Artificial Intelligence (AI) relates to how we observe,
feel, learn, reason and act. This is transforming entire systems of production,
management, healthcare, and governance.
Due to exponential growth in data processing power, AI is continuing to
gain momentum in Medical Diagnostics (MD), Machine Learning (ML), Deep Learning
(DL), document retrieval and processing,
Business Intelligence (BI), Industrial Automation, Research & Development
(R&D). Particularly, in the last 10 years, AI has picked up a fast pace in
innovation and application in many fields. It has a great significance in fast
developing Digital World. AI
is based on well-designed algorithms by a team of experts from
multi-disciplinary areas and it is stored in the equipment/device as embedded software.
AI is no more a threat
to jobs and instead, it is becoming the lifeline of every organization. It is therefore
essential for the top management and all staff to be fully acquainted with AI. They should be enthusiastic to get the best out of this
magic technology which is creating ripples in Digital World. Almost all CIOs, CKOs and top technology
executives of leading companies have identified AI as the most disruptive technology
that is transforming their business models.
However, it does not mean that AI will totally upset the business model.
Instead, AI will help to remodel various processes and make the whole system
highly efficient and more productive. The senior management must be fully involved and
committed to taking full advantage of AI wave which is now sweeping the entire
industrial world at a great speed. The senior
management should review their existing processes and their expectations from
their AI initiatives. You can't really set your business strategy unless you
know what AI technology can do for your business. Educating the top executives, about the
capabilities, limitations, and requirements of AI is the most essential/tricky
job for any CIO/CTO/CKO. You need to cover main AI concepts, guide them on how to identify areas where AI can be gainfully applied in their own
organization. To prepare the new young workforce in AI, most of
the engineering colleges/universities, IITs, NITs and other institutes of
higher learning have introduced UG ( 4 Years) / PG( 2 Years) level programs in
AI and related areas. AI, ML. DL, Data Science is becoming more popular
programs than evergreen Computer Science and Engineering.
Background of AI. AI has
not suddenly emerged but has been there as a subject being taught from 1935
when Turing, a UK scientist who thought of Machine Intelligence (MI) and
published his first report during World War II in 1943. However, a major part
of AI implementation was done during 1950 and 1960s. In those years, most researchers thought AI
use in a two-sided game like Chess and it is still in vogue. In 1995. IBM's Deep Blue; chess
computer could beat Garry Kasparov (Russia) and becoming the
first computer to defeat a human
world champion. However, in 1997, IBM's Deep Blue lost
to Garry Kasparov, then the top-rated chess player in the
world. AI is being taught as a
subject of computer science in most of the engineering colleges, universities /
IITs and NITs in India and abroad at UG/PG level since the 1970s
but no real-life application was thought of. During 1970s LISP was
used for AI programming to solve simple problems related to health care, speech
recognition, image processing. During the 1990s, ALGOL, PROLOG became popular
languages for AI but again AI remained restricted to academic interest in areas
like neural networks, healthcare and speech recognition. Other areas related to
AI are computer science, mathematics, neuron science, psychology, biology,
philosophy, sociology. Research in AI has been focussed mainly on learning, reasoning,
problem-solving, perception, pattern matching, and language understanding.
China plans to catch up with the west in AI utilization by 2020 and lead the
world in AI by 2030. China has very large data with 751 Million Internet users.
In October 2017, Chinese President Xi, Jin-Ping had set the goal of promoting
further integration of Internet, Big Data and AI to become a leading world
economy. Likewise, many other nations like the USA, Russia, Japan, England, France,
Germany, Holland, Israel, Canada, Sweden, Australia, and India have given
priority for various AI initiatives in many fields.
Definitions and Concepts.
Definitions and Concepts.
- Artificial Intelligence (AI). AI relates to where computer stored program is able to perform tasks normally performed by human intelligence ( Natural Intelligence) such as visual perception, face reading, speech recognition, translating between languages, decision-making, moving and positioning of objects. It is a computer/machine’s ability to function as a human and copy some of the human behavior. By embedding suitable algorithms, the machine can learn like a child and soon acquire sufficient knowledge to function in autonomous mode. However, common use of AI is often confused between Data Science, Data Analyst (DA), Analytic, ML and DL which are interrelated specializations. Simple definitions and applications of these terms are briefly given below:
· Data type and
format. Unlike yesteryears, today
data is not just numbers and text in structured form but it includes unstructured data as pictures, sound, videos, multi-media coming from organized/unorganized
sectors, public and social media, flowing across geographical boundaries at a fast speed and
round the clock. This data is characterized by 3Vs - Variety, Velocity, and
Volume. This has greatly impacted
techniques required for data gathering, organizing, processing and interpreting
for decision making in various AI applications.
· Data is Power. As per the Google Cloud platform, over 2.5 quintillion bytes of
data is being generated around the world on a daily basis. Likewise,
other cloud service providers like Adobe, Amazon, IBM, Microsoft, Oracle, Phoenix, Rackspace, Red Hat, and Verizon will have similar large
data. This requires special software tools and techniques to process and extract intelligence. Indeed data is power and the country which can handle their data
efficiently and convert it into actionable intelligence will lead the way.
· Data Science. Data science is
a multi-disciplinary field that uses scientific methods,
processes, algorithms and computing systems to extract intelligence from
structured and unstructured data. At the core of AI is data, raw information, streaming in
and stored in enterprise data warehouses, network nodes and various sensors/computing
devices. This needs data mining to use that outcome in creative ways to
generate required BI.
· Data scientists. Data scientists are
expert in mathematics, domain knowledge, and technology to work at the raw
database level. They are deep thinkers to dig down and discover new solutions. A data scientist has to be very inquisitive asking new questions, making new
discoveries, and learning new things.
· Data Analyst. Data analysts look at data to try to gain insights and they may
interact with data at both the database level and the summarized report level.
- Analytics. The term Analytics has become popular in the past few years. They are specialists in critical thinking and quantitative/ statistical techniques for fast decision making.
- Analytics. The term Analytics has become popular in the past few years. They are specialists in critical thinking and quantitative/ statistical techniques for fast decision making.
Machine Learning (ML). ML is a term closely associated with data science.
It refers to a broad class of methods that revolve around data modeling to
algorithmically decipher patterns in data and make predictions. The Core concept is to use tagged data to
train predictive models. It is similar to teaching a small child how to react
to various objects/events. Tagged data means observations related to
domain knowledge. Training model means automatically characterizing tagged
data to predict tags for unknown data points. It
uses regressions technique to deal with the complexity of neural networks.
ML is also used for pattern recognition. It makes use of clustering technique,
which algorithmically detects what are the natural groupings that exist in the
data set.
Deep Learning (DL) Deep Learning is one of the most highly sought after skills in a technology-driven market. You should learn Python and Tensor Flow to build
neural networks, Convolutional networks, RNNs, LSTM, and Dropout. This will
enable you to plan and execute ML projects related to healthcare, autonomous
driving, sign language reading, music generation, image processing, document
retrieval and summarizing and natural language processing.
AI Data Bases. AI database size and how it works is quite
different than traditional RDBMS which is based on relational algebra and SQL.
AI database combines data warehousing, advanced analytics, and visualizations in an in-memory database.
Principal Software Engineer of Kinetica's Advanced Technology Group has highlighted
that an AI database should be able to simultaneously ingest, explore, analyze,
and visualize fast-moving, complex data within milliseconds. One of the
challenges for training ML and DL
models
is to have large data volume and processing power.
Many Computing companies have developed special AI Chip to handle large volume,
velocity, and complexity of data. The chip has to have an embedded algorithm which
makes the chip layout complex. AI chip will help in pattern recognition, image
classification or natural language processing (NLP).AI databases use special software tools like Hadoop,
map-reduce, and HDFS.
Business Intelligence (BI). Data and BI are two sides of the same coin.
Advancements in storage, processing, analytic techniques and software tools
have reached a point where you don't need to be a database professional or data
scientist to work with massive data sets and derive required intelligence. BI
and data visualization tools
are redefining many businesses processes to achieve high productivity and
efficiency. In global market speed and accuracy of information provides you the
cutting edge and for that AI is the best bet.
Boosters for rapid advances in AI.
- · Participation
by educational institutes in offering courses related to AI skills.
- · Participation
by the industry in indication and validating AI related products and processes for higher productivity, better
efficiency, and improved quality assurance.
- · AI
models have become more comprehensive with easy access to Big Data generated
from e-commerce, businesses, governments, science, wearable, and social media.
- · Response
time for required information has greatly improved due to efficient storage, retrieval
and summarizing of legal proceedings, medical diagnostics, insurance claims.
- · Improvement
in ML algorithms due to the availability of large amounts of data.
- · Availability of greater computing power and processing speed for an affordable cost.
- · The rise
of more efficient and versatile programming languages like Python. It has helped in
developing comprehensive algorithms and embedding those in the microchips.`
- · Easy availability of affordable cloud-based
services which help in running/ validating ML algorithms.
Functional grouping. In today’s economy, AI is a very
versatile and useful software tool. AI
can help to solve complex problems related to various organization/departments
like the judiciary, security, entertainment, education, health, commerce,
transport, agriculture, and defense applications. AI applications can be categorized
into five functional groups:
- Prediction and forecast: Ability of AI tools to quickly sift through large data related to weather, industry production, sales, project management, health care, education, agriculture, and employment.
- Planning: The ability to set and achieve goals. e.g. inventory management, demand forecasting, predictive maintenance, physical and digital network optimization, navigation, scheduling, and logistics.
- Knowledge: The ability to present knowledge about the real world. e.g. financial market trading, purchase prediction, fraud prevention, drug creation, medical diagnosis, the medical recommendation.
- Perception: The ability to infer things about the real world via sounds, images, and other sensory inputs which are needed for medical diagnosis, autonomous vehicles, and surveillance.
- Reasoning: The ability to solve problems through logical deduction in areas like financial application processing, financial asset management, legal assessment, autonomous weapons systems, entertainment games.
- Language Translation: The ability to understand spoken and written language for real-time translation to any specified language. This is very helpful in international conferences, AI can also understand sign language, interpreting gestures and facial expressions.
- Emerging AI trends in various sectors,
- Healthcare. AI and ML technologies have been very useful in the healthcare industry because it generates large data to train and fine-tune algorithms for locating disease patterns faster than human analysts. Some prominent utilization of AI and ML is in diagnostics of cancer, eye ailments, diabetes, drug effect, DNA analysis, patient symptoms and suggested treatment.
- · Entertainment. The service providers of various Cinema screens and TV Channels use ML algorithms to analyze and compare their viewership with that of other service providers and recommend specific shows to retain their customers. This also helps to promote various products and services on new channels and maximize their revenue.
- · Finance. Financial services companies use AI-based NLP tools to analyze brand sentiments from social media platforms and provide actionable advice to the business houses/promoters. Some finance companies use AI-powered B2C robot-advisors for portfolio management. Fraud detection in misuse of Master-Card / Visa-Card is a very important application of AI It can swiftly detect fraudulent transactions and approval or decline, fraudulent transactions.
- · Data security.With faster and affordable internet facilities and easy availability of smartphone-like devices; cyber-attacks are becoming a common feature. As most of the nations are moving toward digital world there are many serious concerns about security and privacy of data. This is being tackled on a war footing since these effects, banking/financial institutions, stock market and even political campaigns like USA presidential election in 2016. AI algorithms can automatically do bug fixing. Likewise, AI could also predict the likelihood of cyber-attacks.AI can also detect and remove malware code.
- · Manufacturing. ML and Robots/Cobots loaded with AI have revolutionized the entire manufacturing sector be it automotive, household appliances, textile, or processing industry. Some very common examples are shown below:
- ML algorithm can assist to find microscopic
defects in objects like circuit boards
- Self-driving
robots may be created which can move finished goods around without
endangering anyone or anything around.
- Collaborative robots (cobots), These robots are enabled
by AI to take fresh instructions from humans and work collaboratively with them.
- Supply chain. AI can help in detecting the patterns of demand for products
across geographical regions, socioeconomic segments, delivery time
- Inventory Management.
AI can help in inventory management and vendors management for raw
material sourcing, financing decisions, human staffing, energy
consumption, and maintenance of equipment.
- Prevention breakdown
of machines. AI can
assist in predicting malfunctions and breakdown of equipment and
taking or recommending pre-emptive actions
- Monitoring work
environment. AI can help in tracking operating conditions and performance of factory tooling.
- Automotive industry.
- Self-driven cars. Google and
Uber has already developed and launched self- driven cars in the USA and
Europe. This is helpful for elderly or handicapped people to reach their definition safely.
- Car Safety. The AI-based system when installed in
the car will enable safety
warnings, related to speed, approaching vehicle, lane changing,
vehicle-to-vehicle communication wherein sensors will be installed along the roads.
- Rail safety. To avoid rail accidents, an AI-based system can
be installed in train driver’s compartment and installing sensors at level crossings, track changing points. It will control train speed as
per track conditions and also avoid any head-on collision due to wrong
track changing.
- Car theft alarm. Some
cars are already loaded with a microprocessor to match the car key with pre-stored
software key and cut off car ignition system so that the unauthorized a person cannot take away the car. In addition, AI software can recognize biometric signs and set the alarm if an unauthorized person attempts to open the car.
Build your career in AI. There are many areas where you can
enter and build and excel in your skill set. AI is now used extensively in many applications like:
- · Experts Systems. AI is being gainfully used in agriculture, road transportation, and weather prediction.
- · Knowledge-Based Systems (KBS). These are being used in many Defence and some Business applications.
- · Fuzzy Logic. It is finding vast applications for household appliances and business applications.
- · Speech recognition. This is quite a common application since the early 1970s. Auto response to queries for travel and hotel reservations.
- · Robots. The heart of the Robot is AI and embedded logic. Robots are already being used in many Military activities, Healthcare, Business applications, Animated movies, and entertainment.
- · Machine Learning (ML). AI forms a major part of ML Systems which are being used for diagnostics and interpretation of laboratory tests. ML is playing a major role in industrial automation.
- Deep Learning (DL). It is being used in medical diagnostics and document summarizing and even preparing draft speeches for political leaders as per their agenda.·
- Navigation and control system. This is particularly useful for aerospace missions.
- · Lie Detector. A lot of research in the USA and UK has been carried out to accurately detect micro-expressions on the face of the accused so that police/law enforcement agency can accurately detect if the suspect is telling a lie. This facility will also facilitate the judge to award punishment grant bail/parole to the convict.
- Prediction in Casinos. In leading casinos like those in Los Vegas, the owners are using AI to recognize and record facial expressions to identify risk appetite of probable losers.
- Guided weapon systems. AI has many applications in military warfare. It is especially needed for surveillance, navigation/tracking, terrain analysis, weapon database, and their lethality index.
- Tracking transport fleet. AI gets integrated with satellite imageries and Global Positioning / System (GPS) to monitor and control the movement of transportation systems.
- Self -learning. There are many tools for language translation and presentation of course content. It helps learning any foreign language at your own pace and pace and adds value to your CV.
- Web translation. A number of internet service providers are offering real-time machine based translation from one language to another language.
- Video gaming and entertainment. Toy industry with embedded AI has a great market for children entertainment. Video games needing higher skills and visualization need AI component suitable embedded.
Sources for AI Courses. To
give impetus to AI and meet the ever-increasing demand for AI qualified workforce, most of the leading universities in the USA, UK, Canada. Australia and Indian IITs,
NITs have revised their curriculum of UG/PG level engineering courses in Computer
Science, Electronics and Communications
, Instrumentation and Control. You can choose full time or part time course at
your convenience and cost. In addition,
you can choose online free AI courses from:
- USA. Learn AI with Google – ML, Stanford University – ML, Columbia University – ML, Nvidia – Fundamentals of DL and Computer Vision, Columbia University – ML, MIT – DL for Self Driving Cars.
- India. IIT Bombay -NLP, IIT, Kharagpur- AI, NIT. Varangal – ML and DL, Cisco Bangalore-ML.
If you undertake the above
on-line AI course and do not pay the fee for certification, the course will be free
but considered an audit. It is therefore recommended to pay a fee of $40 to $60
and get certified, as that will show your competency and value addition to you
CV.
Required Software Tools.
· Programming
Languages for AI Applications. We can
select the best programming language according to its usage, functionalities,
and features. For AI applications, Java, Python, Lisp, Prolog, Ruby, and C++ are the most popular programming languages. All these programming languages are capable of satisfying
different requirements for development and design of different AI related
application software. It
is up to a developer to choose which of the AI languages will satisfy the
desired functionality and features of the application. There are some
advantages and some limitation of these languages when we deal with Mobile
phone devices, very high response time and large heterogeneous data. However, if you have mastery of one of these languages,
migrating to others is quite easy and takes 2-3 weeks. Among most of the AI
developers, Python is the favorite programming languages for AI
development because of its syntax simplicity and versatility. Python is very
encouraging for ML for developers as it is less complex as compared to C++ and
Java. It also a very portable language as it is used on platforms including
Linux, Windows, Mac OS, and UNIX. It is also likable from its features such as
Interactive, interpreted, the modular, dynamic, portable and high level which make
it more unique than Java.
Python supports,
multiprogramming object-oriented, procedural and functional styles of
programming. Python supports neural networks and development of NLP solutions.
It has a library of many simple functions
which speeds up coding.
· Mathematics Expertise. Building analytic models for various business problems involves using applied mathematical/statistical techniques. Hence Data scientist should have sound
knowledge of applied mathematics covering linear algebra, matrix algebra,
numerical methods, and statistical techniques. These are needed for data mining
of vast data and building data product. There are dimensions and correlations in
data which can be expressed mathematically in developing algorithms.
· Strong
Business Acumen. Having good business
acumen is just as important as having the acumen for technology and algorithms. The
real value doesn't come from data, math, or technology itself. It comes from
leveraging all of the above to build an appropriate model and develop a business solution.
· Statistical
Tools. Beside SAS, “R” has become
more popular statistical programming language in handling advanced statistical
tools for forecasting, interpreting and decision making.
· Embedded
Technologies. AI uses the embedded software in microcontroller and microprocessors
which form the heart of Robots, Drones and navigation system. It uses lots of
inbuilt controllers, which are programmable with autonomous built-in control
and/or external remote control system. Embedded software and optical devices are
facilitating many activities in areas of healthcare, diagnostics, physiotherapy, and surgery.
As most of the countries are now becoming Digital Economies, AI has a
big role to play. In it's Budget 2019, India has given a special boost to AI for
agriculture, processing, transportation, education, and research. AI also
supports military warfare, political campaigning /psychological warfare. AI has also a role in controlling global warming.
In the coming years, there will be more
integration of technologies to reap the full benefits of AI and associated
technologies.
An old view about AI is that it is job snatcher and makes people
redundant earlier than expected. A similar
view was expressed by Jack Ma, the founder of Alibaba, at the World Economic
Forum 2018 at Davos “AI and Big Data were a threat to the human workforce as
it would restrict their human (natural) thinking instead of empowering them. People will stop thinking, analyzing and
making a decision. They will depend
heavily on machine-generated information”. However, his view is not fully supported by other economists and technology experts
who promote AI as a software tool for achieving enhanced
efficiency, more productivity, and higher profitability. Given numerous
real-world applications of AI, ML, DL and continuing advancements in AI, it
will transform the way we work and do our business in and outside
the country. AI will
create many new jobs requiring new skills. An AI skilled workforce will draw a better salary. AI will supplement human effort in many fields leaving human
beings to do more creative thinking and innovation, better integration/interfacing
systems and sub-systems. Future will be different and you must acquire AI
skills to make the difference.
For more
articles, visit my blog http://sarbjit-share-knowledge.blogspot.com
For more
information you may visit www.amazon.com and
refer to my books/e-books :
1)
A2Z – 26 steps for assured success
2) Career Challenges during Global Uncertainty
Visit the website :
www.kmowledgeshareindia.com
Dr. Sarbjit
Singh, Former Exec Director, Apeejay College of Engineering & Hony
Advisor, Apeejay Stya University, Gurgaon, Haryana,
122103, India
Comments
Early adoption is the KEY
Great Article
Nice Article written in simple language
AI is becoming way of our life