SAVITRIBAI PHULE PUNE UNIVERSITY

A SEMINAR REPORT ON

Biomimicked Artificial Intelligence

By

Name : Nikhil Agarwal Roll No: 17123

Under The Guidance of

Prof. Gaurav Gupta

DEPARTMENT OF COMPUTER ENGINEERING

Dr. D. Y. Patil Institute of Engineering, Management &
Research

Akurdi, Pune.

DYPIEMR, Department of Computer Engineering 2018-19 Page 1

Dr. D. Y. Patil Institute of Engineering, Management & Research
DEPARTMENT OF COMPUTER
ENGINEERING

CERTIFICATE
This is to certify that the
Seminar Entitled

Biomimicked Artificial Intelligence

Submitted by

Name: Nikhil Agarwal Roll No: 23

is a bonafide work carried out by Students under the

supervision of Prof. Gaurav Gupta and it is submitted
towards the partial fulfillment of the re-quirement of
Bachelor of Engineering (Computer Engineering).

Prof. Gaurav Gupta Prof. P.P. Shevatekar

Internal Guide H.O.D

Dept. of Computer Engg. Dept. of Computer Engg.

Dr. A.V.Patil

Principal

Dr. D.Y.Patil Institute of Engineering, Management & Research

DYPIEMR, Department of Computer Engineering 2018-19 Page 2

 Abstract

Mother nature has always proved to be the inspiration for human race to find
solutions to existent problems. From systems that are modelled on the biological neuronal
networks to those that mimic genetic evolutions, this relentless quest for gaining insights
into the techniques of controlling, regulating and adapting delicate systems has continued
over time. Biomimicry has thus risen as a buzzword for defining solutions to real life
problems by creating artificial systems that mimic the functioning of elements from
Mother Nature. And giving a self-brain to the artificial systems, termed as Artificial
Intelligence, is a natural successor of the same.

The Darwinian Evolution theory simply defines the theory of “survival of the
fittest” and 4.5 billion years since the creation of planer Earth, only the fittest could survive
its conditional challenges. The main motive of this paper is to make a sustainable model
which could survive in this fast changing world based on the principles of nature and
advancement of technology.

Keywords:

Artificial Intelligence, Biomimicry, Fuzzy Modelling, Biological Entities, Technological

Advancements.

DYPIEMR, Department of Computer Engineering 2018-19 Page 3

 Acknowledgments

It gives us great pleasure in presenting the preliminary SEMINAR report on

“Biomimicked Artificial Intelligence”.

I would like to take this opportunity to thank my internal guide Prof. Gaurav
Gupta for giving me all the help and guidance I needed. I am really grateful to
them for their kind support. Their valuable suggestions were very helpful.

I am also grateful to Prof. P. P. Shevatekar, Head of Computer Engineering

Department, Dr. D. Y. Patil Institute of Engineering, Management & Research for
his indispensable support, suggestions.

Nikhil Agarwal
T.E. Computer Engg

DYPIEMR, Department of Computer Engineering 2018-19 Page 4

 Contents

 Chapter 1 - Introduction

 Chapter 2 - Literature Survey

 Chapter 3 - Architecture Design

 Chapter 4 - Use Cases

 Chapter 5 - Applications

 Chapter 6 - Discussions and Conclusions

 Bibliography/References

 Plagiarism Check report

 Seminar log sheet

DYPIEMR, Department of Computer Engineering 2018-19 Page 5

 CHAPTER 1

INTRODUCTION

1.1 Introduction to the Domain:

Biomimicry

Biomimicry, the method of designing artificial systems that mimic natural behaviour,
requires a reverse engineering process; the behaviour of a biological entity is analysed and
further mimicked by an artificial system. In many cases, biologists have already studied the
relevant behaviour and provided a detailed verbal description of it. It is in natural language.
Mimicking the natural behaviour can then be reduced to the following problem: how can
we convert the given natural language or verbal description into a well-defined algorithm
that can be implemented by an artificial system? This problem can be simply and efficiently
solved by the use of Natural Language Processing or Fuzzy Modelling.

1.2 Artificial intelligence:

 Artificial intelligence (AI) is the ability of a computer program or a machine to think
and learn. It is also a field of study which tries to make computers "smart". They work
on their own without being encoded with commands.
 In general use, the term "artificial intelligence" means a machine which mimics
human cognition. At least some of the things we associate with other minds, such
as learning and problem solving can be done by computers, though not in the same way
as we do.
 When most people hear the term artificial intelligence, the first thing they usually
think of is robots. That's because big-budget films and novels weave stories about
human-like machines that wreak havoc on Earth. But nothing could be further from the
truth.
 Artificial intelligence is based on the principle that human intelligence can be defined
in a way that a machine can easily mimic it and execute tasks, from the simplest to
those that are even more complex. The goals of artificial intelligence include learning,
reasoning, and perception.

DYPIEMR, Department of Computer Engineering 2018-19 Page 6

  As technology advances, previous benchmarks that defined artificial intelligence
become outdated. For example, machines that calculate basic functions or recognize
text through optimal character recognition are no longer considered to embody artificial
intelligence, since this function is now taken for granted as an inherent computer
function.
 AI is continuously evolving to benefit many different industries. Machines are wired
using a cross-disciplinary approach based in mathematics, computer science,
linguistics, psychology, and more.

A typical AI analyses its environment and takes actions that maximize its chance of
success. An AI's intended utility function (or goal) can be simple ("1 if the AI wins a game
of Go, 0 otherwise") or complex ("Do mathematically similar actions to the ones succeeded in
the past"). Goals can be explicitly defined or induced. If the AI is programmed for
"reinforcement learning", goals can be implicitly induced by rewarding some types of
behaviour or punishing others. Alternatively, an evolutionary system can induce goals by using
a "fitness function" to mutate and preferentially replicate high-scoring AI systems, similar to
how animals evolved to innately desire certain goals such as finding food. Some AI systems,
such as nearest-neighbour, instead of reason by analogy, these systems are not generally given
goals, except to the degree that goals are implicit in their training data. Such systems can still
be benchmarked if the non-goal system is framed as a system whose "goal" is to successfully
accomplish its narrow classification task.

AI often revolves around the use of algorithms. An algorithm is a set of unambiguous

instructions that a mechanical computer can execute. A complex algorithm is often built on top
of other, simpler, algorithms. A simple example of an algorithm is the following (optimal for
first player) recipe for play at tic-tac-toe.

1. If someone has a "threat" (that is, two in a row), take the remaining square. Otherwise,
2. if a move "forks" to create two threats at once, play that move. Otherwise,
3. take the center square if it is free. Otherwise,
4. if your opponent has played in a corner, take the opposite corner. Otherwise,
5. take an empty corner if one exists. Otherwise,
6. take any empty square.

DYPIEMR, Department of Computer Engineering 2018-19 Page 7

Many AI algorithms are capable of learning from data; they can enhance themselves by
learning new heuristics (strategies, or "rules of thumb", that have worked well in the past), or
can themselves write other algorithms. Some of the "learners" described below, including
Bayesian networks, decision trees, and nearest-neighbour, could theoretically, (given infinite
data, time, and memory) learn to approximate any function, including which combination of
mathematical functions would best describe the world[citation needed]. These learners could
therefore, derive all possible knowledge, by considering every possible hypothesis and
matching them against the data. In practice, it is almost never possible to consider every
possibility, because of the phenomenon of "combinatorial explosion", where the amount of
time needed to solve a problem grows exponentially. Much of AI research involves figuring
out how to identify and avoid considering broad range of possibilities that are unlikely to be
beneficial. For example, when viewing a map and looking for the shortest driving route
from Denver to New York in the East, one can in most cases skip looking at any path
through San Francisco or other areas far to the west; thus, an AI wielding a pathfinding
algorithm like A* can avoid the combinatorial explosion that would ensue if every possible
route had to be ponderously considered in turn.

1.2.1 Salient Features of AI:

Some of common and known characteristics of AI are as follows:

 AI problem has large solution set

Now, we humans are facing a lot of problems like environmental problems, technical problems,
socio-economic problem

 Can manipulate large number of symbolic information to a greater extend

 Learning Ability

 Cope with situational problems and explore solutions at same time

 Solving AI problem with or without AI technique.

DYPIEMR, Department of Computer Engineering 2018-19 Page 8

Some other characterisitics are:

 Rational Thinking

 Fast Decision Making

 Plan Creation and Excecution

Fig: Artificial Intelligence

1.2.2 Applications of AI:

The majority of the industries are already started investing in AI and are able to reap the
maximum out of it. AI is transforming the day to day life of the common man along with the
business world. Now let’s see the add on's in 2019 that has made the AI functioning better
compared to 2018.

 Facial recognition
 Artificial Neural networks
 Deep Learning

DYPIEMR, Department of Computer Engineering 2018-19 Page 9

  Quantum Computing
 Chatbots
Nowadays we are using AI frameworks to communicate with our telephones through Alexa
and Siri. cars like Tesla’s decipher and break down their surroundings to insightfully drive
themselves. Even Google chooses what sort of query items to give us depending on who it
thinks we are. Artificial intelligence is here yet this stage just starts on the grounds that later
on AI is going to make a huge difference.

1.3 Fuzzy Modelling:

Now, a big problem arising in computation of biological entities is that we have the
behavioural structure of the natural elements in verbal format but, the computer system does
not recognise such a structure. It requires a defined mathematical equation to implement the
required command. Fuzzy modelling (FM), with its ability to handle and manipulate verbal
information, constitutes a natural approach for addressing this problem. FM may lead to
systematic approach to a verbal description of an animal’s behaviour (e.g., the foraging
behaviour of ants), and obtain a mathematical model of the said behaviour which can be
implemented to create efficient artificial systems. Recently, FM was applied successfully to
transform verbal descriptions of animal behaviour into well-defined mathematical models.

The method consists of four steps:

(1) Identifying the state-variables.

(2) Restating the given verbal descriptions as an FRB relating these variables.

(3) Defining the fuzzy terms using suitable membership functions.

(4) Inferring the FRB to obtain a well-defined mathematical model.

This approach was used to derive mathematical models for:

(1) The territorial behaviour of fish.

(2) The orientation of a planarian to light.

DYPIEMR, Department of Computer Engineering 2018-19 Page 10

(3) The foraging behaviour of ants.

(4) The mechanisms regulating the population size in flies.

In all these examples, the starting point was a detailed verbal description of the natural
phenomenon. Using FM this was transformed into a well-defined mathematical model.
Simulations and rigorous analysis demonstrated the suitability of the mathematical model.

1.4 Reinforcement Learning:

Reinforcement learning (RL) is an area of machine learning concerned with

how software agents ought to take actions in an environment so as to maximize some notion of
cumulative reward. Reinforced Learning can be performed as a procedure which synthesizes
outputs based on inputs, step-by-step.

To define practical implementation of Reinforced Learning, lets us consider an

example. Suppose we want to train a model based on a specific species. The mathematical
equation set is ready by FM. Then we consider the actions of the species as an input to the
system. The system learns from the various actions of the species and trains to perform the
same actions. We decide that the model would be rewarding (positively or negatively) based
on the output actions on a certain given input. Suppose the activity is performed 10 times and
the model initially just gets 4 out of the 10 actions correct. Then the model is self-trained such
that it would perform more of the correct outputs and neglect the incorrect ones partially. This,
when performed over a period of time would eventually produce a larger number of correct
outputs as compared to the incorrect ones. Thus, the model is ready and is eligible to take
correct decisions by itself. We will see a practical example of the same in the sections ahead.

DYPIEMR, Department of Computer Engineering 2018-19 Page 11

 CHAPTER 2

LITERATURE SURVEY

2.1 Literature Review:

 Biomimicry is a revolutionary new science that analyzes nature's best ideas -- spider
silk and prairie grass, seashells and brain cells -- and adapts them for human use. The
products of biomimicry are things we can all use -- medicines, "smart" computers,
super-strong materials, profitable and earth-friendly business. Biomimicry eloquently
shows that the answers are all around us.
 Micheal Pawlyn discusses states in his book on biomimicry in architehture
“biomimicry-design inspired by the way functional challenges have been solved in
biology – is one of the best sources of solutions that will allow us to create a positive
future and make shift from the industrial age to the egological age of humankind”
 The language of nature is embedded in the way we speak, name things, create
architectural constructions or develop new materials to thrive as specie. The ecological,
sustainable or the biologically-inspired design definitions are signs of a collective
intelligence. Searching new educational ways to rediscover and implement nature´s
strategies is helping us to redefine a new ethic for the designer, a NatureCentered one.
 “Everything we love about civilization is a product of intelligence, so amplifying
our human intelligence with artificial intelligence has the potential of helping
civilization flourish like never before – as long as we manage to keep the
technology beneficial.” - Max Tegmark, President of the Future of Life Institute
(Davey, et. al, 2018).

 Biologically inspired design (also known as biomimicry, biomimetics and bionics) is a

growing movement in modern design that espouses the use of nature as an analogue
for designing technological systems and as a standard for evaluating technological
designs (French 1994; Gleich et. al. 2010; Turner 2007; Vincent & Mann 2002; Vogel
2000). This paradigm has inspired many designers in the history of design, such as
Leonardo da Vinci, the Wright brothers, etc. However, it is only over the last generation

DYPIEMR, Department of Computer Engineering 2018-19 Page 12

 that the paradigm has become a movement, pulled in part by the growing need for
environmentally sustainable development and pushed partly by the desire for creativity
and innovation in design.

DYPIEMR, Department of Computer Engineering 2018-19 Page 13

 CHAPTER 3

ARCHITECTURE DESIGN

3.1 Architecture of Biomimicry:

Architecture of Biomimicry is highly advanced, it has a typical two way approach which
provides diversity in the functioning. Although the end motive of both approaches are the
same i.e. a design solution for a problem.

The two approaches considered while taking biomimicry as a solution are

 Top-Down Approach
 Bottom-Up Approach

Figure 1: Architecture of 5G
The two approached sound similar to the analysis approach to coding languagues but
are rather diverse in their working.

TOP-DOWN APPROACH

In the top-down approach the most important feature is the problem. We have aa
problem at hand which needs to be resolved. Experts in the domain of biomimicry are assigned
to look for solution to the given problem.

 Design Problem – The design problem is the core problem at hand. This is the issue
which we are going to solve by mimicking a particular aspect of a specific entity from
the biological system.

DYPIEMR, Department of Computer Engineering 2018-19 Page 14

  Search for biological analogies – Once the problem is defined, similarities for solution
in biological entities is research. Natural aspects have the same/relatable problem are
studied closely to find analogies.
 Identification of appropriate principles – Once the analogy is identified, the next step
is study of principles from the analogy. Identify the principles which match with the
requirements of the problem and use them.
 Abstraction, detachment from biological model – Once the principles and design
analogies are established, the model is abstracted from the real world and is detached
from the actual biological entity from which it was derived.
 Testing, analysis and feedback – Once the model is ready, it is put up on testing
grounds to check its working in the real world. The functioning is analysed and
feedback is provided to make changes if any.
 Design Solution – Once tested, the model is ready and thus can be put to use to solve
the problem. This is the designed solution for the given problem.

BOTTOM-UP APPROACH

In the bottom up approach, the major difference is that we do not have a problem at
hand. Instead we research the biological entities and find models to go hand-in-hand with
technology.

 Biological research – This is the initiation level. Reseach is done on biological entities
to study their characteristic distinguishing features.
 Biomechanics, functional morphology and anatomy – The mechanical properties of the
entity are analysed and a morph (replica) is made to match with technology. The
anatomy (internal structure) of the entity is also replicated fitting to the model.
 Understanding the principles – Once morphed, the fundamental principles of the
functionality is understood and planned on how to implement it in the model.
 Abstraction, detachment from biological model – Once the principles and design
analogies are established, the model is abstracted from the real world and is detached
from the actual biological entity from which it was derived.
 Technical Implementation – Once all the above steps are completed, the technical
implementation of the biological model happens.
 Design Solution - Once tested, the model is ready and thus can be put to use to solve
the problem. This is the designed solution for the given problem.

DYPIEMR, Department of Computer Engineering 2018-19 Page 15

 CHAPTER 4

CASE STUDIES

4.1 Shinkansen- The Japanese Bullet Train:

In 1989, Japan’s bullet train-Shinkansen, one of the fastest that time, had a problem.
Every time it exited a tunnel, it was really loud creating noises above 70dB. This was because
whenever the train sped into the tunnel, it pushed waves of atmospheric pressure through to
the other end which created a sonic boom. This was a problem in tense residential areas. So a
team of engineers was set up to find an appropriate solution to this problem. Eiji Nakatsu, an
engineer with JR West and a birdwatcher, used his knowledge of the splashless water entry of
kingfishers and silent flight if owls to decrease the sound generated by the trains. Kingfishers
move quickly from air (a low-resistance medium) to water (a high-resistance medium). The
kingfisher’s beak provided an almost ideal shape for suck impacts. As the train faced the same
challenge, moving from low drag open air to high drag air in the tunnel, Nakatsu was able to
relate it and designed the forefront of the Shinkansen train based on the peak of the kingfisher.
And what were the results? The more streamlined Shinkansen train not only travels more
quietly, but also 10% faster and saves 15% electricity.

Ref: Biomimicry: 9 Ways Engineers Have Been 'Inspired' by Nature

DYPIEMR, Department of Computer Engineering 2018-19 Page 16

4.2 Self-cleaning lotus leaves:

Lotus plants have a peculiar way of staying dirt-free and being highly water repellent.
This stands as an obvious advantage as the plant sustains is muddy habitat and clean leaves
would mean more photosynthesis. The leaves’ microscopic structure holds the answer to their
ultrahydrophobic properties.

If we zoom on a leaf, we can see wax crystals and folds in the surface, like tiny pyramids
or traffic cones pointing upwards. Droplets of water are thus only supported by the tips, with
plenty of air in between the tips. The lack of contact surface means the droplets don’t stick very
well. It takes only a light breeze to make them toll off the leaf, taking away dirt particles along
with them.

The study of the structure of these lotus leaves has inspired self-cleaning paints, glass,
textiles and many more materials.

Ref: https://bioneers.org/biomimicry-5-names-to-know/

DYPIEMR, Department of Computer Engineering 2018-19 Page 17

 CHAPTER 5

APPLICATIONS

5.1] Traffic Problem Solution:

Ever been stuck in a traffic jam which lasts for hours and eventually realizing that there
were no reasons whatsoever for it? Well the reason behind such jams is the lack of co-
ordination between drivers. According to a study, an average American driver spends
approximately 50 hrs/year in traffic jams. So what is the solution for this? Well, even seen ants
moving around for food? There is complete uniformity in their motion. There are never traffic
jams created in between them. This is because they motion in a defined motion wherein their
paths are organized and their motion is straight. When worker ants are moving if any random
ant is slow, the ant behind does not overtake it from sides and instead moves with the same
seed so as to work in unison. So the simple method to avoid traffic is to mimic this trait of ants
onto our vehicles. A biomimicry expert could simply explain the motion of the ants and specific
traits in the motion. A fuzzy modelling procedure could simply step up and convert these traits
and points of definition of motion and prepare an equation to define the same. This algorithm
can then be fed to a car’s system and trained accordingly. The concept of self-driving cars
comes into role here. We can use the advancement of AI in the field of automobile, i.e. Self-
driving cars, to implement this computational system. Therefore, automobiles, with a touch of
biomimicry of ants can be our future with no traffic signals.

Fig. Ant Synchronous Motion

DYPIEMR, Department of Computer Engineering 2018-19 Page 18

5.2] Mimicking Ecosystems:

We all have studied about the biological ecosystems. The speciality about such
ecosystems is the circular chaining. Every element of the chain is linked to a predecessor with
is consumed by it and a successor which consumes it. The best part about it is that after reach
the end of the chain, the end product is again transformed into the first product and the cycle
simply continues. So what is we learn this feature and produce an economical system wherein,
there are no by-products produced and the produced ones are again up cycled to be used again
to start the chain. This could completely change the whole economic structure. The harm
implied to nature by these by-products would be minimized or might be even nullified. Circular
economy system could be the future of the global economic system.

5.3] Biomimicked Robots

Robots have been a fascination for technological advancement researchers in the 21st century.
However, a new trend of mimicking biological entities’ behaviour to make robotic models of
the same has gained a lot of interest. The study of biological species and capturing their traits
which could be implemented in the benefit of a structure yields more output then its tradition
friend. A model as such has been already prepared which mimics the kangaroo’s technique to
hop around. The speciality of this model is that at every jump, it could generate energy on its
own which is used for the next jump. Thus, greatly reducing the energy required for the
commutation of the robot. Many such robot can be made which are inspired by mother nature
which can efficiently perform the defined tasks.

Figure 2: Application

DYPIEMR, Department of Computer Engineering 2018-19 Page 19

 CHAPTER 6

DISCUSSION AND CONLUSION

6.1 Discussion of Biomimicked AI:

AI independently has a lot of applications and uses. Although, combined with the
powers of mother nature, it turns out to become an inevitable force to reckon. Together, AI
and Biomimicry can turns the tides of the world to provide an highly technologically advanced
ecosystem with the ecological balance maintained intact.

The right use of technology as a boon along with the traditional natural elements would
become the future of robotics and would result in unimaginable results. Real steps in this
domain would build a greener, better and safer world.

6.2 Challenges:

Although biomimicry and artificial intelligence could possibly bring a revolution that
would bring about prosperity in the livelihood on the human race, it yet has a few concerns
which would be harming. A major issue would be that if we are replicating a specific species
from the environment, and we are able to produce 100% output out of the said species, there
may come a time wherein the real species would be overtaken by the artificial ones and this
would eventually lead to the extinction of the same.

Artificial Intelligence is the use of computation to perform human tasks and beyond,
but the fact is that these systems are sensitive to cyber security and a small malfunctioning
system would create great havoc.

DYPIEMR, Department of Computer Engineering 2018-19 Page 20

 6.3 Conclusion:

While researching on this topic, I realized that the people who are building the world
are amateurs at the biological aspect and thus an integrated work has been brought up to support
the current system with the addition of naturally present elements. If looked upon closely, we
can learn a lot from the surrounding and implement it on the virtual level only to implement it
later to perform the similar task without harming the elemental constituents of out existence.
This paper has been published in the view of promoting the hand-in-hand use of the natural
entities and the technological advancement for the betterment of this planet and specifically the
human race.

DYPIEMR, Department of Computer Engineering 2018-19 Page 21

 BIBLIOGRAPHY / REFERENCES

References:

 Biomimicry and Fuzzy Modeling: A Match Made in Heaven.

 https://en.wikipedia.org/wiki/Artificial_intelligence

 https://asknature.org

 https://biomimicrybe.org

https://en.wikipedia.org/wiki/Reinforcement_learning

 Biomimicry: Innovation Inspired by Nature - Janine M. Benyus

publish date is 2002-09-01

 Biomimicry in Architecture by Micheal Pawlyn

 Nature-Centered Design: Exploring the path to design as Nature.- by David Sánchez

Ruano - 2019

 Deep Learning with Biomimicry – Emily Mendez and Emily Marcum- February 2019

 Biologically Inspired Design: A New Paradigm for AI Research on Computational

Sustainability? – Ashok K. Goel -2015

DYPIEMR, Department of Computer Engineering 2018-19 Page 22

 PLAGIARISM CHECK REPORT

DYPIEMR, Department of Computer Engineering 2018-19 Page 23

DYPIEMR, Department of Computer Engineering 2018-19 Page 24