Use Cases of Machine Learning

Hello connections!!

I am back with another article on Machine Learning about Use Cases of  Machine learning.

What is Machine Learning?

Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it learn for themselves.

The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that we provide. The primary aim is to allow the computers learn automatically without human intervention or assistance and adjust actions accordingly.

Some machine learning methods

Machine learning algorithms are often categorized as supervised or unsupervised.

• Supervised machine learning algorithms can apply what has been learned in the past to new data using labeled examples to predict future events. Starting from the analysis of a known training dataset, the learning algorithm produces an inferred function to make predictions about the output values. The system is able to provide targets for any new input after sufficient training. The learning algorithm can also compare its output with the correct, intended output and find errors in order to modify the model accordingly.
• In contrast, unsupervised machine learning algorithms are used when the information used to train is neither classified nor labeled. Unsupervised learning studies how systems can infer a function to describe a hidden structure from unlabeled data. The system doesn’t figure out the right output, but it explores the data and can draw inferences from datasets to describe hidden structures from unlabeled data.
• Semi-supervised machine learning algorithms fall somewhere in between supervised and unsupervised learning, since they use both labeled and unlabeled data for training – typically a small amount of labeled data and a large amount of unlabeled data. The systems that use this method are able to considerably improve learning accuracy. Usually, semi-supervised learning is chosen when the acquired labeled data requires skilled and relevant resources in order to train it / learn from it. Otherwise, acquiring unlabeled data generally doesn’t require additional resources.
• Reinforcement machine learning algorithms is a learning method that interacts with its environment by producing actions and discovers errors or rewards. Trial and error search and delayed reward are the most relevant characteristics of reinforcement learning. This method allows machines and software agents to automatically determine the ideal behavior within a specific context in order to maximize its performance. Simple reward feedback is required for the agent to learn which action is best; this is known as the reinforcement signal.

1. Machine Learning Use Cases in Transportation

The application of machine learning in the transport industry has gone to an entirely different level in the last decade. This coincides with the rise of ride-hailing apps like Uber, Lyft, Ola, etc.

These companies use machine learning throughout their many products, from planning optimal routes to deciding prices for the rise we take. So let’s look at a few popular use cases in transportation which use machine learning heavily.

Dynamic Pricing in Travel

Do you often get frustrated by the surge pricing cab-hailing companies use? I encounter it on a daily basis for my commute to and from work. Prices seem to be perpetually hiked. Why is this happening?!

So I dug into this a bit more and came across the concept of dynamic pricing – an excellent machine learning use case. To understand this, let’s take a simple example.

Imagine you are starting a ride-hailing business. You need to plan the ride prices for each route in the city in a way that would attract customers but also improve your bottomline. One way to do it is to manually map prices to each route. Not an ideal solution.

This is where dynamic pricing plays a vital role. It means adjusting your prices to changing market conditions. So, prices vary depending on factors like location, time of day, weather, overall customer demand, etc. That is the underlying idea behind why surge pricing was introduced.

Dynamic pricing is a thriving practice in various industries, such as travel, hospitality, transportation, and logistics, among others.

Transportation and Commuting – Uber

Dynamic pricing is not the only machine learning use case ride-hailing companies like Uber use. They rely heavily on machine learning to identify the most optimal route to get the passenger from point A to B.

For us, it appears to be a rather simple solution. Put your location, the destination and the nearest driver will come to pick us up. But what appears to be straightforward is actually a complex web of architectures and services on Uber’s backend.

There are multiple machine learning techniques at play that aim to optimize the route we take.

Google Maps

You must have guessed this one by now. Google Maps is a prime example of a machine learning use case. In fact, I would recommend opening up Google Maps right now and picking out the different features it offers.

Here are some that I can see (and have used extensively):

• Routes: Go from point A to point B
• Estimated time to travel this route
• Traffic along the route
• The ‘Explore Nearby’ feature: Restaurants, petrol pumps, ATMs, Hotels, Shopping Centres, etc.

Google uses a ton of machine learning algorithms to produce all these features. Machine learning is deeply embedded in Google Maps and that’s why the routes are getting smarter with each update.

The estimated travel time feature works almost perfectly. If it shows ’40 minutes’ to reach your destination, you can be sure your travel time will be approximately around that timeline. Got to love machine learning!

2. Machine Learning Use Cases in Smartphones

Did you know that machine learning powers most of the features on your smartphone?

That’s right! From the voice assistant that sets your alarm and find you the best restaurants to the simple use case of unlocking your phone via facial recognition – machine learning is truly embedded in our favorite devices.

Voice Assistants

That example we saw in the introduction about talking to our virtual assistant? That was all about the concept of speech recognition – a budding topic in machine learning right now.

Voice assistants are ubiquitous right now. You must have used (or at least heard about) the below popular voice assistants:

• Apple’s Siri
• Google Assistant
• Amazon’s Alexa
• Google Duplex
• Microsoft’s Cortana
• Samsung’s Bixby

And so on. The common thread between all these voice assistants? They are powered by machine learning algorithms! These voice assistants recognize speech (the words we say) using Natural Language Processing (NLP), convert them into numbers using machine learning, and formulate a response accordingly.

Smartphone Cameras

Wait – what in the world does machine learning have to do with my smartphone camera? Quite a lot, as it turns out.

The incredible images we can click these days and the depth of these images – all of that is thanks to machine learning algorithms. They analyze every pixel in each image to detect objects, blur the background, and a whole host of tricks.

These machine learning algorithms do several things to improve and enhance the smartphone’s camera:

• Object detection to locate and single out the object(s) (or human) in the image
• Filling in the missing parts in a picture
• Using a certain type of neural network using GANs to enhance the image or even extend its boundaries by imagining what the image would look like, etc.

Face Unlock – Smartphones

Most of us have are quite familiar with this. We pick up our smartphone and it unlocks itself by detecting our face. It’s smart, efficient, timesaving and frankly superb.

What a lot of people don’t know about this is that our smartphones use a technique called facial recognition to do this. And the core idea behind facial recognition is powered by – you guessed it – machine learning.

The applications of facial recognition are vast and businesses around the world are already reaping the benefits:

• Facebook uses it to identify the people in images
• Governments are using it to identify and catch criminals
• Airports are using to verify passengers and crew members, and so on

3. Machine Learning Use Cases in Popular Web Services

You’ll love this section. We interact with certain applications every day multiple times. What we perhaps did not realize until recently – most of these applications work thanks to the power and flexibility of machine learning.

Here are four use cases you are ultra familiar with. Now, look at them from a machine learning perspective.

Email filtering

Dealing with way too emails at work? Or is your personal email inbox bursting with utterly random and spam emails? We have all been there. My inbox counts once read 11,000+ unread emails!

Wouldn’t it be easy if we could write a rule that would filter emails according to their subject? A marketing mail would go to that folder. An email about work would come into my primary inbox (and so on). This would make life so much easier.

As it turns out, this is exactly what most email services are now doing! They’re using machine learning to parse through the email’s subject line and categorize it accordingly. Take Gmail for example. The machine learning algorithm Google uses has been trained on millions of emails so it can work seamlessly for the end-user (us).

While Gmail allows us to customize labels, the service offers default labels:

• Primary
• Social
• Promotions

The machine learning algorithms immediately categorize the email into one of these three labels as soon as you receive an email. We get an instant alert if Gmail deems it a ‘Primary’ email.

Of course, Gmail also uses machine learning to figure out if the email is spam or not. A feature we are all truly grateful for. Google’s algorithm has become a lot smarter over the years in deciding if an email is spam or not. This is where getting more data for a machine learning algorithm is so helpful – something Google has in abundance.

Google Search

The most popular machine learning use case in this (or any) list. Everyone has used Google Search and most of us use it multiple times on a daily basis. I would venture to say we take it for granted that Google will serve us the best results up front.

But how does Google Search work?

Google Search has become an impenetrable behemoth that mortals cannot crack. How it works underneath is something only those folks who have designed Google Search know. One thing we can say for certain – Google uses machine learning to power its Search engine.

The amount of data Google has to constantly train and refine its algorithms is a number we cannot fathom. No calculator in the world will tell us the number of queries Google has processed in the last two decades. It is a treasure trove for data scientists!

Google Translate

I’m fluent in Google Translate. I’ve picked up bits and pieces of foreign languages like German, Spanish, and Italian thanks to this wonderful service by Google. Anytime I come across a bit of text in a foreign language, Google Translate immediately offers me the answer.

It won’t surprise you to know that Google uses machine learning to understand the sentence(s) sent by the user, convert them to the requested language, and show the output. Machine learning is deeply embedded in Google’s ecosystem and we are all benefitting from that.

LinkedIn and Facebook recommendations and ads

Social media platforms are classic use cases of machine learning. Like Google, these platforms have integrated machine learning into their very fabric. From your home feed to the kind of ads you see, all of these features work thanks to machine learning.

A feature which we regularly see if ‘People you may know’. This is a common feature across all social media platforms, Twitter, Facebook, LinkedIn, etc. These companies use machine learning algorithms to look at your profile, your interests, your current friends, their friends, and a whole host of other variables.

The algorithm then generates a list of people that match a certain pattern. These people are then recommended to you with the expectation that you might know them (or at least have profiles very similar to yours).

I have personally connected with a lot of my professional colleagues and college friends thanks to LinkedIn’s system. It’s a use case of machine learning benefitting everyone involved in the process.

The ads that we see work in a similar fashion. They are tailored to your tastes, interests and especially your recent browsing or purchase history. If you are a part of a lot of data science groups, Facebook or LinkedIn’s machine learning algorithm might suggest machine learning courses.

Pay attention to this next time you’re using social media. It’s all machine learning behind the curtains!

4.Google's DeepDream

We all know that humans dream? Well, what if computers dream as well?!! This is the premise of Google DeepDream that used convolutional neural networks to find random patterns in various images and amplifies them in different ways. These images can be tweaked in any possible manner using the input data and various parameters so that the results obtained can be funny, weird or even trippy!!!

There are multiple layers in the neural networks in DeepDream wherein each layer extracts more and more high-level features from the input image until the final output is produced by the end layer. To demonstrate this, we have an image from Google DeepDream that is a weird hybrid of a woman and lots of gears. All in all, it’s very difficult to just explain the complicated effects of DeepDream so its best that you just try it yourself by uploading any image you want and then just watching the show!