Module 3 : Artificial Intelligence in Systems Biology: Opportunities in Biomedicine and Healthcare:

OBJECTIVES
• To introduce artificial intelligence in systems biology
• To explain artificial Intelligence methodologies, tools in systems biology,
including machine learning tools used in genome analysis, and variant calling
• To list artificial intelligence applications in systems biology – agriculture,
drug discovery, biomedicine, oncology, identifying variants/mutations from
genetic data, and workflow analysis for genome analysis
• To use case studies to know the importance of artificial intelligence in Systems
Biology towards pharmacogenomics, cancer cure, and COVID-19 pandemic
• To highlight artificial intelligence's future challenges in systems biology,
including gene prediction, protein-protein interaction, and neural networks
INTRODUCTION TO ARTIFICIAL INTELLIGENCE (AI)
The artificial intelligence (AI) concept began in the mid-1950s to understand intelligence, its origin, and creation, in
natural and artificial environments.
It is a multidisciplinary field that extends out and is stimulated by a great diversity of other fields in perpetual
motion.
AI involves ideas, approaches, and systems as a part of the computer program that exhibits characteristics similar
to intelligent behavior. Some degree of intelligence is observed in animals often the same scope of intelligence was
observed by human action.
Ex: the creation of aircraft that is capable of mimic and express human intelligence, thought, consciousness, and
emotions.
Advances in diverse fields created the environment to apply artificial intelligence to accomplish complicated tasks
logically and easily.
US National Research Council of the National Academies affirmed numerous sub-disciplines of biology (Figure 7.1)
and the integration of AI with diverse sub-disciplines of science to gain knowledge on bio-logical systems (Shukla
Shubhendu and Vijay, 2013).
Recent application of AI technology in cybernetics deals with two complementary issues:
(1) AI in systems biology have helped to formulate theories, methods, algorithms, and helps to generate big data
and its implications for a given task;
(2) Biological systems are so complex it needs inter- and transdisciplinary expertise to understand the principles.
“Systems biology is a universal approach to decipher the intricacy of biological systems while synthetic biology or
artificial biology is a blend of engineering and biological sciences with an idea to generate novel biological models”
INTRODUCTION TO ARTIFICIAL INTELLIGENCE (AI)
FIGURE 7.1 Sub-disciplines of artificial intelligence (adapted from Shukla, Shubhendu,
and Vijay, 2013; with copyright permission from IJSER publisher).
INTRODUCTION TO ARTIFICIAL INTELLIGENCE (AI)
Systems biology can solve the most complex problems in medicine, drug discovery,
and biomedical engineering.
AI technology application in systems biology includes text-mining, qualitative physics, statistical
interpretations, machine learning, and deep learning algorithms, etc.
Hence, AI technology in systems biology is helping to do quality research leading to solutions that
are unresolved until now.
Characteristics of AI. :
AI technology has been widely used to resolve complex problems in diverse areas such as drug
discovery and coral reef systems biology.
AI also makes a significant contribution to capturing, distributing, developing, and usage of these
resourceful data for human welfare.
AI includes machine learning, deep learning, knowledge management, and mul-tiagent systems.
AI provides a scientific platform by incorporating and boosting the skills of humans and computers
via interaction between the human-machine processor, computational ingenuity analysis, and array
recognition and practical knowledge via computer theories and data mining
Characteristics of AI. :
• AI plays a significant role in the pharmaceutical industries for the discovery of novel
drugs.

• AI researchers should help in understanding systems biology as traditional statistical

approaches have limitations.

• Systems biology in life science has resulted in the construction of a “complete”

computational model to capture knowledge about important systematic properties,
3Dstructures, and functions of each biological unit.

• It operates on a basic understanding of the system via simulations.

• To bridge the technological gap between man and machine, computational models play a
vital role in better perception of biological systems, collection of integrative data,
developing new methods, and tools for systems biology.
Characteristics of AI.
AI METHODOLOGIES AND ALGORITHM FOR SYSTEMS BIOLOGY
• Microarray technology is useful in analyzing the expression of multiple genes that leads to various
diseases. Using computational methods, it is possible to detect de novo mutations that are generated via
sequencing technology.
• Mutations in the exonic region have been identified through reading processing, aligning, mapping, and
variant calling.
• Algorithm for read processing such as Next Generation Sequencing (NGS) Data QC toolkit, Cutdapt, and
FastX have been used to trim out the low-grade adapters in sequencing.
• To identify the sequence similarity, the processed reads were compared with the reference genome,
followed by multiple sequence alignment using NovoAlign, BWA, STAR, TWAP, TOPHAT, and RNA
sequencing.
• GATKs (Genome Analysis Tool Kits) are widely utilized for intel alignment base quality recalibration, variant
calling, and also to find precise mutants from sequencing, mapping, genomic libraries, and sample
enrichment.
• Numerous germline and somatic mutant calling algorithms were developed for detecting low-frequency
mutants from the complex system through an error correction technique.
• Genomic mutants fall under four major groups like insertion, deletion (del), structural variants like
duplication, translocation, copy number variants, and SNVs (Single Nucleotide Variants).
• SNV is ≤10 bp and it is to check non-reference nucleotides from the pile of sequence that covers each
position. The tools used for germline variant calling involves GAP and MAG integrated by GATK, whereas
CRISP and Thunder are used for the analysis of variant calling and pooled samples data.
MACHINE LEARNING (ML)
• Machine learning (ML) in artificial intelligence enables the machine to make decisions by
themselves without explicit programming.
• Machine learning includes the implementation of genetic algorithms (GAs).
• Genetic algorithms are defined as a perfect algorithm which includes Darwinian
evolutionary theory (survival of the fittest) and the easiest form of genetic process.
• Depending upon the biological operators such as mutation, crossover, and selection,
GAs are widely used to produce good solutions to biological problems.
• operator instigates a change in the final solution of the problem so that the diversity in the
population could be maintained, while a crossover operator mimics biological
recombination
MACHINE LEARNING (ML)
Relationship between ML, DL, and AI (adapted from Moore and Raghavachari, 2019;
copyright permission from Frontiers publishers).
ML ALGORITHM APPLIED IN SYSTEMS BIOLOGY
Cell Profiler: Usually, biological images are analyzed by a single parameter from a group of images.
Cell Profiler is a software developed by Anne Carpenter and her colleagues at MIT and Harvard for
the quantitative analysis of cell numbers using a fluorescent microscope. Presently, by applying
deep learning techniques, a cell profiler can be utilized in an array of biological fields.
Deep Variant: Data mining of whole-genome sequence (WGS) is possible only using Deep learning
(DL) algorithm. With the application of a deep learning algorithm, a systems tool called “Deep
Variant” was developed by Verily life science and Google to predict common genetic variations in a
more accurate and precise manner.
Atomwise: “Atomwise” is a biotech firm in San Francisco that has developed an algorithm with
atomic accuracy to convert molecules into three-dimensional shapes of amino acids .
In biomedicine, Using DNN, the genomic and proteomic data were analyzed to detect the novel
biological markers. In cheminformatics, deep learning is used in identifying the target
drug (Mamoshina et al., 2016).
ML ALGORITHM APPLIED IN SYSTEMS BIOLOGY
• In single cells, DNA methylation stages were envisaged using a computational tool called
“Deep CpG” that was recently established by CNN.
• A Deep CpG tool is used for predicting the motifs appropriate for DNA methylation.
• Deep CpG is predominantly used in studying epigenetic markers (Angermueller et al.,
2017).
• Tensor Flow: It was first established by Google researchers using a deep learning (DL)
algorithm.
• Tensor Flow is involved in accelerating neural network strategy.
• Tensor Flow is employed in various fields for time management and graphical
predictions.
• Recent advancements in Tensor Flow entrenched towards a supporting tool called
“Tensor Board” that is used for the visualization of complex models (Rampasek and
Goldenberg, 2016).
ML ALGORITHM APPLIED IN SYSTEMS BIOLOGY
COMPUTATIONAL NEURAL NETWORKS
• Neural networks are the second major bottom-up approach in attaining artificial
intelligence, while deep learning comprises an array of neural network that
transmits the data from one neuron to the other via hidden layers
• These hidden layers help to resolve complex problems by naturally mimicking the
human brain
• Artificial Neural Networks (ANNs) require machine learning algorithms to significantly
analyze the input data.
• Artificial neurons comprise various inputs and a single output. Simultaneously, one
neuron activates the other neuron using precise activation functions like
(1) sigmoid function
(2) step function
(3) linear function
(4) ramp function and
(5) hyperbolic tangent function.
COMPUTATIONAL NEURAL NETWORKS
ANN comprises
(1) an input layer that receives the input signals or values;
(2) hidden layers, which are a group of neurons separating the input and output layer; and an
(3) output layer, which has one neuron and its output value range between 0 and 1
(Shanmuganathan, 2016).
• A hidden layer in ANN screens the pooled data interconnects with each layer and
provides an accurate output.

• DL techniques are vastly used in biomedicine, cellular imaging, and analysis of whole
genome by the mining of huge data sets and creating a computational model.

• Hence, deep learning in biological big data analysis provides a better knowledge of the
high dimension data set, detecting regulatory mutants, and also analyzing the rate of
mutation via DNA sequencing of the whole cell or tissue samples.
COMPUTATIONAL NEURAL NETWORKS
COMPUTATIONAL NEURAL NETWORKS
PROS AND CONS OF ARTIFICIAL NEURAL NETWORK (ANN)
APPLICATIONS OF ARTIFICIAL INTELLIGENCE (AI) IN AGRICULTURE, BIOMEDICINE, AND HEALTHCARE
• AI IN AGRICULTURE :
• AI technology is precisely used in agriculture to detect diseases, pests, and proper nutritional
supply to the plants in the field.
• In agriculture, the use of AI falls under three main categories.
• (1) Agricultural robots are developed to harvest the crops at higher volume and a faster rate than
humans; For instance: Initiation of Blue-River Technology has led to the development of “See and
Spray Robot” which control weeds in the cotton field.
• (2) Crop and soil health is monitored using deep learning algorithms and software-based
technologies.
• Ex :The development of “Harvest CR00 Robots” has benefited the strawberry growers to gather
and box the fruits with less labor cost. In recent times, PEAT a Berlin-based agricultural firm has
established an application named “Plantix” which detects diseases, pests, and certain soil
defects.
• (3) AI-based machine learning robots were established to forecast the crop yield and weather
changes. “aWhere” is a Colorado-based firm that is involved in analyzing crop sustainability,
climatic conditions and to detect diseases and pests in the farm using machine learning
algorithms
• AI technology is also employed to analyze altered stress levels in plants. To provide valuable
results, the AI machine learning approach under stress (Gambhire and Shaikh Mohammad,
2020).
• Disease detection in crops using AI technology.
• Sensing of meteorological data using machine learning algorithms.
AI IN BIOMEDICINE
• Implementation of machine learning (ML) algorithms in biomedicine help in disease
diagnosis, biomedical data processing, analyzing the function of living organisms, and
biomedical research.

• In disease diagnosis integration of machine learning algorithms, Point-of-Care Testing

(POCT), and biochips can easily spot cardiac ailments at a very early stage (Rong et al., 2020).
• AI can also be used to analyze the survival rate of cancer patients.

• In biomedical imaging, AI technology can be implemented with electroencephalography

(EEG) for an exact prediction of the epileptic seizure (Usman and Fong, 2017).

• AI technology is adapted to process huge complex clinical data, by incorporating logical

reasoning and providing valuable conclusion within a short period.
AI IN DNA EXPRESSION PROFILING
• Microarray technology is extensively used for studying the expression pattern of multiple genes
concurrently.
• Series of experimental conditions enabled us to gather a huge amount of gene expression data at a varied
time point.

• it is atough challenge for the researchers to scientifically interpret the gene expression data, owing to their
robustness and huge data sets.

• By using computational tools, it is possible to convert the DNA expression profiles into theoretical data.

• Like clustering algorithms that display metabolic regulation of certain genes, concept learning is also vital
for analyzing the sequential expression pattern of genes.

• The use of AI technology in gene expression profiling displays the progressive stages of the microarray data
and it also defines the actual behavior of the model.

• It also provides a license for the users to identify the variants that naturally exist in the model via statistical
analysis (Dubitzky and Azuaje, 2004).
AI FOR IDENTIFYING EXONIC REGIONS
• In the era of genomics, the development of high-throughput techniques like Next-
Generation Sequencing (NGS) has fastened the sequencing of a whole genome in ashort
period.
• At present, AI with machine learning has been implemented to recognize the exonic
regions in a genome.
• the application of machine learning algorithms in gene prediction tools would be more
complex than traditional homology-based predictions.
• Presently “MethBank” database is used for wholegenome sequencing that affords
configurable and interactive data analysis.
• It is a user-friendly database that works well on diverse computer platforms like Java,
Linux, and MySql (Zou et al., 2015).
• Exonic regions are segments of a gene that are found in the final mature
messenger RNA (mRNA) molecule and are conserved during RNA splicing,
where non-coding regions called introns are removed. These regions contain
the nucleotides that code for proteins and also include untranslated regions
(UTRs). The collective set of all exons in a genome is known as the exome.
• Whole exome sequencing (WES) is a genetic test that analyzes the protein-
coding regions (exons) of an individual's DNA to identify genetic variations
associated with diseases
• A missense variant is a change in the DNA sequence that results in a different
amino acid being incorporated into a protein.

Phylogenetic refers to the study of evolutionary relationships among different biological entities, like species, individuals,
or genes, through a process called phylogenetic inference

A deleterious missense mutation is a single DNA change that results in a different amino acid being incorporated into a
protein, leading to a non-functional or harmful protein that can cause or increase the risk of a genetic disorder. Unlike
benign

Mendelian diseases are genetic conditions caused by a mutation in a single gene, following the inheritance patterns
described by Gregor Mendel.

Pairwise genome alignment compares two genomes to find homologous regions, while multiple genome alignment
extends this to three or more genomes, often using a phylogenetic tree to guide the process. Creation of gene trees uses
alignment data to infer evolutionary relationships between genes or entire genomes, reconstructing a history of
duplication, loss, and divergence over time.
AI IN IDENTIFYING VARIANTS/MUTATIONS FROM GENETIC DATA
• Genetic variants and various human diseases were detected using high-throughput genome
sequencing technology like Whole-Exome Sequencing (WES).

• In humans, missense variants cause genetic disorders like cancer, Mendelian diseases, and
other unidentified ailments.

• Experimental validation of these genetic diseases is quite complex, which is laborious and needs
huge resources, while this problem can be solved using computational approaches like analysis
of sequence similarity, phylogenetic relationship, and 3D protein structural homology.

• VarCards is a database that holds all the potential information associated with exonic mutants
and the classification of human genetic variants with consequences of allele frequencies.

• studied by Receiver Operating Characteristic (ROC) curves that include diverse factors like
precision, sensitivity, and Area under the Curve (AUC).
AI IN IDENTIFYING VARIANTS/MUTATIONS FROM GENETIC DATA
• The cutoff values are utilized to detect the deleterious missense mutants that were
observed from the ANNOVAR and the dbNSFB database.

• CNN (Convolutional Neural Network) is a machine learning technique that is commonly

used to detect genetic mutants.

• Campagne and Torracina software use the CNN method to analyze genomic data and
identify indel’s (Nagarajan et al., 2019).

• Genomic data were categorized into three groups as training data (60%), modeltesting
data (30%), and model validation data (10%).

• The deep variant method is used for detecting SNPs while the mutation in a gene is
identified using deep sequence software.
• Steps involved in predicting genetic disorders via computational methods.
AI WORKFLOW METHOD FOR GENOMIC ANALYSIS
• Ensemble project updates comparative genomics resources every year and generates
alignments and protein homology predictions.

• eHive, an AI tool to maintain comparative genomics, is built on blackboard systems,

network distribution, graphs, and block branch diagrams.

• MySQL database functions as a core processing unit while Perl script inquiries the
system and performs the job.

• It works on three main steps

(1) pairwise whole-genome alignments,
(2) multiple whole-genome alignments,
(3) creation of gene trees.
AI WORKFLOW METHOD FOR GENOMIC ANALYSIS
• the pairwise alignment pipeline consists of the following steps for comparing human and pika genomes.
• The first step involves the alignment of raw data;
• unit Chunk and to collect information on each genome using Group DNA.
• The other two programs Create Pair Aligner Jobs and Create Filter Duplicate Jobs using BlastZ, Filter Duplicates,and Target
Filter Duplicates run for the input DNA query.
• The BlastZ analysis runs the entire BLAST program (Altschul et al., 1990).
• The variations in the chunk sequences are removed via Query Filter Duplicate and Target Filter Duplicate analysis.
• The efficient query regions were identified using the updated Max Alignment Length module in a MySQL database.
• the raw data was connected and netted using the pairwise sequence alignment method.
• In the axtChain program, the input data are formatted using Dump Large Nib for Chains module.
• The Create Alignment Chain Job generates one Alignment Chain job for each genomic region.
• The netting is also executed using the above-mentioned chain alignment strategy.
• Latter, the sanity test analysis with the input query sequence was performed using Pairwise Health Check analysis.
• The third and final step includes the alignment of human and pika genomes (Severin et al., 2020).
AI IN PHYLOGENY
• A phylogeny is a branching tree model that confines the evolutionary origin of any
biological entity.
• Clade and clade length signifies the evolutionary time taken for a precise evolution.
• Multiple steps implicated in the construction of an evolutionary .
• The distance between each sequence is calculate using the Distance Matrix method.
• It works by building a correlation matrix in two comparable modes, i.e., Unweighted Pair
Group Method with Arithmetic Mean (UPGMA) and Neighbor Joining (NJ) method.
• A recent development in computational biology has provided the base for automated
analysis and tree building using AI algorithms.
• AI algorithms for a standard tree construction are k-nearest neighbor, Bayes theorem,
decision trees, and neural network.
• Steps involved in the construction of an evolutionary tree.
AI IN HEALTHCARE
• AI technology along with the machine learning approach is widely used in healthcare centers to
satisfy the patients, predict the correct diseases, and to improve the patient's health via
personalized drugs.
• AI in healthcare can be used to predict the functions of various organs such a urinary
bladder volume, and stroke.
• It is also being used to accelerate drug discovery methods and supply in a cost-effective
• manner (Rong et al., 2020).
• AI plays a vital role in the expert system (ES) that mimics the decision-making ability of a human
expert (Patra et al.,2010).
• ES aims in solving complicated biological problems with the usage of IFTHEN rules rather than
using the traditional coding method.
• ES is widely used in diverse areas of medicine like MYCIN, CADUCEUS, eye, hypertension,
pregnancy, leukemia, and several other human diseases to provide the solution for human
health.
• The MYCIN system was used for the diagnosis of blood-clotting disease, while CADUCEUS
was developed to diagnose thousands of diseases (Oluwafemi et al., 2014).
AI IN HEALTHCARE
• Architecture of the expert system.
AI TECHNOLOGIES IN SYSTEMS BIOLOGY TOWARDS PHARMACOGENOMICS
• In pharmacogenomics, AI technology plays a vital role in converting the accumulated raw data into a
readable format.
• pharmaceutical and nutraceutical industries have started using systems biology and machine learning
algorithms to determine the molecular mechanism of the drug thereby reducing the time and cost of the
experiment.
• The development of a novel drug is generally based on clinical studies, electronic medical imaging,
medical records, and DNA expression profiling.
• For instance, the application of computational tools and machine learning techniques in
pharmacogenomics helps in the discovery of a precise drug for cancer treatment with the help of patient
details, drug resistance, NGS data, and gene expression profiling.
• Accurate and efficient drug molecules are being developed by a virtual screening approach.
• Machine learning methods and AI technology have been functionalized in diverse stages of virtual
screening.
• There are two stages of virtual screening:
• (1) ligand-structure based virtual screening and (Ligand-based virtual screening methods use the
information present in known active ligands rather than the structure of a target protein for both lead
identification and optimization)
• (2) ligand-receptor based binding. (the specific, high-affinity interaction between a ligand (a signaling
molecule) and a receptor (a target molecule), which triggers a cellular response. )
AI TECHNOLOGIES IN SYSTEMS BIOLOGY TOWARDS PHARMACOGENOMICS
• Potential AI algorithm generates ligand-structure based virtual screening methods, which
are based on non-parametric scoring functions.
• To identify the exact ligand site, various nonpredetermined AI-based scoring functions
are utilized (RF-score, ID score, and NN score).
• Intending to improve the performance of scoring functions in AI technologyuses four
algorithms: (1) SVM, (2) Bayesian, (3) RF, and (4) DNN methods
• Nagarajan et al. (2019) detected the mutants and identified a precise drug target using a
computational approach
CASE STUDIES ON AI IN SYSTEMS BIOLOGY
AI TECHNOLOGIES IN SYSTEMS BIOLOGY TOWARDS PHARMACOGENOMICS
AI IN SYSTEMS BIOLOGY FOR CANCER CURE
The incorporation of AI technology in cancer imaging would produce abundant scientific
records, which would help the oncologist to speed up the treatment process for life-
threatening cancer diseases (Kantarjian and Yu, 2015).
In cancer imaging, AI is used to discriminate against normal cells and metastatic cancer
cells.
AI in medicine brings a societal revolution in healthcare centers by analyzing the key
regulatory genes involved in lung, breast, and thyroid cancer.
The Predictive Modeling for Pre-Clinical Screening mainly focuses on developing novel
drugs to treat cancer patients at the pilot-scale.
AI technology is also implemented for the discovery of cancer cells and lymph node
metastasis in women breast cancers.
With the aid of suitable computational models, AI in oncology uses Ordinary Least Square
methodology for medical imaging.
AI IN SYSTEMS BIOLOGY FOR CANCER CURE
In silico machine learning approach offers a better vision into the omics concept, next-
generation sequencing, molecular dynamics simulation, imaging, and clinical data.
there are certain limits in designing a machine learning approach for cancer,
(1) to access, share, label, and integrate multi-modal data’s of diverse types of cancer;
(2) to develop AI models for cancer research that can be scaled up using next-generation
processors;
(3) to assess complexity and consistency in the AI models.

• identification of a desired lung cancer type is either by mutations of epidermal growth factor
receptor and anaplastic lymphoma receptor tyrosine kinase rearrangements or via
sequential mutations of T790 is a serious process in the diagnostic treatment of critical
cancer patients with gefitinib, erlotinib, nivolumab, and atezolizumab
AI IN SYSTEMS BIOLOGY FOR CANCER CURE
• Convolutional Neural Network (CNN) technology is utilized for differentiating the type of
cancer like breast, bladder, and lung via complex visual recognition.
• helps the oncologist to detect the type of cancer and provide suitable treatment for the sufferers.
• During laboratory screening, the performance levels of cancer cells were detected using
computational modeling.
• Depending upon the lung pattern of CT scan (Computerized tomography), CNN's technology
can be applied.
• ImageNet developed by the GoogleNet architecture that would increase the robustness of the
biological systems thereby decoding and improving their dissimilar functions.
• Coudray et al. (2018) observed that DL algorithms can be utilized for automated examination
of tumor slides with the help of The Cancer Genome Atlas (TCGA) database.
• novel AI technology can help pathologists to detect precise gene mutation or cancer subtypes
quickly.
APPLICATIONS OF AI FOR COVID-19 PANDEMIC
• Preventive medicine requires the provision of novel technologies like artificial
intelligence(AI), machine learning (ML) approach, Big Data, and the Internet of Things
(IoT) to combat today’s emerging deadly diseases (Haleem et al., 2020).
• COVID-19 (Coronavirus) is one such pandemic disease that has a threat to the world.
• The medical industry is now looking for new technologies that can be used to monitor
and control the spread of the COVID-19 pandemic.
• AI is one such technology that can be implemented to detect the spread of coronavirus,
mortality rate, diagnosis, monitoring and it is also useful in controlling the infection (Huet
al., 2020).
• AI technology can boost the design strategy, medications, and testified data outcomes of
the COVID-19 patient.
FUTURE CHALLENGES IN ARTIFICIAL INTELLIGENCE
• Artificial intelligence uses computing machines that accomplish the tasks
intelligently and expressively.
• It succeeds in driving robots, filtering mails, medical diagnosing, predicting the
weather, and playing a game, etc. (Crevier, 2002).
• Practitioners and researchers of AI have claimed that machines are working
ashumans as they can think, feel the response, and have emotions.
• It accepts data,makes some process, stores, manipulates, and returns the
information.
• it is a great competitor to the human brain (Thomson, 1999).
• Machines are the simulation of humans where they can talk, walk, and some
operations.
• Artificial Intelligence of any device, computer, robot, or machine that cannot
substitute the human and it is preprogrammed (Crevier, 2002).
• AI has a great challenge in understanding the human brain and mimics it.
FUTURE CHALLENGES IN ARTIFICIAL INTELLIGENCE
• AI researchers face a lot of encounters in mimicking the human brain and turning it into devices
(Tatchell and Bennett, 2009).
• Artificial Intelligence research is implemented in an expert system, robotics, cybercrime,
agriculture, and biomedicine.
• The expert system is a program that mimics the human in a specific area.
• MYCIN, an example of an expert system, diagnoses the blood infection.
• The first expert system was designed in 1970 (Dennet, 2008).
• Robotics is another application of AI, which is extensively used in science and technology in the
manufacture, application, and design (Hunt and Shelley 2009).
• “Artificially Intelligent Robots can mimic humans in real-life (Perez et al., 2018).
• A perfect robot has certain abilities like selfevolve, self-diagnose, filtering mails, medical self-
repair, sensing, perception, and reasoning, etc.”
• It is growing with the understanding of the human mind and works in two dimensions
(1) It understands the human brain working process and
(2) it mimics the human mind artificially.
Recommended Study
• Gefitinib and erlotinib are targeted therapies. They directly interfere with
specific molecular pathways (EGFR) that drive cancer cell growth and are
only effective in tumors with specific mutations.
• Nivolumab and atezolizumab are immunotherapies. They work by
unleashing the patient's own immune system to fight cancer cells by
blocking the inhibitory signals sent by tumors.
• sigmoid function:A sigmoid function is a mathematical function that
produces an "S" shaped curve, mapping any input value to an output
between 0 and 1
• a step function (also called as staircase function) is defined as a piecewise
constant function, that has only a finite number of pieces.
• A linear function is a function that represents a straight line on the
coordinate plane.
• A ramp function is a mathematical and signal processing function that is
zero for negative input values and increases linearly (like a ramp) for non-
negative input values.
• The hyperbolic tangent (tanh) is a mathematical function that maps any
input value to an output between -1 and 1, forming an "S"-shaped curve.