What is deep learning?

Deep learning is a subset of machine learning, where computers process large datasets using neural networks modeled after the human brain.

In deep learning, the focus is primarily on the au­tonomous learning process of these neural networks. They consist of an input layer, one or more hidden layers, and an output layer. In­for­ma­tion enters the input layer as a vector, is weighted through ar­ti­fi­cial neurons in the hidden layers, and finally produces a specific pattern in the output layer. The more layers a neural network has, the more complex the tasks it can handle, enabling ar­ti­fi­cial in­tel­li­gence to tackle intricate problems.

How does deep learning work?

Sorting images according to whether dogs, cats or people can be seen in them is a chal­leng­ing task for a computer. Something that is im­me­di­ate­ly clear at a glance to humans requires a computer to analyze in­di­vid­ual image char­ac­ter­is­tics.

With deep learning the raw data input, in this case the image, is analyzed layer by layer. In the first layer of an ar­ti­fi­cial neural network, for example, the system examines the colors in the in­di­vid­ual image pixels. Each image pixel is processed with its own neuron. In the following layer, edges and shapes are iden­ti­fied and, in the layer after that, more complex char­ac­ter­is­tics are examined.

The in­for­ma­tion collected is displayed in a flexible algorithm. The results from one layer are carried onward into the following layer and change the algorithm. In this way, the computer is able to use a variety of op­er­a­tions to come to the con­clu­sion of whether an image can be cat­e­go­rized as a dog, cat or human.

At the start is the training period, errors in cat­e­go­riza­tion are corrected by humans, allowing the algorithm to adapt. After a short time, it can improve its image recog­ni­tion in­de­pen­dent­ly. As the in­ter­link­ing between the neurons in the network changes and the weighting of variables within the algorithm is adapted, certain input patterns (different kinds of cat pictures) lead more and more ac­cu­rate­ly to the same output patterns (the cat being rec­og­nized). The more image material is available to the system for it to learn from, the better.

With deep learning, it isn’t always possible for humans to un­der­stand which patterns the computer rec­og­nized in order to reach its con­clu­sions, par­tic­u­lar­ly since the system con­tin­u­ous­ly optimizes its own decision-making rules.

History of deep learning

Deep learning is actually quite a recent term – it was first used in 2000 – yet the method of using ar­ti­fi­cial neural networks to enable computers to make in­tel­li­gent decisions is several decades old.

Basic research in the field goes all the way back to the 1940s. Ar­ti­fi­cial neural networks were first developed in the 1980s. Back then, though, the quality of the decisions was dis­ap­point­ing, because machines’ in­de­pen­dent learning – deep learning – requires large quan­ti­ties of data, which at the time just weren’t available digitally. Only around the turn of the mil­len­ni­um did the age of big data begin, making deep learning in­ter­est­ing again for science and business.

Strengths and weak­ness­es

Compared with earlier AI tech­nolo­gies, deep learning is sig­nif­i­cant­ly more effective. Before the tech­nol­o­gy can reach its full potential, though, some weak­ness­es still have to be overcome.

Strengths of deep learning

One of the most important arguments is the quality of its results. In image recog­ni­tion and speech pro­cess­ing, in par­tic­u­lar, the tech­nol­o­gy is clearly superior to all others. Provided with high-quality training data, deep learning can carry out routine work much more ef­fi­cient­ly and much faster than any human – without any signs of fatigue either, and with no change in quality.

With other forms of machine learning, de­vel­op­ers analyze the raw data and pe­ri­od­i­cal­ly define ad­di­tion­al features that the algorithm is to take into account while learning in order to improve the AI’s fore­cast­ing power. With deep learning, the system itself rec­og­nizes useful variables and in­cor­po­rates these into its learning process. After the initial training period it can learn without any human guidance, saving both time and money since skilled employees aren’t necessary for future de­vel­op­ment.

Up to now, large quan­ti­ties of data had to be labeled manually in order to make machine learning possible. In image recog­ni­tion, for example, employees were required that would assign the label dog or cat to the images. With deep learning, the manual training period is sig­nif­i­cant­ly shorter. Above all this is relevant because, while general corporate practice certainly does involve col­lect­ing large quan­ti­ties of data, only in rare cases does it exist in the form of struc­tured data (telephone numbers, address, credit cards, etc.). In most cases it is stored as un­struc­tured data (images, documents, emails, etc.). Unlike al­ter­na­tive methods of machine learning, deep learning can evaluate different sources of un­struc­tured data while con­sid­er­ing the task at hand.

The argument that the tech­nol­o­gy is too costly in practice for it to be ap­plic­a­ble on a large scale is losing traction. Services like Google Vision or IBM Watson are in­creas­ing­ly emerging, allowing companies to build on existing neural networks instead of having to develop them from scratch. With this, in the future deep learning will be more and more capable of playing on its strengths in corporate practice.

An overview of the strengths

• Better results than with other methods of machine learning
• No feature de­vel­op­ment and no data labeling necessary
• Efficient execution of routine tasks without affecting quality
• Problem-free handling of un­struc­tured data
• More and more services to make it easier to use ar­ti­fi­cial neural networks

Weak­ness­es of deep learning

Deep learning requires an enormous amount of pro­cess­ing power. This largely depends on the com­plex­i­ty and dif­fi­cul­ty of the task to be ac­com­plished and the size of the data set used. Up to now, that made the tech­nol­o­gy expensive and only prac­ti­ca­ble for research and a handful of mega-cor­po­ra­tions.

There has indeed been ob­serv­able progress in this respect. What won’t change in the fore­see­able future, though, is the fact that decisions made by deep learning are no longer trans­par­ent to humans. The neural network is (so far) a black box. For some ap­pli­ca­tions where trans­paren­cy is decisive, this makes the tech­nol­o­gy ir­rel­e­vant.

For deep learning to work at all, large sets of training data are required. If these quan­ti­ties of data aren’t available, computers aren’t yet able to deliver reliable results with the help of deep learning. The first libraries of neural networks are indeed being published, making the ap­pli­ca­tion of deep learning easier for the general public. However, the services are not suitable for every ap­pli­ca­tion, meaning that the de­vel­op­ment of learning al­go­rithms for deep learning still demands a lot of time in­vest­ment, and po­ten­tial­ly takes more time than using al­ter­na­tive methods.

An overview of the weak­ness­es

• Requires high pro­cess­ing power
• De­vel­op­ing learning al­go­rithms is rel­a­tive­ly time-consuming
• A large data pool is necessary
• More training data needed than with other methods of machine learning
• Decisions difficult or im­pos­si­ble to un­der­stand (black box)

Ap­pli­ca­tion areas for deep learning

Deep learning is already being im­ple­ment­ed in various sectors, and in the future we will come across it in many more areas of our day-to-day lives.

• User ex­pe­ri­ence: Some chatbots are already optimized using deep learning and leverage Natural Language Pro­cess­ing to respond better to customer inquiries, easing the workload on human customer support teams.
• Voice as­sis­tants: As mentioned, deep learning is used in various voice as­sis­tants like Alexa, Google Assistant, or Siri through speech synthesis. These systems au­tonomous­ly expand their vo­cab­u­lary and improve their language com­pre­hen­sion.
• Trans­la­tions: Deep-learning-powered trans­la­tors, such as DeepL, produce high-quality trans­la­tions. Thanks to this tech­nol­o­gy, dialects and text from images can be au­to­mat­i­cal­ly trans­lat­ed into other languages.
• Content creation: LLMs like ChatGPT use deep learning to generate text that is not only gram­mat­i­cal­ly correct but can also mimic an author’s style—provided they have suf­fi­cient training material. Early ex­per­i­ments have seen AI systems create Wikipedia articles and re­mark­ably authentic Shake­speare­an texts using deep learning.
• Cy­ber­se­cu­ri­ty: Deep learning-powered AI systems are par­tic­u­lar­ly suited for detecting ir­reg­u­lar­i­ties in system activity, helping to identify potential hacker attacks.
• Finance: The ability to detect anomalies is es­pe­cial­ly useful in financial trans­ac­tions. Properly trained al­go­rithms can help prevent attacks on banking networks and credit card fraud more ef­fec­tive­ly than tra­di­tion­al methods.
• Marketing and sales: AI systems can use deep learning to perform sentiment analysis and au­tonomous­ly implement defined actions to restore customer sat­is­fac­tion.
• Au­tonomous driving: While fully au­tonomous vehicles remain a vision for the future, the tech­nol­o­gy already exists. It combines various deep learning al­go­rithms: one to recognize traffic signs, another to detect pedes­tri­ans, and so on.
• In­dus­tri­al robots: Robots equipped with deep learning AI could be deployed across numerous in­dus­tri­al sectors. By simply observing a human operator, these systems could learn how to operate machines and optimize their own per­for­mance.
• Main­te­nance: Deep Learning offers sig­nif­i­cant potential in in­dus­tri­al main­te­nance, where complex systems require con­tin­u­ous mon­i­tor­ing of numerous pa­ra­me­ters. Ad­di­tion­al­ly, it can predict which com­po­nents of a system are likely to require servicing soon (Pre­dic­tive Main­te­nance).
• Medicine: Deep learning AI systems can scan images for anomalies far more ac­cu­rate­ly than even a trained human eye. As a result, diseases can be detected earlier than ever on CT or X-ray images using these in­tel­li­gent systems.

Deep learning has great potential but isn’t a universal solution

In public discourse to some extent there is the im­pres­sion that deep learning is the only tech­nol­o­gy of the future for AI. It’s true that, in many ap­pli­ca­tion areas, deep learning makes much better results possible than previous pro­ce­dures did.

However, deep learning is not the best tech­no­log­i­cal solution for every problem. There are other strate­gies to make computers “in­tel­li­gent” – solutions that can also work with small datasets and where the decision-making is trans­par­ent for humans.

Some AI re­searchers view deep learning as a tran­si­tion­al phe­nom­e­non and believe that better ap­proach­es, not based on the human brain, will emerge. Google’s company strategy proves that these critical voices are not to be ignored: There, deep learning is just one part of the AI strategy. Alongside it are also further methods of machine learning, like the de­vel­op­ment of quantum computers.