Machine Learning Algorithms

We are heralding a new dawn with 2020 – a world where neural networks, deep learning, and NLP are fast coming up as competitive differentiators. Businesses across the world are embracing Machine Learning technologies with gusto. Hence it doesn’t come as a surprise that Boston Consulting Group predicts 25% job loss due to machine learning algorithms and automation.

While data is essential for machine learning solutions, the algorithm behind it is equally essential. If you are looking for the most popular machine learning algorithms in today’s times, then check out our compilation:

1. Naïve Bayes Classifier Algorithm

This algorithm performs well with huge data spanning millions of records in the data set. There are two main forms of this algorithm – Gaussian Naïve Bayes (Apply normal distribution to continuous attribute values) and Multinomial Naïve Bayes (for data that shows multinomial distribution)

This type of algorithm shows immense value in –

1. Sentiment Analysis (used by Facebook to assess status updates)
2. Document Categorization (Google uses it for Indexing for PageRank)
3. Spam filtering (used by Google)

2. K Means clustering Algorithm

This is one of the simplest unsupervised learning algorithms that can solve the common clustering issue. It displays better clustering performance than hierarchical clustering. Its key application is by search engines like Yahoo and Google.  They use this technique to group the web pages based on the ‘relevance rate’ of the search queries.

3. Support Vector Machine Learning Algorithm

SVM algorithm continues learning from the classified data set to understand the classification pattern and apply it to new data. On training data sets, SVM is known to provide high accuracy and incredible classification performance. It uses the concept of hyperplane (a line) to classify data. The more it is away from the hyperplane, the better is its classification accuracy.

This algorithm is prominent in financial sectors where analysts try to compare stocks and its movement vis-à-vis competitive stocks or benchmark stocks.

4. Linear regression

In simple terms, it compares the inter-relation between two variables and assesses how a change in one impacts the other and to what extent. It is used by small and medium businesses in their revenue forecasting and team growth predictions.

For instance, if the sales are following a linear trend, this algorithm can accurately forecast the possible sales in the upcoming months.

5. Artificial Neural Networks

This is a class of algorithms modeling that mimics the biological neural structure. Some of the popular algorithms in this category include Perceptron, LSTM Recurrent Neural Network, Boltzmann Machine, and Radial Basis Function Network (RBFN). They are routinely used for regression and classification problems.

These networks can combine with other networks through the input layer, output layer, and hidden layer. This, in turn, forms multiple layers which give rise to deep artificial neural networks. Their architecture expands the realm of machine learning to delve into deep learning.

Which other machine learning algorithms have you been using for meeting your specific business objectives? Do write to us and let us know.

Skills Artificial Companies should invest in

If we take a look at the present market trends, it will be easier to get an idea of immense growth of AI . This growth spans companies across industries, maturities, and competencies. With many startups, established ventures, and tech giants gearing up to embrace AI technology, there’s no denying its bright and promising future.

Amidst all this hype around AI, one can identify its rapid adoption. It’s true that a lot of companies belonging to diverse sectors have adopted AI methodologies.

Let’s take a look at some market stats that explain the situation better!

Market statistics

According to reliable market reports, 38% of ventures are already leveraging AI. This is done with a goal to accelerate growth and development. By the end of 2018, the figures will rise to 62%. That explains why AI is important and crucial for ventures across the globe.

However, with the ever-growing importance of Artificial Intelligence and increasing market demand, it’s high time to identify ways to improve it.

The top 4 AI skills

If you are running an AI organization, you must be aware of the skills that empower ventures across the globe. Here’s a quick look at the top four skills every AI Company should invest in:

1. Natural Language Generation or NLP

Natural Language Processing is an innovative technology that helps machines to understand language the way it is spoken or understood by humans. Different approaches like distributional, frame-based semantic analysis, or interactive learning will help upsell your NLP experts’ skills within the AI ecosystem.

2. Chatbots

When it comes to ensuring excellent customer service, most of the ventures opt for chat bots. Chat bots can connect with human intelligence thus helping organizations develop a unique relationship with their customers. Knowledge on AWS S3, AWS ML, jQuery, JavaScript, SQL, and PHP will be vital in developing chatbots.

3. AI-centered hardware

Appliances, graphic processing units, and other hardware units should be designed, developed, and created with AI in mind. Companies should develop hardware systems that are appropriate to perform AI-centered computational tasks. Hardware units are as important as software skills. Quite naturally, it becomes imperative to develop and design them with AI in mind.

4. Biometrics

You must have heard the term ‘biometrics’ while checking out the staff attendance systems in enterprises. AI is growing and that too at a rapid pace. Quite naturally, organizations should enable increased interactions between machines and humans. The ideas shouldn’t be only restricted to speech, body language, and touch recognition. Biometrics encompasses a broader perspective and positive interaction is a crucial part of it.

Final thoughts

Artificial Intelligence processes, technologies, and methodologies are talks of the town. There’s no denying the buzz they are creating in the professional arena. AI companies should strive hard towards developing skills that empower organizations. This will help them deliver immense business value.

You have to identify development opportunities and incorporate the right strategies for the right set of business needs. Develop these skills if you are running an AI organization; they will take your venture to unsurpassed heights!

Meet Mitra – The Made in India Robot

Mitra – a robot created under the mission of ‘Make in India’ recently attracted many eyes at the Global Entrepreneurship Summit (GES) 2017 where she greeted the honorable Prime Minister – Narendra Modi and Ivanka Trump. Invento Robotics is the smart brain behind creating this 5 foot robot and Mitra is becoming highly popular among the robotics enthusiasts and other industries.

Image reference: Hindustan Times – Mitra Robot at GES 2017

Want to know why Mitra is being considered a one-of-its-kind robot in the global robotics arena? Here are 5 good reasons:

1. Quick Development

While all the other humanoids took years for production, Mitra was created in just one year. And even with such quick creation, it contains exceptional and revolutionary features that make the robot extremely spectacular.

The features, the design, and the performance capacity have allowed Mitra to add value to customer interactions at places like Canara Bank as well as PVR Cinemas as well. The robot is being used in many branches of Canara Bank and also assists people with their basic banking queries and surprises guests with its multi-lingual skills. It also interacts and helps people in the PVR cinemas of Bangalore.

2. Better than the best

When it comes to comparison with other humanoids, Mitra is beating the best. In fact, the quality and the performance level of Mitra is being compared to ASIMO, a Honda project in Japan. While ASIMO took about 17 years to get ready, Mitra has matched its performance level within just one year of work.

3. Highly interactive

The name ‘Mitra’ means ‘friend’ and the humanoid justifies the given name completely. The robot has an intuitive capability that allows it to talk and interact with people conveniently. The facial recognition feature helps Mitra instantly recognize customers and provide the desired serrvices. This definitely bumps up the user experience at any store.

4. Memorizes faces

Another great reason for Mitra being an awesome humanoid is its ability to remember faces. The robot has the capacity to memorize the face through different identity signals. This allows the robot to provide a personalized approach to the interactions. This ability is also the reason why Mitra successfully steals the highlight during events and public appearances.

5. Extremely versatile

Apart from all the features, it is the active nature of the robot that makes it highly valuable for various industries. The combined features and mechanical performance allows the robot to perform the given tasks. The ease of leveraging the performance of this humanoid is the reason why Mitra has already started delivering outcomes for stated objectives in banking and entertainment sectors. It won’t be a wonder if it is soon embraced by other industries as well.

In closing..

Thus the above points depict why the humanoid Mitra has obtained the green signals in terms of funding and time. This, in turn, will help it evolve and get more advanced in the coming years. It would be interesting to see the evolution of this humanoid in the coming future across a broader scope of practical applications.

Deep Learning in Marketing and Sales

The buzz and enthusiasm about deep learning has significantly increased over the past few years. With numerous business ventures embracing this technology for good, deep learning is earning huge popularity now. If you are aware of Machine Learning and its implications, understanding Deep Learning won’t be that tough.

Here’s a quick look at what is Deep Learning and its serious applications in the business arena.

What is Deep Learning?

DL or Deep Learning happens to be a significant part of ML. It can be referred to as a subset or subdivision of Machine Learning that maps artificial neural networks. The mapping takes place to recreate or replicate processes performed by human brain.

That’s not all. Deep Learning also plays a crucial part in the identification of speech patterns, algorithms, images, and data analytics.

In spite of this simple and easy-to-understand introduction, there’s no denying the complexities involved in Deep Learning. Implementing DL strategies and incorporating them into existing business processes isn’t an easy affair. It becomes imperative to keep numerous aspects in mind thus devising effective DL strategies.

Transforming the business world with Deep Learning

Deep Learning can be revolutionary. If implemented in the right way, this particular technology can transform business processes to a great extent. Deep Learning helps to decode complex unstructured data and derive consumer insights that are crucial for creating sales and marketing strategies.

From retail and transport to healthcare and manufacturing, DL has started making a mark in various sectors. Let’s take a look at how it’s transforming sales and marketing for businesses:

1. Automating end-to-end customer journey

As mentioned earlier, deep learning will allow marketers to access insights from unstructured data sets such as image, video analytics, speech recognition, facial recognition, text analysis tools and much more. In short, deep learning becomes a way to accurately understand the voice of a customer.

Customer feedback and expectations can be gauged on a real-time basis and business organizations can get information to upgrade their products and services. Based on these premium insights from deep leaning, Brands can articulate the right messaging to the right customer.

2. Understanding analytics from IoT products 

Home automation is creating profitable avenues for organizations across the globe. Deep learning can help businesses understand the analytics for IoT products. It helps to capture data from the machines in different scenarios and monitor them in an easy and cost-effective way.

Through these analytics, deep learning can help to understand the interactions between machines and customers better and based on the data, the performance of IoT products can be enhanced from time to time.

3. Using Chat Bots to enhance CX

The presence of chat bots has revolutionized business marketing to a great extent. Chat bots leverage data mining, artificial intelligence, and natural language processing, thus creating new ways to interact with the end user.

Through chat bots, customers have the opportunity to engage in personalized communications. Apart from ensuring unmatched consumer experience, Chat Bots help an organization to have timely conversations with users and give them product recommendations or suggestions. You can create your marketing strategies with consumer preferences in mind thus offering them targeted products.

4. On-going predictive analysis

Deep Learning plays a highly significant role in the data analysis process. Whether it’s a small or big organization, entrepreneurs will have the chance to perform successful data analytics. Predictive analysis becomes easier, and you can develop crystal clear ideas of customer preferences.

Signing off

If market reports and figures are anything to go by, Google is running 1000 deep learning projects as of now. The number of projects was only two in 2012, and this will succinctly explain the massive importance of DL in business marketing.

Artificial Intelligence and Machine Learning in the Retail Industry

While the world is busy talking about Artificial Intelligence powered technologies such as self-driven cars, machines challenging human intelligence at a game of chess, and AI technologies in recruitment, there still remains an untapped potential for AI in the retail industry.

With most retailers now focusing towards providing an omnichannel experience for their shoppers, AI can play a crucial role in disrupting the retail industry.

Creating Smart Shops with Artificial Intelligence

While AI assistants such as Siri and Google Home help us maintain our day-to-day groceries list and change our shopping experience, an ideal situation would be to walk into a smart store without any shop attendees or long check-out queues.

Imagine a shopping experience where you can just enter a shop, pick up the stuff you want, a bunch of facial recognition algorithms process your purchase, and automatically money gets deducted from your digital wallet when you leave the store! Suddenly, shopping seems to become more of technology experience.

Amazon has been marginally successful in replicating this experience for customers with its Amazon Go grocery store. In order for shoppers to use this, they need to install the Amazon Go app. There are digital sensors on the shelves that detect when purchases are made and the money gets deducted from the customer’s account when he or she exits the shop.

Amazing isn’t it? This is what a future-ready AI store should ideally look like!

Predicting online customer behavior with Machine Learning

We live in a digital world today where every minute truckload of data about customer behavior is recorded and stored. But most companies struggle in extrapolating this data into actionable steps that can increase their ROI by 10X times.

You may have luxurious marketing and advertising budgets to target your customers, to get awesome click-through rates, but how are you making your customers tip over the fence and make a sale?

With Machine Learning tools, you can create an intelligent and automated marketing system that helps you with:

• customer segmentation
• predicting customer value and
• designing product recommendations.

But it doesn’t end with just that. Machine learning also helps us optimize our Ad budgets and invest them in the right customers.

In a world where resources and time are always limited, we are forced to make quick and smart decisions. But machine learning algorithms are programmed to analyze this huge pool of customer data within a fraction of seconds and identify customers who are 65% more likely to make a purchase. Thus, in this way, ML helps to justify and set optimum ad bids to target customers.

Product recommendation is another favorite pick for top marketers. ML helps to analyze online customer behavior in depth, such as the products they are interested in, the blogs they read, the price ranges they operate at etc. Based on all these interactions, a business can invest in curating content that makes the shopping experience as “personal” as possible.

Artificial Intelligence and Cognitive Computing in the Retail industry:

AI and cognitive computing are adding the “innovation” to the retail industry. Creating an omnichannel experience for customers has become the top most priority for retail brands.

Out of the many retail sections, in this blog post we are going to concentrate mainly on two AI integrated retail categories:

1) Product Recommendations:

IBM’s Watson, powered with its cognitive computing abilities is an excellent choice for brands who are looking to provide both in-store and online product recommendations.

• AI-powered retail store

In this example, we see how a New York based wine company called Wine4Me decided to choose the AI technology of IBM Watson to make in-store wine recommendations for its customers. The goal was to make wine shopping an easy and personalized experience.

But the first step was to provide Watson with ample data from wine tasters and also teach Watson how people ask and shop for wine. Based on the occasion, price, sweetness,  brand, and age of the wine, Watson should be able to bring up a list of wine recommendations that would suit the customer’s demand.

Now, this whole AI-powered shopping technology works as a win-win for both the retailer and the customer. Customers, on one hand, enjoy a hassle-free and engaging shopping experience while retailers can make better inventory decisions by tracking customer preferences.

• AI-powered digital store

In this example, we will talk about how the San-Francisco based premium brand The North Face integrated Watson on its digital platform to create an unparalleled shopping experience for its customers. The North Face is a premium outdoor product company specializing in outerwear, coats, backpacks etc

Though SEO filters and keyword phrases help in attracting customers to the website, most brands forget to communicate with their customers to help them find what they want. This is where AI can help you establish that user connect.

Let us assume that a user wants to buy a jacket online, then the AI platform asks the user a series of questions such as the purpose of the jacket, the time period during which he will use it, which place will be visiting, color preferences, material preferences and so on. The user is free to answer all these questions in his own style and type in any response.

With every response, the AI will trigger a series of product recommendations that closely match the user’s requirement. Now, this sounds fancy, doesn’t it?

But it all boils down to data and also literally teaching the AI what would the possible use case scenarios be that users would identify themselves with. Even the most non-obvious statements and taken for granted business scenarios need to be keyed in so that the AI helps to generate the right response.

This next-generation online shopping experience is sure to make both customers and retailers conscious of their expectations. Thus, if used in the right way, AI has great potential to provide product recommendations that can increase revenue from sales by 10x.

2. In-store Sales:

AI is all set to change the traditional shopping experience when a customer walks into the store. With e-commerce booming, you find lesser customers who are willing to enter a store and interact with a store assistant.

So, how can AI help in increasing the customer foot-fall at a brick and mortar retail store? The answer is simple – by using AI-powered robots.

Pepper,  a humanoid robot from Softbank’s robotics company has been in the news for enriching the customer shopping experiences at retail stores and serving as a retail concierge. Stores across the US, who have adopted Pepper as their smart shop assistant, have recorded almost 70% increase in customer walk-ins and sales.

Now that’s huge, isn’t it?

But what does this humanoid Pepper do differently from a salesman?

Firstly, Pepper acts as a brand ambassador for the store that drives in customers. Based on sheer curiosity, customers are more likely to stop by your store to experience an interaction with a robot who could help them with their shopping decisions. It almost feels like having a friendly celebrity with whom you can take a tour of the store.

Secondly, the humanoid robot is programmed to sense movement, emotion, offer shopping advice, product information, etc. Thus, it helps the user make an informed purchase. It has a tablet positioned on its chest where the user can key in his preferences and based on that Pepper will offer suggestions.

This definitely boosts user engagement and conversations inside a store. And the best part is that Pepper is always connected to the internet and all the data thus collected is stored in the cloud. So, next time when the same customer enters the shop, it becomes easy to showcase products based on his/her interests.

Thus, Pepper helps to automate all the obvious and mundane tasks at a retail store and transform in-store shopping into an engaging and joyful experience for customers.

shopping experience while retailers can make better inventory decisions by tracking customer preferences.
add: The rise of retail AI solutions is further transforming the industry by enabling real-time analytics, personalized recommendations, and demand forecasting, leading to increased efficiency and customer satisfaction

The future of Artificial Intelligence in Retail

While we are positive and hopeful that the future of retail is likely to be dictated by AI, we also notice that there is still time until this technology gets widely accepted in stores around the world.

One of the major constraints involves high budgets, thus making it easy only for the bigger players such as Amazon, Walmart, Target etc to become early adopters of such technologies.

But having said that, it is time that retailers identify the potential that AI and Machine Learning can make to their business and take steps to make the shift soon.

If you would like to build an AI-powered solution for your retail business, then drop us a short message with your requirements

Designing for Artificial Intelligence

A new age design manifesto needs to grow beyond its traditional scope. As an outcome, it needs to factor in the growing space and time ecosystem enabled by new age technologies like Artificial intelligence (AI) and robotics. Hence the design principles of the past have evolved significantly to include new policies and principles.

Interested to explore what these new age design principles for AI are? Then read on..

1. It solves a real life human problem

The modern day design principles around AI should focus on solving a specific human problem. Going beyond the buzz and hype, a well-designed AI system has to concentrate on resolving a human problem (for e.g. delivering true value in service or product). The intent has to come out clearly when looking at using design to increase the value proposition of the AI system.

An example is the supportive body suit from SuperFlex. It mimics the natural body and muscle movement and helps out elderly people who have muscle or bone issues preventing them from carrying out routine tasks like moving hands or standing up.

2. Design for AI need not follow historical context

As mentioned earlier and as depicted by the emerging design trends, design need not conform to historical context. With new technologies it is obvious that design too needs to move beyond what we have experienced in the past and open our eyes to something totally new. This is essential when designers are working on truly ‘smart’ objects and should not be limited to just AI-based robots.

3. Design needs to understand the utility value of AI

AI was never designed or promoted to replace humans. It was instead designed to add value to human lives and make it more efficient/productive. If designers keep this basic difference in mind the resultant AI system would have better success potential in the market. When you are brainstorming for design ideas you need to ask yourself “Can AI  complement human lives rather than replicate it?”

4. Good AI Design needs to help everyone

A smart AI based product needs to be embraced equally by the tech lover and the senior persons of the family. Typically it is seen that one person who brings the system inside the house loves it while others aren’t easily swayed by its prowess. Designers need to figure out how they can have the entire household to get to use the product and derive benefits from it.

5. Good AI design doesn’t get in your way

A designer needs to understand that the AI product has to be subtle and discreet in its functioning so that it delivers a stellar experience without getting in your way. Such a well-designed AI system needs to generate subtle signals about the action being performed without disturbing the activity that you are doing.

August Smart Lock is one good example here.  It allows the user to unlock the door automatically when he/ she is nearby. You need not stop to take out keys from the bag or retrieve the smartphone from your pocket to unlock the door.

To conclude

With these principles in place, designers for AI systems will be in a truly remarkable place in the near future. This will be important as design will definitely be playing an increasingly influential role in building complex Artificial Intelligence solutions and systems.

Artificial Recruitment

In its most basic form, Artificial Intelligence is a computer system designed to learn, make decisions and carry out tasks that would normally require human intervention.

Speech recognition software, self-driving cars, chatbots that talk to the public and manufacturing robots all rely on Artificial Intelligence.

In the future, perhaps we can look forward to robot butlers that can cook and clean, police that patrol the streets 24–7, robotic friends and if you believe Tesla boss Elon Musk, an eventual Terminator-style apocalypse where the machines become self-aware and decide to wipe us out…

You’ve also probably read the headlines about artificial intelligence and how “robots are going to take all of our jobs” one day. Administrators, production line workers, customer service reps, professional drivers and perhaps even surgeons are all set to be casualties of the machine learning age.

You need that gut instinct and judgement to know whether someone will actually fit into the team.

And we all know hard skills aren’t everything. We also need to assess cultural fit, personality and soft skills like confidence and emotional intelligence. Machines can’t do that.

Plus, if you were a job-seeker, you’d not  like to get interviewed and hired by a machine…

What AI can and will do, however, is help us track down great candidates, faster!

We’re not quite there yet, but that is the future of AI-powered recruitment.

Here’s what AI will be able to do.

• Scour the internet to find great candidates.
• Make contact with them.
• Conduct first stage interviews. (Automated video interviewing already exists.)
• Help eliminate bias from the process.
• Standardise interviews and CV assessment.

Obviously, you will still have to make a final decision on the right hire for you… but if artificial intelligence can manage the process up until that point, imagine how much time you could save?!

The good news is that AI is already here in its most basic form.

The systems are clunky, but these are simply beta versions. The AI systems will learn fast, so don’t judge them too harshly right now!

Here are three example.

1. Beamery

Beamery is a candidate relationship management system that uses machine learning to enable proactive recruitment, “build talent pools, power collaboration and drive better decisions with predictive analytics.”

The start-up works with Facebook, among others, and analyses interactions between candidates and employers to identify candidates you should target and helps recruiters to build relationships with them.

2. Mya

You can encourage the right candidates to apply in the first place with the help of Mya, which parent company FirstJob claims will automate approximately 75% of the recruitment process.

It’s a combination of a chatbot at the front end, which effectively answers queries and gives feedback through messenger apps like Facebook, and an AI-powered search at the backend.

That search tool can eliminate irrelevant resumes and help find the needle in the haystack that is the perfect candidate.

If there’s information missing, the chatbot can get in touch and ask the right questions, thanks to Natural Language Processing, and plug the gaps in a resume that you might have rejected previously.

If Mya gets stuck, it refers the question to your HR department, but it can save you a huge amount of time and potentially rescue an application that simply did not cover all the bases.

It also learns, based on its past conversations, so the system will get better with time.

3. ThisWay Global

ThisWay “does in seconds, what it takes an HR professional 40 hours to do” apparently.

It’ll track down the most skilled candidates for your business, whilst removing any unconscious (or conscious) bias from the process, altogether.

And it also takes into consideration personality, culture, goals and motivations.

The Depth Of Deep Learning

Artificial Intelligence (AI) and Machine Learning (ML) are some of the hottest topics right now.

The term “AI” is thrown around casually every day. You hear aspiring developers saying they want to learn AI. You also hear executives saying they want to implement AI in their services. But quite often, many of these people don’t understand what AI is.

Once you’ve read this article, you will understand the basics of AI and ML. More importantly, you will understand how Deep Learning, the most popular type of ML, works.

The first step towards understanding how Deep Learning works is to grasp the differences between important terms.

Artificial Intelligence vs Machine Learning 

Artificial Intelligence is the replication of human intelligence in computers.

When AI research first started, researchers were trying to replicate human intelligence for specific tasks — like playing a game.

They introduced a vast number of rules that the computer needed to respect. The computer had a specific list of possible actions, and made decisions based on those rules.

Machine Learning refers to the ability of a machine to learn using large data sets instead of hard coded rules.

ML allows computers to learn by themselves. This type of learning takes advantage of the processing power of modern computers, which can easily process large data sets.

Supervised learning vs unsupervised learning

Supervised Learning involves using labelled data sets that have inputs and expected outputs.

When you train an AI using supervised learning, you give it an input and tell it the expected output.

If the output generated by the AI is wrong, it will readjust its calculations. This process is done iteratively over the data set, until the AI makes no more mistakes.

An example of supervised learning is a weather-predicting AI. It learns to predict weather using historical data. That training data has inputs (pressure, humidity, wind speed) and outputs (temperature).

Unsupervised Learning is the task of machine learning using data sets with no specified structure.

When you train an AI using unsupervised learning, you let the AI make logical classifications of the data.

An example of unsupervised learning is a behavior-predicting AI for an e-commerce website. It won’t learn by using a labelled data set of inputs and outputs.

Instead, it will create its own classification of the input data. It will tell you which kind of users are most likely to buy different products.

You’re now prepared to understand what Deep Learning is, and how it works.

Deep Learning is a machine learning method. It allows us to train an AI to predict outputs, given a set of inputs. Both supervised and unsupervised learning can be used to train the AI.

We will learn how deep learning works by building an hypothetical airplane ticket price estimation service. We will train it using a supervised learning method.

How Deep Learning can build an AI  to estimate Airplane ticket prices

We want our airplane ticket price estimator to predict the price using the following inputs (we are excluding return tickets for simplicity):

• Origin Airport
• Destination Airport
• Departure Date
• Airline

Like animals, our estimator AI’s brain has neurons. They are represented by circles. These neurons are interconnected.

The neurons are grouped into three different types of layers:

1. Input Layer
2. Hidden Layer(s)
3. Output Layer

The input layer receives input data. In our case, we have four neurons in the input layer: Origin Airport, Destination Airport, Departure Date, and Airline. The input layer passes the inputs to the first hidden layer.

The hidden layers perform mathematical computations on our inputs. One of the challenges in creating neural networks is deciding the number of hidden layers, as well as the number of neurons for each layer.

The “Deep” in Deep Learning refers to having more than one hidden layer.

The output layer returns the output data. In our case, it gives us the price prediction.

So how does it compute the price prediction?

This is where the magic of Deep Learning begins.

Each connection between neurons is associated with a weight. This weight indicates the importance of the input value. The initial weights are set randomly.

When predicting the price of an airplane ticket, the departure date is one of the heavier factors. Hence, the departure date neuron connections will have a big weight.

Each neuron has an Activation Function. These functions are hard to understand without mathematical reasoning.

Simply put, one of its purposes is to “standardize” the output from the neuron.

Once a set of input data has passed through all the layers of the neural network, it returns the output data through the output layer.

Nothing complicated, right?

Training the Neural Network

Training the AI is the hardest part of Deep Learning. Why?

1. You need a large data set.
2. You need a large amount of computational power.

For our airplane ticket price estimator, we need to find historical data of ticket prices. And due to the large amount of possible airports and departure date combinations, we need a very large list of ticket prices.

To train the AI, we need to give it the inputs from our data set, and compare its outputs with the outputs from the data set. Since the AI is still untrained, its outputs will be wrong.

Once we go through the whole data set, we can create a function that shows us how wrong the AI’s outputs were from the real outputs. This function is called the Cost Function.

Ideally, we want our cost function to be zero. That’s when our AI’s outputs are the same as the data set outputs. 

Thus at this juncture, with the help of deep learning we have almost trained the AI to project an output that is in line with the input data. This ladies and gentleman to be honest is just the basics, but deep learning along with precision is used to train the AI for supervised learning.

We hope that at this juncture you have some idea of how the different elements of Machine Learning, Deep Learning and Neuron networks all come together to create Artificial Intelligence.

Stay tuned for more interesting facts on AI and Machine Learning!

Language For The Machines

Machine learning is all about making a machine capable of learning codes and producing them automatically by analyzing the input data. This process includes the development of AI-enabled algorithms which help machine learn and produce codes. So, in a way we can say that machine learning is a part of AI (Artificial intelligence) which has been very viably used in many fields, like math, psychology, etc.

Well, this is just an intro; a more technicalities are still hidden far behind the curtains. The biggest one being learning of a machine language because only a good programmer can make use of all tools and bring out the best in the job.

So, in this article, we shall discuss about the 3 must-know languages which a programmer should know while dealing with machine learning.

1. Python

Python is the highly flexible and multi-purpose language in nature. Due to its these features, it has gained a lot of popularity among the developers, programmers and data scientists. This language has its own libraries for the purpose machine learning – Numpy and Scipy. These two libraries are enough to learn about the Linear Algebra and Kernel methods of the phenomenon of machine learning. The biggest ease with this language is that- it has easy syntax. Easy syntax, in turns, makes the whole machine learning process easy and understandable. Those who want to excel in machine learning, should start with python.

2. C Language

Developed by Denise Ritchie, C is the mother of all languages. Therefore, if you are thinking of building a predictive algorithm, this language will help you superficially. But as said, this language is the mother of all other programming language and hence, learning this would not be a cakewalk. To get started with C, you will have to have great fundamentals of basic C and its syntax. However, having mastery in this language does not mean to hold a PhD. It simply means to have strong concepts and clear fundamentals. Also, once you are nicely through the learning of C language, you can even give a try to other functional languages like Erlang, Scala, Julia, and Haskell.

3. R Language

R language is the modern version of the S language developed by the Bell Labs. This language, when combined with lexical scooping, helps in enhancing the flexibility of statistical models. In terms of machine learning, it is one of the strongest language to master. In this language, many GNU packages are available. This language can be used for creating the powerful algorithms and giving statistical visualization of those created algorithms. Currently, R language in very popularly used in educational industry, but soon it will exact famous in other fields as well.

So, as you can see the 3 most powerful languages for machine learning, it is time for you to decide which language you would start with. It would be a recommendation to start with C, which is the base of all other languages. After that, one can jump to python and R language.

AI Knows From A To Z

Target found out the pregnancy of a teenager before her parents did.

An angry father walks into a Target store in Minneapolis, demanding to talk to the manager:

“My daughter got this in the mail!” he said. “She’s still in high school, and you’re sending her coupons for baby clothes and cribs? Are you trying to encourage her to get pregnant?”

A few days later:

“I had a talk with my daughter,” he said. “It turns out there’s been some activities in my house I haven’t been completely aware of. She’s due in August. I owe you an apology.”

Target had a system which assigns each shopper a “pregnancy prediction” score based on the products they buy. The system could also estimate their due date to within a small window, so Target could send coupons timed to very specific stages of their pregnancy.

This happened in 2012 and it’s hardly state-of-the-art “AI”, but it just goes to show that anything creepy a machine learning model does, is just a product of how and with what data it is trained.

Programmer and CMU PhD Tom Murphy created a function to “beat” NES games by watching the score. How it worked was that the program would do things that increased the score, and then learn how to reproduce them again and again, resulting in high scores. It came up with novel techniques and strategies for playing games and even exploited glitches humans didn’t know about, or at least hadn’t told it about. The program, called a “technique for automating NES games,” can take on nearly every NES game. Nearly.

Tom made the program play Tetris. Most of us have played this game and needless to say, we all know that it gets tricky after a certain point. The program struggled to figure out what to do. The choices of Tetris blocks is entirely random, so it’s not surprising that the computer wasn’t able to consider future repercussions far enough ahead to notice that stacking those blocks in a certain ways made a big difference.

On one such run, when faced with imminent defeat, the computer did something eerie. Rather than lose, and receive a ‘game over’ message, it just paused the game. Forever.

Tom describes the computer’s reasoning like this: “The only winning move is to not play.” And that’s right. If you pause a game for ever you will never lose that game.

An Artificial Intelligence program developed by Elon Musk’s Team called Open AI created a lot of buzz as well. Musk believes that development of AI should be regulated and AI safety should be a prime concern of every developer. To put weight to his idea, he started a project called OpenAI. The team used DOTA 2 as a test means to develop their AI.

Now what’s special is how they trained this BOT. They didn’t write any code about the rules of Dota 2 or any strategies that professional players use. They just gave basic instructions(eg: Winning is good, losing is bad, Taking Damage is Bad, Giving Damage is good, etc) and made the BOT play with a copy of itself. In the beginning, the BOT made very stupid decisions. But slowly, it started to learn, devise its own strategies and make novel moves. It took the BOT 2 hours to beat the existing Dota 2 BOT and 2 weeks to reach the level of a professional Dota player!

Finally, OpenAI put its BOT to test against many of the world’s top Dota 2 players(1v1 match) and it was easily able defeat them. Then came The International 2017, one of the biggest eSports event in the world. Here, OpenAI was tested against what people consider the best Dota 2 player in the world: Danylo “Dendi” Ishutin. To everyone’s surprise, OpenAI defeated Dendi in a solid 2–0 before Dendi gave up!

A Few More Worthy Mentions 

The blink recognition software in Nikon’s camera kept asking “Did someone blink” when an Asian would pose in front of the camera. The camera perceived the *small* eyes of Asians as closed.

Recently a report announced that Facebook had to abandon their experiment after two AIs went out of control and supposedly started interacting with each other in a language other than English which made them easier to work! below is what they said to each other.

Bob: i can i i everything else . . . . . . . . . . . . . .

Alice: balls have zero to me to me to me to me to me to me to me to me to

Bob: you i everything else . . . . . . . . . . . . . .

Alice: balls have a ball to me to me to me to me to me to me to me

Bob: i i can i i i everything else . . . . . . . . . . . . . .

Alice: balls have a ball to me to me to me to me to me to me to me

Bob: i . . . . . . . . . . . . . . . . . . .

Alice: balls have zero to me to me to me to me to me to me to me to me to

Bob: you i i i i i everything else . . . . . . . . . . . . . .

Alice: balls have 0 to me to me to me to me to me to me to me to me to

Bob: you i i i everything else . . . . . . . . . . . . . .

Alice: balls have zero to me to me to me to me to me to me to me to me to

You, i everything else (dot)(dot)(dot)(dot), looks like Bob was devising a plan to kill everyone other than him and Alice.

And that ladies & gentleman is how Artificial Intelligence has evolved over the years.

Fun fact, it always is on the move. It is always evolving.