Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Proposal for Establishment of
Centre for Artificial Intelligence & Machine
Learning in Agriculture
(CAIML)
Taking a lead towards intelligent and customised solutions
for SMART agricultural practices in J & K
Directorate of Planning & Monitoring
SKUAST-Kashmir
Shalimar Srinagar- 190025 [email protected]
0194-2461493, 7780882991
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
1
Artificial intelligence is like the new electricity and it will transform every sector. NitiAayog CEO Amitabh Kant
NATIONAL STRATEGY FOR
ARTIFICIAL INTELLIGENCE 2018
NITI Aayog
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
2
Preamble
Jammu and Kashmir has missed the first three industrial revolutions, that have shaped the
economies and the prosperity of the societies around the world. The outcome is an economically
fragile and technologically backward state. The first industrial revolution witnessed the
emergence of machines and mechanization, a process that replaced agriculture with industry as the
foundations of the economic structure of society since the 2nd half of 18th century (1765). This
revolution used water and steam to mechanize production. The 2ndindustrial revolution from
1870s used electric and chemical energy to create mass production. The 3rd revolution from 1969
used nuclear energy for production, and electronics and information technology to automate the
production. Besides the biotechnology emerging as the core sector that transformed health and
agriculture. We unfortunately continue to remain as the consumer based society than a contributory
one, and therefore a fragile economy.
The 4thindustrial revolution is unfolding before our eyes. It would be largely driven by artificial
intelligence, machine learning, robotics, 3-D printing, automation and bio-economy. With
exponential expansion, it is characterized by merging technology that blurs the lines between the
physical, digital and biological spheres to completely disrupt the way we think, communicate, act;
and also the entire production, management and governance systems. Envisaging AI as the single
largest technology revolution of our live times, with the potential to disrupt almost all aspects of
human existence, NITI Aayog in its policy document on National Strategy for Artificial
Intelligence in 2018, has identified 5 core areas viz Education, Agriculture, Healthcare, Smart
Cities and Smart Mobility.
The limiting factors of the state of J & K that kept the first three revolutions off its shores, have
become irrelevant in the global village swapped by 4thindustrial revolution, which is rooted in a new
technological phenomenon—digitalization, that enables us to build a new virtual world from which
we can steer the physical world. The Euro 92 billion agri-economy of the Netherlands is driven by
this technology, where the machines are trained to communicate with plants (Tulip) and soil and
weather to predict the most accurate requirements for plant growth in terms of the nutrients and the
pest management, and thereafter linking the processes from farm to fork.
SKUAST-Kashmir can be a path breaker by being bold and disruptive in venturing first in AI and
ML in the state and in the entire NARES. The proposal on establishment of the Center for Artificial
Intelligence and Machine Learning in Agriculture (CAIML) is therefore a strategic move to infuse
higher order skills of AI & ML in students and researchers to assure that we catch the bus of 4th
industrial revolution in right time, and secure and empower our future generations by compensating
the losses of the earlier opportunities.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
3
Mandate:
1. R &D : Develop AI driven customized Decision support systems for farmers, processors, markets
2. HRD : Develop Next Gen Human Capital equipped with 4th Generation skills who shall drive the technology based agri-economy in J&K
3. Collaborations: Explore and Establish Partnerships with Organizations having propensity with AI,
ML and IoTs in Agriculture.
Taking a lead towards intelligent and
customized solutions for SMART
agricultural practices in J & K V I S I O N
Intelligent IoT based Farm management System for Sheep
Intelligent IoT based Fertigation System for Apple orchards
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
1
Prelude
The world population by 2050 is projected to increase by over 2.3 billion and only 4% additional
land is expected to come under cultivation by then (FAO, 2009). Therefore, there is a need to double
food production by 2050 by intensifying production on almost the same amount of land (FAO,
2018). This is an enormous challenge since nearly 124 million people across 51 countries already
face acute food shortage. Other issues like the global climate change and dwindling labour force also
add substantially to the problem which our conventional approaches are unable to handle. There is a
need for an out of the box solution which could eventually lead to the 2nd Agricultural Revolution.
The dawn of the 4th Industrial Revolution is already blurring the lines between the physical, digital
and biological spheres and this may prove to be the sine qua non of the next Agricultural Revolution
as well (World Economic Forum, 2016).
This collective revolution would largely be driven by Artificial Intelligence, Machine Learning, Big
Data Analytics, Heuristics, Fuzzy Logic Systems, Automation, Robotics and Bio-Economy, thus
leading to intensification of agriculture and smart interpretation of big data. Intelligent data driven
systems and digital fabrication technologies interacting with the biological world, would make it
possible to adopt new and innovative approaches to deliver food security, economic opportunity and
environmental sustainability.
Problem Statement
Jammu and Kashmir presents tremendous potential for agriculture and allied sectors. Despite this, the
economy of the State is currently import based. Traditional crops and indigenous breeds, the
products of which could fetch a high market price, are dying out slowly due to negligence and a lack
of technological intervention. Residential houses are rapidly erupting over farming areas and people
are moving out of agriculture in hoards. This is because agriculture no longer seems to be a viable
career choice. The agricultural land in the state has shrunk by more than 18,000 hectares in the last
two decades. The total agricultural area in 1998-99 was 878 thousand hectares, which is now 859.52
thousand hectares. To top it all, unprecedented weather patterns often wreak havoc and cause huge
loses to the State economy. Also, unemployment in the State is a major problem. The youth lack
exposure to latest ground breaking arenas of science and therefore graduates are often unemployable
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
2
under the current breakneck competitive environment. Therefore, it may be said that the two most
powerful resources for any society viz. its youth and its agriculture, are underperforming.
Global Trends in Agriculture Using AI and Machine Learning
The emergence of new age technologies like Artificial Intelligence (AI), Cloud Machine Learning,
Satellite Imagery and advanced analytics is creating a global ecosystem for Smart or Precision
Farming. Such cognitive solutions are benefiting agriculture in a myriad of ways. Data on historical
weather pattern, soil reports, new research, rainfall, pest infestation, images from Drones and
cameras etc. is sensed by Cognitive IOT solutions to provide strong insights to improve yield by in-
depth field analysis, crop monitoring, scanning of fields, managing adverse weather conditions, crop
health management, and yield management.Drones are now being used to produce a 3-D field map of
detailed terrain, drainage, soil viability and aerial spraying of pods with seeds and plant nutrients into
the soil provides necessary supplements for plants.
Proximity Sensing, Hyper-spectral Imaging, 3d Laser Scanning and Remote Sensing are two
technologies are finding their use in soil testing, soil characterization and for building crop metrics
across thousands of acres. Hardware solutions pairing data-collecting software with robotics can
automate irrigation and prepare best fertilizers, herbicides etc. in addition to other activities to
maximize output while optimizing input. Self-driving tractors and combine harvesters trundle round
fields without human involvement.
Computer vision technology, IOT and drone data can be combined to ensure rapid futuristic actions
by farmers. Pre-processing of images is being used in pest identification, nutrient deficiency
recognition and disease diagnosis. Images of different crops under white/UV-A light are captured to
determine how ripe the green fruits are. Farmers can create different levels of readiness based on the
crop/fruit category and add them into separate stacks before sending them to the market.
Based on multiple parameters, cognitive solutions make personalized recommendations to farmers
on the best choice of crops and hybrid seeds. This technology also helps to monitor crops along their
entire lifecycle including report generation in case of anomalies.
Neural network-backed computer vision, wearable devices, and predictive analytics algorithms are
helping scientists reimagine animal farming too. Animal identification and monitoring is now
possible using intelligent integral control systems. Tags, implantable transmitters and receivers for
low level signals from animals by telemetry are now commonly used (Puers et al., 1988). AI is also
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
3
making it possible to analyse animal behaviour for animal health management, pregnancy diagnosis,
heat identification and early disease diagnosis. Intelligent animal identification systems identifying
animals automatically from their sounds (Chesmore, 2001) have also been developed. Robots are
now being used for scientific feed management, auto drafting, auto floor cleaning, high-tech milking
and robots to diagnose disease and administer remedies/cures to the affected animal (Ishaq, 2018).
AI is also being used for the prediction of the genetic merit and futuristic projections of the animal’s
production potential (Ghazanfari et al, 2011). Machine learning algorithms are also actively being
used for gnomic selection of animals. Data-centric farming powered by AI algorithms collect and
analyses massive farm data, regarding humidity, temperature, and food requirements.
All these technologies are producing a giant leap in production and this may help farmers produce 70
percent more food by 2050 (Jones, 2017). It may thus be safely concluded that AI promises a food
revolution, from farm to supermarket.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
5
Primary Objectives of the Centre
• Make the study of AI and ML an integral part of agricultural education curriculum at all
levels.
• Creation of an education, research and extension hub at SKUAST-K to lay the foundation of
AI and ML programs and initiatives in the State.
• Attract top AI& ML education, research and entrepreneurial talent to the University for
improvement of Agriculture.
• Develop organic home-grown personalized AI solutions for temperate agriculture.
• Promote the use of AI& ML as part of an integrative solution for skill development,
upgradation of living-standards and agricultural transformation in the State and
consequently in the entire country.
Immediate/ Short Term Applications
The primary goal of the University is promotion of agriculture in all forms to improve food security
and transform lives. Therefore, the establishment of an Artificial Intelligence cum Machine Learning
Center would be strengthening all primary mandates of the University viz. education, research and
extension. It would be of both long and short term benefit. The immediate implications of the
AI&MLCenter would include the following:
• Provide support to all research activities at the University and increase technology transfer
ability towards major stakeholders (farmers, entrepreneurs etc.)
Keeping all this in view, the establishment of the Centre for Artificial
Intelligence (AI) & Machine Learning(ML) in Agriculture at SKUAST-
Kashmir (CAIML) is proposed, to build and sustain AI-based,
innovation, growth and productivity in Agriculture and Allied Sectors.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
6
• Collaboration with internationally renounced institutes (MIT, IISc and IITs, etc) and
industries (Microsoft, Google, Amazon etc) through MOUsthereby opening up new
opportunities to the enrolled students and youth in general.
• Development and testing of cutting-edge tools and technologies to coordinate and monitor the
work force activities across diverse verticals(keeping the unique conditions and challenges of
temperate agriculture in view).
• Creation of Early Warning Systems for weather changes, epidemicsetc.
• Use of Farm Management Information Systems for planning, capture of thousands of farm
variables and recording of the workforce activities across diverse fields
• Creation of separate, centralized databases for crops, fruits and animals for the State.
• Use Automatic Learning Systems like Artificial Neural Networks for data analysis for
personalized farm projections and advisories.
• Modification of available technologiesto better suit the temperate agro climatic conditions.
• Enhance production per plant though micro-climatic data, soil condition data and real-time
controls for irrigation, seed behavior to make it conducive for a better crop growth for sustainable
agriculture.
• Upgrade traditional farming equipment (tractors, tillers, milking machines etc.) for
automatizing mundane tasks while improving precision and minimizing losses.
• Utilize smart sensors and software to monitor diseases and pest and also to track and trace a
perishable’s condition for supervision of the supply chain to minimize spoilage.
• Create an online intelligent, integrated demand-supply based market place with customized
projections regarding the rise or fall of demand in the future
• Accurate identification of animals and tracking and monitoringof feed intake, growth,
pregnancy diagnosis, heath through readily available devices and sensors like RFID tags,
pedometers, smart cameras etc.
• Automatic cloud-based data capture and analysis through automatic milking machines, smart
on-body devices, automatic weighing and drafting systems for animal farms.
• Providing better farming conditions and controls through the application of Internet of Things
(IoT) in animal farms for temperature control, humidity condition sensors, automated feeders,
waters and automated milking machines (cows) in animal and poultry houses.
• Integrate the breeding process with modern genomic tools to organize the hybridization of
desirable characteristics in the crop. Utilize bioinformatics and AI for making superior breeding
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
7
decisions both for animals and plants. Coordinate plant and animal breeding process from pre-
breeding to commercialization phase.
• Boost conservation efforts for domestic and wildlife species especially Hangul (the State
Animal) through aerial monitoring systems, real-time sensor technologies etc.
• Creation of an innovation driven ecosystem by the provision of state of the art IoT labs and
facilities for leveraging start-ups, enterprises, institution and individuals for testing and
developing the technology solutions.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
8
IOT/ AI in Course Curriculum
Program Courses to be Offered Duration
UG Courses
(for BSc/ BVSc & AH/
BFSc/ BTech)
• Update the Course on “ Biostatistics and
Computer Science 3+1” with chapters on AI
and ML, The revised structure could be: o Elementary Statistics.
o Concepts and Techniques of Data Science /
Data Mining
o Introduction to AI and ML techniques
1 Semester
Elective and Certificate
Courses
(for UG, PG, PhD
Outside students)
• Data Science in Agriculture
• IoT and Cloud Computing
• Crop and animal Modelling
• Automation and robotics in agri-operations
• Machine Learning
• Block Chain Technology
1 month, 2 month,
and 3 months
PG Course
(forMSc/ MVSc/ MFSc/
MTech)
Machine Learning and Artificial Intelligence and
its Applications in Agriculture 1 Semester
Diploma (AI and
Machine Learning)
Post Graduate Diploma in Machine Learning and
Artificial Intelligence in Agriculture 1 Year
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
9
Course Syllabus
Data Science Tool Kit Python: Basics Data Types, Conditional Statements, Looping, Control
Statements,String, List and Dictionary Manipulations, Python
Functions, Modules and Packages, Object Oriented Programming in
Python,Regular Expressions, Exception Handling.
Introduction to Database Management System & SQL, Database
Interaction in Python.
Data Analysis & visualization – using numpy, matplotlib, scipy
R Programming: Basics - Vectors, Factors, Lists, Matrices, Arrays, Data
Frames, Reading data.
Data visualization - barplot, pie, scatterplot, histogram, scatter matrix
Statistics and
Exploratory Data
Analytics
Analytics problem solving, Inferential statistics, Hypothesis testing,
Exploratory data analysis, Case study
Machine Learning Supervised and Unsupervised Learning, Classification and Regression,
Linear Regression, KNN, K Means, Logistic Regression, Support
Vector Machines (SVM), Decision Tree, Naïve Bayes, Ensemble
Methods, Random Forest, Boosting and Optimization
Natural Language
Processing
Basics of text processing, Lexical processing, Syntax and Semantics,
Other problems in text analytics
Deep Learning Deep Learning Concepts, Basics of Artificial Neural Network, Deep
Neural Networks, Convolutional Neural Network (CNN), Recurrent
Neural Network (RNN), Tensorflow, Keras, Introduction to Generative
Adversarial Networks(GAN),OpenCV
Reinforcement
Learning
Introduction to AI/Cognitive platforms and Understand the basics of
Reinforcement Learning and its applications in AI
Graphical Model Introduction to Bayesian Methods, Graphical Methods, Learning and
Inference
Computational Biology sequencing and assembling of the genomes of the niche crops like
saffron, kala zera, Mushkbudgi, Red rice varieties, Ambri apple, Sea
Buck thorn, Apricot, Pashmina goat, Bakerwal goat, local poultry, and
end number of the medicinal and aromatic plants. This shall include
gene expression and regulation •DNA, RNA, and protein sequence,
structure, and interactions • molecular evolution • protein design •
network and systems biology • cell and tissue form and function •
disease gene mapping • machine learning • quantitative and analytical
modeling
Mini Project
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
10
Staff Requirements
Area Faculty Position Expertise
Machine Learning
(Machine learning (ML) is a category of
algorithm that allows software
applications to become more accurate in
predicting outcomes without being
explicitly programmed)
Application: yield prediction, disease
detection, weed detection crop quality,
and species recognition; animal welfare
and livestock production)
Assistant Professor
(Computer Science)
Expertise in the following domains and
teach the relevant courses in the same.
a) Machine Learning
b)Computer Vision
c)Natural Language Processing
d) Deep Learning
e) Reinforcement Learning
Expertise in following tools/ languages
a) Programming Languages like C,
C++, Python, JAVA
b) Tools like MATLAB/ OCTAVE/
PYTORCH/ TENSOR FLOW
Data Optimization / Deep Learning
(Data Optimization is a process that
prepares the logical schema from the data
view schema. involves maximizing the
speed and efficiency with which data is
retrieved)
Assistant Professor
(Computer Science)
Expertise in the following domains.
a) Linear Optimization, Convex
Optimization, Integer Optimization
b) Optimization problems and Algorithms
c) Improving Deep Neural Networks
d)Multi objective Optimization Problems
and Algorithms.
e) Bayesian Statistics
Further expected to teach the
mathematical courses like:
a)Statistics and Exploratory Data
Analytics
b) Applied Mathematics
c) Operational Research etc.
Expertise on implementing the algorithms
in the following tools:
MATLAB/ OCTAVE
Lab Courses like
Optimization with Metaheuristics in
Python, Improving algorithms and
implementation.
IoT
(Internet of Things (IoT) is an ecosystem
of connected physical objects that are
accessible through the internet.)
Collar tag sensor deliver temperature,
health, activity, and nutrition insights on
each individual cow as well as collective
information about the herd.
Assistant Professor
(E&C)
Expertise in the following domains and
teach courses like:
a)Programming the Internet of things
b)Wireless and cloud Emerging
Technologies
c)Architecting smart Devices
d)Fog Computing
e)Data Analytics
Expertise in following tools/ languages
a) Programming Languages like C,
C++, Python, JAVA,
b) Tools like MATLAB/ OCTAVE/
PYTORCH/ TENSOR FLOW
Big data/
Computational biology/
Data Analytics
Assistant Professor
Expertise in
Computer programming
Bioinformatics and computational biology
Algorithm development
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
11
(Big data analytics examines large
amounts of data to uncover hidden
patterns, correlations and other insights.)
Big Data and analytics enable monitoring
and supervision for growth rate and
nutrient requirements on a plant-by-plant
basis.
Pipeline development for managing and
analysis of big data
Database development and applied
statistics
Automation (Automation is the creation of technology
and its application in order to control and
monitor the production and delivery of
various goods and services.)
Remote-piloted Drones that can spray
fungicide, Harvesting fruit and vegetable
robots
Assistant Professor
(Mechatronics)
Expertise in
Mechanics
Mechatronics
Automation
Designing and manufacturing of
machines/ devices
Subject Expert
Remote Sensing & GIS
(Technique of obtaining geographical
informationof any part of surface(land)
from space using Satellite Technology,
drones, aircrafts, etc. and processing of
data into useful information by using
computer-based tools.)
Fire detection, flood prediction, soil
properties prediction.
Assistant Professor
(RS & GIS)
Expertise in image processing, remote
sensing, pattern recognition, gig Geo-data
processing, and machine learning
Crop and Animal Modeling
1. Simulation modelling describes
processes of crop growth and
development as a function of weather
conditions, soil conditions, and crop
management.
2. Animal Modeling is the use of
advanced biometrical tools like
BLUP technology for predicting the
true genetic worth, alongwith the
algorithm writing languages like R
and Python
Assistant Professor
(Agronomy/Soil
Science/ Environmental
Science)
Assistan-*t Professor
Animal Breeding &
Genetics
Expertise in
Data and input file preparation for crop
models, assessing crop growth and yield
using crop simulation models (InfoCrop,
DSSAT and other models): Experiment
setup and running simulations, simulating
climate impacts, simulating adaptation
strategies and interpretation, Soil carbon
balance model, Pest and disease
modelling, Simulating GHG mitigation
options, Yield forecast using simulation
models
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
12
Staff Rationalisation
Area Faculty Position No of
positions
Availability through
Rationalization / up-grading
Machine Learning
Asstt Prof
(Computer Science )
1
Upgradation of 1 post of STA
( Statistics)
Data Optimization /
Deep Learning
Asstt Prof
(Computer Science ) 1 Upgradation of 1 post of computer
programer
IoT Asstt Prof
E & C engineering
1 Upgradation of 1 post of STA
Automation
(Drones / robotics)
Asstt Prof
Mechatronics
1 Upgradation of 1 post of STA
Data Analytics /
Computational
biology
Asstt Prof
Computational Biology
2 Rationalization: Asstt professor
Plant Genetics & Breeding
Agri-Statistics
Remote Sensing and
GIS
Asstt Prof
RS & GIS 1 Rationalization: Asstt professor
Soil Science
Crop and Animal
Modeling
Asstt Prof
Modeling 2 Rationalization: Asstt professor
Agronomy
Animal Genetics & Breeding
Supporting Staff
Position No Availability through
Rationalization
Stenographer 1 out of 72 available posts
ACT 1 out of 145 available posts
Accounts Asstt 1 out of 38 available posts
Lab Astt 2 out of 221 FCLA available posts
Computer Assts 2 out of 12 available posts
OCC 2 out of 301 available posts
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
13
Literature Cited
Chesmore E.D. Application of time domain signal coding and artificial neural networks to passive
acoustical identification of animals. Applied Acoustics 62 (2001) 1359–1374
FAO. 2018. Food Security & Nutrition around the World. http://www.fao.org/state-of-food-
security-nutrition/en/
FAO. 2009. Global agriculture towards 2050.How to Feed the World 2050; High Level Expert
Forum. Rome, 12-13 Oct, 2009.
http://www.fao.org/fileadmin/templates/wsfs/docs/Issues_papers/HLEF2050_Global_Agriculture.pdf
Ghazanfari S, Nobari K and Tahmoorespur M. 2011. Prediction of egg production using artificial
neural network. Iranian Journal of Animal Science 1: 11–16.
ICAR. 2015. Vision 2050. Indian Council of Agricultural Research, New Delhi, India.
Ishaq M. A. 2018. Artificial Intelligence in animal health and veterinary sciences. vHive.
https://vhive.buzz/artificial-intelligence-animal-veterinary-sciences/
Jones C.O. 2017. AI promises a food revolution, from farm to supermarket. Raconteur.
https://www.raconteur.net/sustainability/ai-promises-a-food-revolution-from-farm-to-
supermarket
NITI Aayog. 2018. Discussion Paper National Strategy for Artificial Intelligence. NITI Aayog, New
Delhi. India
Puers, B, Michiels, L., Sansen, W., Geers, R., Goedseels, V. 1988.Temperature telemetry for
domestic animals. 10th International Symposium on Biotelemetry, Jury 31 - August 5,
Fayetteville.
World Economic Forum. 2016. The Fourth Industrial Revolution: what it means, how to respond.
https://www.weforum.org/agenda/2016/01/the-fourth-industrial-revolution-what-it-means-
and-how-to-respond/
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
15
ARTIFICIAL INTELLIGENCE IN AGRICULTURE
Introduction
The world of Artificial Intelligence (AI) is quickly on the rise as it makes its way into many different
industries. From manufacturing to automotive, we can likely see AI used for many different purposes;
and as time goes on, we will only see it even more. One of the most interesting industries that AI is
breaking into is agriculture. Agriculture is a major industry and a huge part of the foundation of the
economy of various countries especially ours. In India, around 60–70% of population (directly or
indirectly) depends upon Agriculture sector and currently it contributes to 16–17% of the GDP. The
productivity in Indian Agriculture sector has increased over the years but the Growth in productivity
is still slower than Manufacturing/Services. On global level also, our productivity is much lower in
Agriculture sector because of which is reflected in 50% work force contributing only around 17% of
GDP. This is largely because of the fact that farmers still stick to traditional ways which leads to
lesser productivity. The traditional ways need to be replaced by a more efficient and stable way of
farming and one of the ways to do so is the introduction of AI.
What is AI?
Artificial Intelligence is a way of making a computer, a computer-controlled robot, or a software
think intelligently, in the similar manner the intelligent humans think.AI is accomplished by studying
how human brain thinks and how humans learn, decide, and work while trying to solve a problem,
and then using the outcomes of this study as a basis of developing intelligent software and systems.
The philosophy of AI is to make the machines do what humans door in other words Artificial
Intelligence pursues creating the computers or machines as intelligent as human beings.
Goals of AI
i) To Create Expert Systems -The systems which exhibit intelligent behavior, learn,
demonstrate, explain, and advice its users.
ii) To Implement Human Intelligence in Machines- Creating systems that understand, think,
learn, and behave like humans.
AI in Agriculture
i) Agricultural Robots
Companies are developing and programming autonomous robots to handle essential agricultural tasks
such as harvesting crops at a higher volume and faster pace than human laborers. AI companies are
focusing much of their efforts on developing autonomous robots that can easily handle multiple
agricultural tasks. These robots are capable of harvesting crops at a much faster pace and higher
volume than human workers. The robots are designed to assist in picking and packing crops while
also combating other challenges within the agricultural labor force. Additionally, agricultural robots
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
16
have the ability to protect crops from harmful weeds that may be resistant to herbicide chemicals that
are meant to eliminate them.
ii) Image-based Insights for Crop
Many farms are taking advantage of drone technology to provide high-quality imaging that can help
monitor crops while scanning and analyzing fields to collect necessary agricultural data. This imaging
technology can also assist in the identification of crops and their progress, including their health, and
the determination of their readiness. For example, these images can provide farmers with the ability to
determine how ripe their crops are, and if and when they will be ready for harvest. Additionally,
imaging technology can assist with overall field management, providing estimates in real-time
identifying where specific crops may require more water, fertilizer, soil or pesticides. Machine
learning is also used to provide an analysis of crop or soil health. Innovative AI and machine learning
companies have developed technologies that use machine learning to provide farmers and laborers
with insights about the strengths and weaknesses of their soil. This is done with the intention of
preventing and eliminating bad crops and increasing the potential for healthy crops to grow.
iii) Precision Farming
Precision farming uses AI to generate accurate and controlled techniques that help provide guidance
and understanding about water and nutrient management, optimal harvesting and planting times as
well as when the right times for crop rotation would be. These processes make farming more efficient,
and can even help predict ROI on specific crops based on their costs and margin within the market.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
17
iv) Sowing the Seeds
High-tech agriculture starts at the very second that the seed is first placed in the ground. Experts in
the field are familiar with “variable rate planting equipment” that does more than just planting a seed
down into the dirt somewhere.
All sorts of artificial intelligence work is being done behind the scenes on predictions – where a seed
will grow best, what soil conditions are likely to be, etc. The power of artificial intelligence is being
applied to agricultural big data in order to make farming much more efficient – and that's only the
beginning.
v) Pest and Weed Control
Farmers are quickly adopting new high-tech ways of protecting plants against weeds and various
kinds of pests outdoors. Another alternative is to grow in greenhouses, which is being done as well,
but some of the most amazing farming technology is being deployed outside. Airborne surveillance
engines can look for stunted crops, signs of pest or weed damage, dryness and many other variables
that are part of the difficulty of farming in general. With all of this data in hand, farmers can enhance
their production models and their strategies across the lay of the land to decrease risk, waste and
liability.The “See and Spray” model acquired by John Deere recently is an excellent example of
harnessing the power of artificial intelligence and computer vision.
Further, researchers have developed an artificial intelligence (AI) capable of identifying diseases in
plants; specifically, the cassava plant, which is the most widely grown root on the planet and a decent
source of carbohydrates. In their research published to arXiv, the team explains how they used a
technique known as transfer learning to teach the AI how to recognize crop diseases and pest damage. As
reported by Wired, they utilized TensorFlow, Google’s open source library, to build and train a neural
network of their own, which involved showing the AI 2,756 images of cassava leaves from plants in
Tanzania. Their efforts were a success, as the AI was able to correctly identify brown leaf spot disease
with 98 percent accuracy. Best of all, the AI can be easily loaded onto a smartphone, and operate without
accessing the cloud.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
18
vi) Yield Boosting Algorithms
When we talk about machine learning and artificial intelligence, we often talk about algorithms. The
mathematical models behind the computer science are the fundamental basis for how we deal with
big data to make decisions.
Companies are now quickly developing agricultural yield boosting algorithms that can show farmers
what's going to be best for a crop. Despite some concerns about the difficulty of doing this type of
analysis in nature, farmers and others have been able to make quite a lot of headway in maximizing
crop yield, simply by applying the algorithms and intelligence generators that we've built to help
computers imitate our own cognitive abilities.
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
19
Likely impact of AI on different areas of Agriculture
Agriculture and Artificial intelligence is a growing collaboration and will highly optimize the
way the things are conventionally done in the field of agriculture. There is immense potential of
this collaboration between the two fields. Some of the likely impact of AI on agriculture can be
best realized by analyzing how AI based systems can be applied to optimize the processes in the
field of Agriculture. Some of these are included as:
1. Using GIS (Geographic Information System) based visual precision agriculture to use
agricultural pesticides and fertilizers more efficiently
2. Using neural networks for improving agricultural processing output (analyzing and
predicting mechanical parameters)
3. Using CAD (Computer Aided Design) for improving agricultural equipment design and
performance
4. Using hybrid network to build information sharing platforms (e-governance, climate,
agricultural practices, expert pools, etc.)
5. Using laser induced imaging for detection of contamination of agricultural products
6. Using sequence alignment in molecular biology to improve crop quality
7. Developing e-commerce (seed sales, produce sale, prices info, etc.) platforms to boost
agricultural economy
8. Working towards improvement in low altitude image acquisition systems
9. Optimization of irrigation models using genetic algorithms
10. Developing design for automated sorting systems of farm produce
11. Building GIS and MCE (Multi-Criteria Evaluation) mapping for land quality evaluation
12. Using Support Vector Machine (SVM) technique for soil water content forecasting
13. Using 3S (Remote Sensing, GPS, and GIS) techniques for predicting landforms and
climatic changes
14. Building simulation algorithms for optimization of pesticides and fertilizers spray methods
15. Helping development of sustainable wireless monitoring systems (long-range, automatic,
etc.)
16. Developing spatial distribution maps of soil nutrients
17. Developing geospatial computation grid for digital agricultural experiments
18. Creating crop disease leaf image segmentation using imaging analysis techniques
Centre for Artificial Intelligence & Machine Learning in Agriculture (CAIML)
20
Major sub disciplines of AI and Agriculture
1. Computer vision
2. Image Processing
3. IoT
4. Robotics
5. Computational biology
6. Machine Learning
7. Data mining
8. Data science
9. IT for Agricultural systems
10. Pattern Classification
11. System Biology
12. Bioinformatics
13. Computer Programming
14. Applied Mathematics
15. Algorithms