot’s nose, eyes, and so on. [35]

Optimization of formulation using ANN


Design of new network help in, making choice of the Self organizing feature maps (SOFMs) network model.[29] To forecast the
output responses use of the functions like TanhAxon, SigmoidAxon, LinearTanAxon, Linear SigmoidAxon, and Axon can used.[30]
Differing activations task can be presented by performance parameters like mean squared error (MSE), minimum absolute error,
correlation coefficient and predicted output of the selected network models.[31]

7. Partnerships between artificial intelligence (AI) and pharmaceutical companies and areas of collaboration in drug
development.[32]

SR.NO PHARMA AREA OF COLLEBRATION ARTIFICIAL


INTELLIGANCE
1 Astellas pharma Drug repurposing Biovista

2 Bayer pharma To track real-time data via smartphones and Xbird


other wearable technologies

3 Roche to target personalized medicine using medicine Bina


learning and large scale genome sequencing .
4 Sumitomo dainippon To identify new tratments for psychiatric Exscientia
Pharma diseases
5 GlaxoSmithKline To discover novel and selective small molecule Exscientia
6 GlaxoSmithKline To identify novel biological target and pathway Insilico medicine
7 Abbvie To announce the mechanism of AI based AiCure
patient monitoring platform can improve
adherence.
Table no 2; Partnerships of artificial intelligence (AI) and pharmaceutical companies and areas of collaboration in drug
development.

8. Conclusion:

Human being is most creative and unpredicted knowledge creative and storage machine that can be ever come into an exit. But this
above line can be agreed only for few years and decades ago until A.I comes into existence. A.I 1000 times faster, error free, time
saving, cost saving, efficient work ability than human beings. But take into note than A.I technology created by human beings only
and let’s see in feature that it becomes hell or heaven as we know words spoken by Stephen Hawking said, “This may mean the end of
human race”. Everything has its two side one GOOD and another BAD. But evolution of A.I has result in decrease human load.

References:

1. Flasiński M. Introduction to Artificial Intelligence. 1st ed. Switzerland: Springer International Publishing; 2016. p. 4.
2. Cellan Jones R. Stephen Hawking Warns Artificial Intelligence could End Mankind. Available from: http://
www.bbc.com/news/technology-30290540. [Access 12 may. 2019

JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1471
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
3. Castrounis A. (2017). Artificial Intelligence, Deep Learning, and Neural Networks,
Explained.[online]Kdnuggets.com.Availableat:http://www.kdnuggets.com/2016/10/artificial-intelligence-deep-learning-
neural- networks-explained.html [Accessed 28 Sep. 2017]
4. Statistical Artificial Intelligence (AI). Available from https://www.statista.com/study/38609.[Accessed 1 june 2019]
5. Rowe R. C, Colboum E. A, Phaem. Tech. Eur. 9 (1996)46-55.
6. Qwek S.S, Lim C.P, Peh K.K, Int. J. Comp. Intel. Appl.1 (2001) 187-292.
7. Goh W.Y, Lim C.P, Peh K.K, K Subari, Neural Comput. Appl. 10 (2002) 311-317.
8. Holland J.H, Adaptation in Natural and Artificial Systems, The University of Michigan Press, Ann Arbor, MI, USA (1975).
9. Berenji H. R. and Khedkar P., Learning and tuning fuzzy logic controllers through reinforcements. IEEE Trans. Neural
Networks 3 (1992) 724--740.
10. Bezdek J. C, Tsao E. C, and Pal N. R, Fuzzy Kohonen clustering networks, in: Proc. IEEE Int. Conf. on Fuzzy Systems (San
Diego, 1992) 1035--1043.
11. Buckley J. J. and Hayashi Y, Fuzzy neural networks: A survey, Fuzzy Sets and Systems 66 (1994) 1-13.
12. Buckley J. J. and Hayashi Y, Neural networks for fuzzy systems, Fuzzy Sets and Systems 71 (1995) 265-276.
13. Halgamuge S. K. and Glesner M, Neural networks in designing fuzzy systems for real world applications, Fuzzy Sets and
Systems 65 (1994) 1--12.
14. Jang S.R, ANFIS: Adaptive-Network-Based Fuzzy Inference Systems, IEEE Trans. Systems, Man & Cybernetics 23 (1993)
665-685.
15. Nauck D, Klawonn F, and Kruse R, Foundations of Neuro-Fuzzy Systems, (Wiley, Chichester, 1996, to be published).
16. Nauck D. and Kruse R, Designing neuro-fuzzy systems through back propagation, in: W. Pedrycz, Ed., Fuzzy Modeling:
Paradigms and Practice, (Kluwer, Boston, 1996) 203-228.
17. Dasta JF. Application of artificial intelligence to pharmacy and medicine. Hosp Pharm 1992; 27:312-5, 319-22.
18. Vuorimaa P, Fuzzy self-organizing map, Fuzzy Sets and Systems, 66 (1994) 223--231.
19. Tschichold German N, Rule Net - A New Knowledge--Based Artificial Neural Network Model with Application Examples
in Robotics, PhD thesis (ETH Zürich, 1996).
20. Kit-kay M and Mallikarjuna Rao P. Artificial intelligence in drug development: present status and future prospects. Drug
discovery today. 2019:24(3):773-780.
21. Wang Q. et al. (2017) A novel framework for the identification of drug target proteins: combining stacked auto-encoders
with a biased support vector machine. PLoS One 12, e0176486.
22. Ferrero E. et al. (2017) in silico prediction of novel therapeutic targets using gene-disease association data. J. Transl. Med.
15, 182.
23. Bakkar N. et al. (2018) Artificial intelligence in neurodegenerative disease research: use of IBM Watson to identify
additional RNA-binding proteins altered in amyotrophic lateral sclerosis. Acta Neuropathol. 135, 227–247.
24. Reymond J.L. et al. (2010) Chemical space as a source for new drugs. Med. Chem. Commun. 1, 30–38.
25. Okafo G. et al. (2018) Adapting drug discovery to artificial intelligence. Drug Target Rev. 2018, 50–52.
26. Segler M. S. et al. (2018) Planning chemical syntheses with deep neural networks and symbolic AI. Nature 555, 604–610.
27. Mayr A. et al. (2016) DeepTox: toxicity prediction using deep learning. Front. Environ. Sci. 3, 80.
28. Gayvert, K.M. et al. (2016) A data-driven approach to predicting successes and failures of clinical trials. Cell. Chem. Biol.
23, 1294–1301.
29. Rodrigues T. et al. (2018) Machine intelligence decrypts b-lapachone as an allosteric 5-lipoxygenase inhibitor. Chem. Sci. 9,
6899–6903.
30. Luechtefeld T. et al. (2018) Machine learning of toxicological big data enables read across structure activity relationships
(RASAR) outperforming animal test reproducibility. Toxicol. Sci. 165, 198–212.
31. Bain E.E. et al. (2017) Use of a novel artificial intelligence platform on mobile devices to assess dosing compliance in a
Phase 2 clinical trial in subjects with schizophrenia. JMIR Mhealth Uhsealth 5, e18.
32. Perez Gracia, J.L. et al. (2017) Strategies to design clinical studies to identify predictive biomarkers in cancer research.
Cancer Treat. Rev. 53, 79–97.
33. Ryu, J.Y. et al. (2018) Deep learning improves prediction of drug–drug and drug– food interactions. Proc. Natl. Acad. Sci. U.
S. A. 115, 201803294.
34. Martin E, Mukherjee P, Sullivan D, Jansen J. Profile-QSAR: A novel meta-QSAR method that combines activities across the
kinase family to accurately predict affinity, selectivity, and cellular activity. J. Chem.Inf. Model. 2011, 51, 1942–1956.
35. Merget B, Turk S, Eid S, Rippmann F, Fulle S, Profiling Prediction of Kinase Inhibitors: Toward the Virtual Assay. J. Med.
Chem. 2017, 60, 474–485.
36. Varnek A, Baskin I, Machine Learning Methods for Property Prediction in Chem informatics: Quo Vadis? J. Chem. Inf.
Model. 2012, 52, 1413–1437.
37. Lo Y.C, Rensi S.E, Torng W, Altman R.B. Machine learning in chemo informatics and drug discovery. Drug Discov. Today
2018, 23, 1538–1546.
38. Lima A.N, Philot E.A, Trossini G.G, Scott L.B, Maltarollo V.G, Honorio K.M. Use of machine learning approaches for
novel drug discovery. Expert Opin. Drug Discov. 2016, 2016 11, 225–239.
39. Ghasemi F, Mehridehnavi A, Pérez A, Pérez H, Neural network and deep-learning algorithms used in QSAR studies: Merits
and drawbacks. Drug Discov. Today 2018, in press.
40. Keiser M.J, Roth B.L, Armbruster B.N, Ernsberger P, Irwin J.J, Shoichet B.K, Relating protein pharmacology by ligand
chemistry. Nat. Biotechnol. 2007, 25, 197–206.

JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1472
© 2020 JETIR September 2020, Volume 7, Issue 9 www.jetir.org (ISSN-2349-5162)
41. Pogodin P.V, Lagunin A.A, Filimonov D.A, Poroikov V.V, PASS Targets: Ligand-based multi-target computational system
based on a public data and naïve Bayes approach. SAR QSAR Environ. Res. 2015, 26, 783–793.
42. Mervin L.H, Afzal A.M, Drakakis G, Lewis R, Engkvist O, Bender A, Target prediction utilizing negative bioactivity data
covering large chemical space. J. Cheminform. 2015, 7, 51.
43. IBM. IBM Watson Health. Available from: https://www.ibm.com/watson/health/oncology and genomics/oncology/.
44. Abrar P. IBM’s Supercomputer Helps Doctors Fight Cancer Available from: http://www.thehindu.com/business/IBM’s-
Supercomputer-helps-doctors-to-fight-cancer/ article14556945.ece.
45. University of California San Fransisco. New UCSF Robotic Pharmacy Aims to Improve Patient Safety. Available from:
https://www.ucsf.edu/ news/2011/03/9510/new-ucsf-robotic-pharmacy-aims-improve-patient-safety.
46. McHugh R, Rascon J. Meet MEDi, the Robot Taking Pain Out of Kids’ Hospital Visits. Available from: http://
www.nbcnews.com/news/us-news/meet-medi-robot-taking-pain-out-kids-hospital-visits-n363191.
47. Pantozzi J. This Robot can help Kids through Chemo, Vaccinations and other Scary Medical Procedures. Available from:
https://www.themarysue.com/medi-robot-for-kids-medical-procedures.
48. Trynacit K. MEDi Robot to Comfort Patients in Stollery Children’s Hospital. Available from:
http://www.cbc.ca/news/canada/edmonton/medi-robot-to-comfort-patients-in-stollery-children-s-hospital-1.3919867.
49. Eye for Pharma. Artificial Intelligence - A Brave New World for Pharma. Available from: http://www.social.
eyeforpharma.com/clinical/artificial-intelligence-brave-new-world-pharma.
50. McCurry J, ‘most intelligent’ Android, Leads Japan’s Robot Revolution. Available from: http:// www.thehindu.com/todays-
paper/tp-national/ Erica-%E2%80%98most-intelligent%E2%80%99-android-leads-Japan%E2%80%99s-robot-revolution/
article13974805.ece.
51. Montgomery D.C, Response surface methods and other approaches to process optimization, in; Design and analysis of
experiments, wiley, new york, 2001,pp. 427-500
52. Yichun S, Yingxu P, Yixin C, Application of artificial neural networks in the design of controlled release drug delivery
systems, Advanced drug delivery reviews, www.elsevier.com/locate/addr, page no 1202-1209.
53. Hossam M. Z, Jakub S, Crina G, Computational intelligence modeling of the macromolecules release from PLGA
microspheres.- focus on feature selection. PLOS one, 17 june 2016, page no 1-17.
54. Svetlana I, Jelena D, Zorica D, Artificial neural network in evaluation and optimixation of modified release solid dosage
forms, MDPI, ISSN 1999-4923. 18 oct 2012, Page no 32-47.
55. Jin-Sol P, Jung-Ryul, Non-compartmrntal data analysis using SimBiology and MATLAB, Transl Clin Pharmacol, ISSN
2289-0882, 19 sep 2019, Page no 89-91,

JETIR2009198 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 1473

You might also like