ARTH - Task 20 👨🏻💻

Task Description📄

✍🏻 Research for industry use-cases of Neural Networks and create a blog, Article, or Video elaborating how it works.

Hello, Connections !!!

This is an important article on Neural Networks and their industrial use-cases. In this article, you will get to know about Neural Networks in detail.

Neural Networks have become a success in the recent Machine Learning craze due to their significantly better performance than traditional Machine Learning algorithms in many cases. Today, neural networks (NN) are revolutionizing business and everyday life, bringing us to the next level in artificial intelligence (AI). By emulating the way interconnected brain cells function, NN-enabled machines (including the smartphones and computers that we use on a daily basis) are now trained to learn, recognize patterns, and make predictions in a humanoid fashion as well as solve problems in every business sector.

The art and science of Deep Learning are built on the foundation of Neural Networks and how they work. Hence demystifying Neural Networks is going to be the first step in demystifying Deep Learning. Let’s dive in!

A Brief History of Neural Networks

Neural networks date back to the early 1940s when mathematicians Warren McCulloch and Walter Pitts built a simple algorithm-based system designed to emulate human brain function. Work in the field accelerated in 1957 when Cornell University’s Frank Rosenblatt conceived of the perceptron, the groundbreaking algorithm developed to perform complex recognition tasks. During the four decades that followed, the lack of computing power necessary to process large amounts of data put the brakes on advances. In the 2000s, thanks to the advent of greater computing power and more sophisticated hardware, as well as to the existence of vast data sets to draw from, computer scientists finally had what they needed, and neural networks and AI took off, with no end in sight. To understand how much the field has expanded in the new millennium, consider that ninety percent of internet data has been created since 2016 That pace will continue to accelerate, thanks to the growth of the Internet of Things (IoT).

What Are Neural Networks?

An artificial neural network learning algorithm, or neural network, or just neural net, is a computational learning system that uses a network of functions to understand and translate a data input of one form into the desired output, usually in another form. The concept of the artificial neural network was inspired by human biology and the way neurons of the human brain function together to understand inputs from human senses. 

It’s not a very accurate representation but it tries to replicate some of the methods our brain uses to learn from its mistakes. Let’s look at how our brains work from a simplified perspective and then compare it with a Neural Network.

Brain Vs Neural Networks

The brain is essentially a bunch of neurons connected to each other in a huge interconnected network. There are a lot of neurons and even more connections. These neurons pass a small amount of electrical charge to each other as a way to transmit information. Another important feature of these neural connections is that the connection between two neurons can vary between strong and weak. A strong connection allows more charge to flow between them and a weak one allows lesser. A neuron pathway that frequently transmits charge will eventually become a strong pathway.

Now as the brain takes input from any external source, let’s say for example we touch a hot glass. The nerves from our hands transmit signals to certain neurons in our brain. Now there is a pathway from these neurons to the neurons which control our hand. And in these cases, our brain has learned that the best option is to move our hand from the glass ASAP. Hence this certain pathway between the neurons taking input from the hand and the neurons controlling the hand will be strong.

Neural pathways become stronger upon frequent usage, and our brain essentially tries to use pathways that have proven to give us better results over time. So essentially as we humans live our lives and decide whether our actions are good or bad, we are training our brain to make sure we don’t repeat our previous mistakes or keep doing things that we think resulted in a good outcome.

How Does a Neural Network Work?

Machine learning algorithms that use neural networks generally do not need to be programmed with specific rules that define what to expect from the input. The neural net learning algorithm instead learns from processing many labeled examples (i.e. data with "answers") that are supplied during training and using this answer key to learn what characteristics of the input are needed to construct the correct output. Once a sufficient number of examples have been processed, the neural network can begin to process new, unseen inputs and successfully return accurate results. The more examples and variety of inputs the program sees, the more accurate the results typically become because the program learns with experience.  

This concept can best be understood with an example. Imagine the "simple" problem of trying to determine whether or not an image contains a dog. While this is rather easy for a human to figure out, it is much more difficult to train a computer to identify a dog in an image using classical methods. Considering the diverse possibilities of how a dog may look in a picture, writing code to account for every scenario is almost impossible. But using machine learning, and more specifically neural networks, the program can use a generalized approach to understanding the content in an image. Using several layers of functions to decompose the image into data points and information that a computer can use, the neural network can start to identify trends that exist across the many, many examples that it processes and classify images by their similarities. 

After processing many training examples of dog images, the algorithm has a model of what elements, and their respective relationships, in an image are important to consider when deciding whether a dog is present in the picture or not. When evaluating a new image, the neural net compares the data points about the new image to its model, which is based on all previous evaluations. It then uses some simple statistics to decides whether the image contains a dog or not based on how closely it matches the model.

In this example, the layers of functions between the input and the output are what make up the neural network. In practice, the neural network is slightly more complicated than the image above shows. The following image captures the interaction between layers slightly better, but keep in mind that there are many variations of the relationships between nodes, or artificial neurons:

How Do You Train a Neural Network?

Once you’ve structured a network for a particular application, training (i.e., learning), begins. There are two approaches to training. 

Supervised learning provides the network with desired outputs through manual grading of network performance or by delivering desired outputs and inputs.

 Unsupervised learning occurs when the network makes sense of inputs without outside assistance or instruction.

There’s still a long way to go in the area of unsupervised learning. “Getting information from unlabeled data, [a process] we call unsupervised learning, is a very hot topic right now, but clearly not something we have cracked yet. It’s something that still falls in the challenging column,” observes Université de Montréal’s Yoshua Bengio in the article “The Rise of Neural Networks and Deep Learning in Our Everyday Lives.”

Bengio is referring to the fact that the number of neural networks can’t match the number of connections in the human brain, but the former’s ability to catch up maybe just over the horizon. Moore’s Law, which states that overall processing power for computers will double every two years, gives us a hint about the direction in which neural networks and AI are headed. Intel CEO Brian Krzanich affirmed at the 2017 Computer Electronics Show that “Moore’s Law is alive and well and flourishing.” Since its inception in the mid-20th century, neural networks’ ability to “think” has been changing our world at an incredible pace.

Why Do We Use Neural Networks?

Neural networks’ human-like attributes and ability to complete tasks in infinite permutations and combinations make them uniquely suited to today’s big data-based applications. Because neural networks also have the unique capacity (known as fuzzy logic) to make sense of ambiguous, contradictory, or incomplete data, they are able to use controlled processes when no exact models are available.

Attributes of Neural Networks

With the human-like ability to problem-solve — and apply that skill to huge datasets — neural networks possess the following powerful attributes:

• Adaptive Learning: Like humans, neural networks model non-linear and complex relationships and build on previous knowledge. For example, software uses adaptive learning to teach math and language arts.
• Self-Organization: The ability to cluster and classify vast amounts of data makes neural networks uniquely suited for organizing the complicated visual problems posed by medical image analysis.
• Real-Time Operation: Neural networks can (sometimes) provide real-time answers, as is the case with self-driving cars and drone navigation.
• Prognosis: NN’s ability to predict based on models has a wide range of applications, including for weather and traffic.
• Fault Tolerance: When significant parts of a network are lost or missing, neural networks can fill in the blanks. This ability is especially useful in space exploration, where the failure of electronic devices is always a possibility.

Types of neural networks

There are different kinds of deep neural networks – and each has advantages and disadvantages, depending upon the use. Examples include:

• Convolutional neural networks (CNNs) contain five types of layers: input, convolution, pooling, fully connected, and output. Each layer has a specific purpose, like summarizing, connecting, or activating. Convolutional neural networks have popularized image classification and object detection. However, CNNs have also been applied to other areas, such as natural language processing and forecasting.
• Recurrent neural networks (RNNs) use sequential information such as time-stamped data from a sensor device or a spoken sentence, composed of a sequence of terms. Unlike traditional neural networks, all inputs to a recurrent neural network are not independent of each other, and the output for each element depends on the computations of its preceding elements. RNNs are used in fore­casting and time series applications, sentiment analysis, and other text applications.
• Feedforward neural networks, in which each perceptron in one layer is connected to every perceptron from the next layer. Information is fed forward from one layer to the next in the forward direction only. There are no feedback loops.
• Autoencoder neural networks are used to create abstractions called encoders, created from a given set of inputs. Although similar to more traditional neural networks, autoencoders seek to model the inputs themselves, and therefore the method is considered unsupervised. The premise of autoencoders is to desensitize the irrelevant and sensitize the relevant. As layers are added, further abstractions are formulated at higher layers (layers closest to the point at which a decoder layer is introduced). These abstractions can then be used by linear or nonlinear classifiers.

Tasks Neural Networks Perform

Neural networks are highly valuable because they can carry out tasks to make sense of data while retaining all their other attributes. Here are the critical tasks that neural networks perform:

• Classification: NNs organize patterns or datasets into predefined classes.
• Prediction: They produce the expected output from the given input.
• Clustering: They identify a unique feature of the data and classify it without any knowledge of prior data.
• Associating: You can train neural networks to "remember" patterns. When you show an unfamiliar version of a pattern, the network associates it with the most comparable version in its memory and reverts to the latter.

Neural networks are fundamental to deep learning, a robust set of NN techniques that lends themselves to solving abstract problems, such as bioinformatics, drug design, social network filtering, and natural language translation. Deep learning is where we will solve the most complicated issues in science and engineering, including advanced robotics. As neural networks become smarter and faster, we make advances on a daily basis.

Key advantages of Neural Networks:

ANNs have some key advantages that make them most suitable for certain problems and situations:

1. ANNs have the ability to learn and model non-linear and complex relationships, which is really important because in real-life, many of the relationships between inputs and outputs are non-linear as well as complex.

2. The input is stored in its own networks instead of a database, hence the loss of data does not affect its working.

3. These networks can learn from examples and apply them when a similar event arises, making them able to work through real-time events.

4. ANNs can generalize — After learning from the initial inputs and their relationships, it can infer unseen relationships on unseen data as well, thus making the model generalize and predict on unseen data.

5. Even if a neuron is not responding or a piece of information is missing, the network can detect the fault and still produce the output.

6. They can perform multiple tasks in parallel without affecting the system performance.

7. Unlike many other prediction techniques, ANN does not impose any restrictions on the input variables (like how they should be distributed). Additionally, many studies have shown that ANNs can better model heteroskedasticity i.e. data with high volatility and non-constant variance, given its ability to learn hidden relationships in the data without imposing any fixed relationships in the data. This is something very useful in financial time series forecasting (e.g. stock prices) where data volatility is very high.

Neural Networks Case Study: Waymo, Google’s self-driving car division

Waymo has driven more than 20 million miles on public roads in over 25 cities. They also drove tens of billions of miles in simulations. Additionally, Waymo is operating a taxi service in the United States, transporting passengers for real without a driver.

As with every self-driving vehicle, Waymo implements its tech using 4 main steps: perception, localization, planning, and control.

• Perception is about finding obstacles, traffic lights, and roads. Waymo uses active learning to collect data, and AutoML to generate architectures and select the more efficient one (accuracy and inference time).
• Localization is mainly perception with the task of finding where you are. Waymo is leveraging the knowledge of Google Maps to do so.
• Prediction is using recurrent neural networks and reinforcement learning in simulators to train their agents to estimate trajectories really well.
• Planning is generating trajectories based on feasibility, staying on the road, and avoiding collisions. The vehicles also learn from human labelers to generate more realistic trajectories.

Waymo’s system is the result of 11 years of research and experimentation in self-driving cars. In the autonomous tech world, their way of building cars has some enemies, as people have voiced a preference for Tesla’s system and realized that on-the-road experience is very valuable.

Whatever your opinion may be, you can’t deny the insane level of work and tech Google and Waymo have invested in their autonomous vehicle efforts.

The future of AI at Waymo isn’t sentient vehicles. (Sorry, Knight Rider fans.) It’s in cutting-edge research like automated machine learning, in which the process of building machine learning models is automated. “Essentially, the idea that you have AI machine learning that’s creating other AI models that actually solve the problem you’re trying to solve,” Arnoud says.

This becomes extremely useful for driving in areas with unclear lane markings. These days, the most challenging driving environments require self-driving cars to make guidance decisions without white lines, Botts Dots, or clear demarcations at the edge of the road. If Waymo can build machine learning models to train its neural nets to drive on streets with unclear markings, then Waymo’s self-driving cars can put the Phoenix suburbs in its rearview and eventually hit the open road. 

Waymo still has a long way to go…

One of their main problems is the way they use Maps: Waymo can’t drive without a map. They can map the whole world more precisely, but this is a huge challenge for scale.

Waymo’s principal vision system being made of LiDARs is also actually a problem—LiDARs are completely blind during snow, rain, or fog.

Consequentially, Waymo drives a lot in places like Phoenix, Arizona, or San Francisco, California, where the conditions are perpetually dry and sunny.

You did an amazing job reading all of this, and I hope you understand Waymo a bit better! 

Other Use-cases in the field of Neural networks and deep learning are :

Cybersecurity

ANNs can also be used to protect organizations from several types of attacks, such as DDoS and malicious software. The malware itself is a huge problem, with at least 325,000 new malicious files being generated every day. Yet, no more than 10 percent of the files change from iteration to iteration, so algorithm-based learning models that can predict these variations are able to detect which files are malware with amazing accuracy.

AI is better than humans at cybersecurity because they automate the most complex processes required for detecting attacks and analyzing the best way to react to breaches. More in general, neural nets could be used to detect any change or anomaly in network traffic, including the newest 5G networks. AI can avoid the risk of false positives and identify potentially malicious activities such as brute-force attacks, unusual failed logins and file exfiltration with some experiments reaching a 96.4% detection rate.

Obviously, hackers started developing their own adaptive AI to deceive security software and exploit vulnerabilities, in a never-ending arms race between attackers and defenders. However, all of this actually benefits AIs, which get smarter and smarter every day they are deployed in the battlefield. (To see how AI is fighting crime in the real world, see How AI Is Helping in the Fight Against Crime.)

Fighting Against Future Pandemics

The 2020 COVID-19 pandemic shocked the whole world, forcing us to rethink our entire society and face a cataclysm that no one ever expected. Whether this specific viral pandemic is over not, there’s one thing that we learned the hard way: that we cannot be found unprepared anymore in the future if another biological threat emerges.

AI technologies can help us (and already helped us) in many different ways. Complex neural networks can be used to coordinate the efforts of multiple cameras simultaneously and send a warning whenever a person with elevated temperature is found. The AI can take immediate action such as stopping the person from accessing critical places and ensure better workplace safety.

Radiological imaging can be coupled with advanced AI to allow for immediate recognition of X-ray images suggesting that a patient is affected by the disease, even in remote or less served areas. Articulated machine learning models can be leveraged to determine the impact of quarantine measures to minimize the effects of restrictive measures on national economy.

Business and Advertising

This can be summed up in just one word (well... three): personalized product recommendations. Every time we search for something on Google or any other search engine, eventually we start seeing a ton of precisely targeted ads about these things. How could the software understand so well what our interests are and how to entice us into buying those extremely cheap goods we want so badly?

Once again, deep learning is the answer. These highly reactive programs learn by watching our behaviors, such as when we skip paging two of the search results when none of those found on page one satisfied our needs. Machines can crunch demographic data about customers' habits and preferences at a speed that no human analyst can ever hope to reach, and can consume it to optimize pricing, offers, customer experience, and profitability. It should not surprise anyone that one of the biggest lovers of AIs and smart algorithms is none other than Amazon itself.

Yet, the retail giant is using advanced heuristics to optimize its services in many other ways. One of the reasons why Jeff Bezos' creature is so successful, is, in fact, the amazing efficiency of its logistics planning. Other giants such as Walmart and Honda as well as many small-to-medium businesses and factories vastly improved their efficiency by implementing machine learning in the management of orders, stocking, inventory control, and warehousing.

AIs are incredibly good at detecting quality issues inside and beyond the assembly line, for instance by identifying patterns in the free-text fields of warranty registration cards.

Conclusion

Deep learning is rapidly transforming many industries including healthcare, energy, fintech, transportation, and many others, to rethink traditional business processes with digital intelligence. Early adopter industries have witnessed a profound effect on the workplace and great potential in terms of developing deep learning applications, which can be used for yielding forecasts, detecting fraud, attracting new customers, and so much more. The opportunities and capabilities are substantial and that’s why many enterprises are investing in deep learning for building out their existing applications as well as developing new solutions.

So, here we learned that how an organization like WAYMO LLC uses Neural Networks to solve the challenges.

Thanks for reading the article.

Hope this article helpful to you!!!!

You can appreciate the article by giving it a like and posting comments about your feedback.