Deep Learning in Practice: 5 Use Cases for Deep Learning Algorithms

from | 8 May 2020 | Basics

Deep Learning is a data science method that plays a major role especially in the context of Artificial Intelligence and Machine Learning. More and more applications in everyday life, such as facial recognition software in smartphone cameras, use this method. In this blog article, we explain the high relevance of Deep Learning using five practical use cases.  

What is Deep Learning?

Deep learning is a method of learning that is special class of optimisation methods of artificial neural networks. This is why they are sometimes called "deep neural networks". The main difference is the complexity of the intermediate layers, the so-called "hidden layers".

Deep Learning has become one of the central development drivers in the field of artificial intelligence in recent years for two reasons: firstly, because it achieves particularly good results when large amounts of data (Big Data) are available with which neural networks are trained. And secondly, because deep-learning algorithms have made intellectual and mental processes representable that were long assumed to be reserved for humans.

a deep learning algorithm or a deep neural network
Source: rsipvision.com 

In a deep learning algorithm or deep neural network, there are numerous hidden layers between the input and output layers.

Two of the most prominent Examples are Speech and face recognition. Siri, Cortana & Co, Chatbots or the new Google Image Search are examples of applications that would not exist without DL. The algorithms of chatbots, for example, learn with every question they are asked and thus improve themselves. It is precisely this learning ability of deep-learning algorithms that distinguishes them from "normal" chatbots. Artificial neural networks off.

Explanation example

On film"Her", Spike Jonze confronts his viewers with a form of Artificial intelligencethat people can not only talk to in a very natural way, but that you can even fall in love with. Deep Learning is a central key that can make it possible for us in the future to actually interact with digital personalities can.

YouTube

By loading the video you accept YouTube's privacy policy.
Learn more

load Video

Because anyone who tries to have a conversation with digital assistants like Siri or Alexa today will notice how quickly the limits of what is possible are reached here. Understanding and imitating human speech is still one of the greatest challenges for computers. At the same time, the progress being made in this area is enormous.

Deep Learning Study
© [at]

How deep learning algorithms can understand language and images

0 and 1, yes or no - that is the binary base operationon which the entire IT is based. In order to understand language or images, extremely many shades of grey, ambivalences and complex processes of understanding are necessary. That is why DL is a promising path that could make it possible to achieve exactly that in the future.

One of the great strengths of these algorithms has been demonstrated in recent years, for example, in the area of Image recognition and video analysis. The learning ability of this class of Algorithms enabled them to continuously learn to understand the content of images. The following illustration shows in simplified form how a deep learning algorithm "sees" something.

Image recognition with Deep Learning
Deep-learning algorithms break down images into the smallest meaning-distinguishing components and thus learn to recognise displayed content. Source: spektrum.de

Example 1: Analysis of image data for disease diagnoses

One of the most prominent areas of application of deep learning algorithms is the field of image recognition. Impressive progress has been made here in recent years, especially in the field of medicine. These algorithms can be trained with image data of the most varied kinds - this opens up completely different areas of application and algorithms can be trained on them, Examine X-rays or CT scans for abnormalities.

In this way, they can assist doctors in Support diagnosis of diseases. Because even if specialists have many years of experience, they can never sift through the same amount of image data used in training. Data sets used to train deep learning algorithms can comprise many millions of images. No wonder, then, that intelligent analysis programmes based on Deep Learning see better than humans.

Example 2: Sales and aftersales

In the area of sales and aftersales, Deep Learning uses speech and sentiment analysis for Improving the customer experience. But not only images in static or even moving form can be used as a basis for training deep learning algorithms. Intelligent algorithms can also understand language in the form of texts or natural, spoken language better and better through deep learning methods. For example, it is extremely difficult to understand irony - not only for machines or programmes.

At first glance, sentences meant ironically are no different from sentences meant seriously. Sentences such as "You did a great job!" or "Today you looked especially smart." can be meant both seriously and ironically without the sentence itself changing. The context often provides the information that is crucial for interpretation.

For the service, for example, it is important to know whether an email from a customer is a normal enquiry or whether the customer is clearly angry. Here it can happen that hundreds of emails from customers arrive in one day. A filter that filters the Requests pre-sorted according to priority, helps to improve customer service enormously.

With suitable training data, algorithms can be trained in such a way that they can be used as a intelligent filter system can be used. From a mass of thousands of messages, for example, they can identify those from angry customers. Sentiment analysis, which virtually measures the feelings of customers, is used for customer care and can minimise the risk of customers dropping out.

Example 3: Improving safety architectures

The more interconnected the world becomes, the more important the issue of cyber security and Data security. Deep Learning can contribute to this, Close security gaps in systems. The learning capability makes the method particularly suitable for distinguishing normal activities from attacks or other irregularities. This ability makes Deep Learning interesting, for example, in securing sensitive locations such as airports.

Real-time monitoring is all about monitoring the live video footage and Identify conspicuous eventsn. The longer a deep-learning algorithm watches normal airport activity, the better it learns to distinguish which behaviours are atypical or conspicuous.

Example 4: Minimising risk in financial transactions

Deep Learning sees things that humans cannot see. But deep learning algorithms also have advantages in other sensitive areas that can become the target of attacks. For example, in the monitoring of bank transactions and securities trading. With anonymised training data, algorithms can be trained to specifically recognise unusual activities that occur within a banking network. In this way, in addition to Credit card fraud attacks by malware and other malicious software can also be warded off.

This makes one of the greatest challenges in the field of cyber security manageable: the Detecting first-time attacks by previously unknown malware or individual attackers. Even spam filters can be trained in this way to identify emails with malicious attachments. Alone due to the amount of Datawhich need to be verified, humans cannot keep up with systems based on Deep Learning in this task.

Example 5: Industry 4.0: Tool for mastering Big Data

Deep Learning is also used when it comes to analysing Big Data and the questions are very complex at the same time. Due to its strengths, the data science method is an important tool for mastering Big Data. For example, in the evaluation of sensor data, as in the case of Maintenance data that are generated in a wind farm. With this technology, measurements are taken at various points, sometimes every second, so that the amount of data quickly reaches the petabyte range.

In such complex industrial ecosystems, the algorithm facilitates the Unit forecastthat need to be maintained. You can find out more about this in our free Whitepaper for: Predictive Maintenance. Deep Learning also makes it possible to establish complex connections between industrial processes, data on interaction with customers and sales data.

The full effectiveness comes with time

In very general terms, Deep Learning is a data science method that is used in particular when large amounts of unstructured data are available in which certain patterns be recognised should. Especially large amounts of audio data, video data or image data have been the focus of interest in recent years.

The intelligent solutions in this field open up areas that were long reserved for humans. What distinguishes Deep Learning from most other methods and makes it so intelligent is the learning aspect.

For this is therefore not "Out-of-the-box solutionsthat are programmed once and are then immediately ready for use. Rather, the algorithms inevitably need a certain training phase during which they learn to fulfil their respective tasks. To do this, the algorithms make some assumptions and check these assumptions against the test data. In this way, they learn from their mistakes or successes and become better and better over time.

For this reason, for example, chatbots or digital assistantsbased on Deep Learning have become better and better the more often they have received feedback from users as to whether their answer was helpful. It is true that the capabilities of Siri, Cortana & Co are still relatively limited today.

However, if you look at the successes achieved by Deep Learning alone in recent years, it is only a Question of timeuntil they will talk to us naturally and maybe even have a sense of humour.

Author

Michaela Tiedemann

Michaela Tiedemann has been part of the Alexander Thamm GmbH team since the early start-up days. She has actively shaped the development from a fast-moving, spontaneous start-up to a successful company. With the founding of her own family, a whole new chapter began for Michaela Tiedemann at the same time. Hanging up her job, however, was out of the question for the new mother. Instead, she developed a strategy to reconcile her job as Chief Marketing Officer with her role as a mother.

0 Kommentare