Artificial intelligence and machine learning techniques no longer represent a fertile material for science fiction but have gone beyond them to become a real and growing presence in all sectors. Think of all the things that didn’t exist a few years ago, such as the ease of big data processing, instant machine translation, chatbots that can automate personal conversations with large-scale customers, and even the famous DeepFake counterfeiting technique used to juggle clips Video difficult to detect.
As well as deep learning and natural language processing (NLP) systems and their applications, which are considered to be one of the most important trends in the technical field during the recent period, where large technology companies rely on them to develop their services, for example, Google relied on them to develop many services from the search engine via the YouTube platform. , Gmail, Maps and even an audio assistant that recognizes the language and provides interpretation. Facebook has also relied on him to provide recommendations that suit you in the news feed, and Amazon has also relied on him to develop its voice assistant Alexa.
The year 2019 saw a growing dependence on artificial intelligence and machine learning and their leading position in many areas, as companies increasingly seek to take advantage of these technologies in multiple ways. Many companies have tried to improve their user experience by integrating artificial intelligence into almost all of their solutions. Scientists teach machines to think and make decisions the way humans do. There is no doubt that the development of artificial intelligence and its impact on humans will further dominate the technology sector in the coming years.
The most significant developments in artificial intelligence in 2019:
1- Automatic learning and deep learning:
Deep learning is a type of Machine Learning, two of the most important applications of artificial intelligence. Deep learning is defined as an artificial intelligence application that allows systems to automatically improve their functions by acquiring knowledge through experience, and then using the same thing in data processing and complex calculations. As a result, machines will not need to be programmed separately for each job, as deep learning becomes possible using access to the data collected by the devices, thereby improving their learning ability. Companies choose deep learning of their systems to improve their performance and obtain precise results, to identify the risks to which they may be exposed and to work to avoid them.
Because the AI approach allows machines to make quick decisions, the most important areas in which companies use machine learning include prediction and classification systems, speech recognition, computer vision, and autonomous cars.
One of the most important examples of the development of machine learning and deep learning in 2019 is the success of OpenAI in training a robotic system called Dactyl to solve the Rubik’s cube puzzle of a hand based on the acquisition of knowledge from the real world, where the robot was fully trained by simulation and was able to transfer knowledge to New successful mode.
The biggest challenge faced by OpenAI researchers was to create simulation environments diverse enough to capture all movement in the real world, so they used a technique called Automatic Domain Randomization to improve hand skills to solve the Rubik’s Cube puzzle because increasingly difficult environments have been created endlessly in the simulation forming neural networks.
2- Deepfake technology converts a still image into the video:
In May, Samsung developed a system that can create fake videos for someone using a still image. They used machine learning technology known as the Generative adversarial network (GAN) to create fake videos by taking only one frame as input. Samsung researchers used high-resolution natural image synthesis to allow machine learning models to recognize the basic geometry of a person’s face so that they can be added as a mask on the face of a person. other person speaking in a video clip, and they were able to produce a video from the Mona Lisa board itself.
3- Content writing:
On February 14, OpenAI released a language model called (GPT) that can create coherent paragraphs of text, achieve advanced performance in many language model standards, perform machine translation, and answer questions.
On November 5, the company released the full version of the model called (GPT-2), which was able to define the context and create strong text by itself by writing a few sentences. The model was trained on more than 8 million web pages, creating content that is difficult to determine whether a text was written by a human or artificial intelligence system.
If AI has been able to analyze huge amounts of data and imitate humans to make decisions based on this data, and even decisions that can be considered better, given that machines with high abilities can possess a greater amount of knowledge more than any person in reality, Can’t he also write content at the same level as a person writes or even better than he.
It happened in the world of journalism in 2015, when the Associated Press published an economic headline “Apple exceeding Wall Street expectations for the first quarter of the year” Once you’ve read this article, it looks like it is written by a real human being, but if you read to the end, you will find that it was created using automated analyzes in other words: this sentence was written by a so-called “reporter robot”.
4- Processors based on artificial intelligence:
A common trend in 2019 was the development of AI processors based on AI chips, allowing companies to integrate AI trends, such as facial and voice recognition and machine learning into their systems.
To make these processors available to consumers, large companies such as Intel, NVIDIA, Qualcomm, ARM, and AMD are working on the use of artificial intelligence to develop their processors, so that features such as voice and facial recognition are integrated into the devices. Among the most important industries that will depend heavily on these treatments are cars and healthcare, where it can save many lives.
5- Artificial intelligence techniques that adapt to human uses:
2019 was marked by strong competition between speaker devices, in particular between Google Home devices and Amazon Alexa. The competition was mainly for the advantages of artificial intelligence added in the two devices, for example, Google Assistant it executes 3 consecutive commands, simultaneous interpretation in different languages, instant conversation in up to 27 languages, while ‘Alexa supports many features that make it easier to order goods from Amazon and pay for it with voice commands.
Voice assistance has been the most exciting category this year as it grows very quickly until it almost reaches the level of previous fast-growing categories such as smartphones, tablets, computers, and smart TVs.
6- Improvement of facial recognition technology:
Facial recognition technology has evolved considerably in recent years, as it is now used everywhere in airports, train stations, shopping malls, financial services and even by law enforcement agencies. The market for this technology is, therefore, growing exponentially.
Facial recognition technology faces many problems, perhaps the most important of which are: inaccurate identification and concerns about ethnic bias, which researchers are seeking to overcome through artificial intelligence.
7- Cloud computing:
The field of cloud computing has seen significant development in recent years thanks to the integration with artificial intelligence technologies which have facilitated analyzes, forecasts and data extraction, which has reduced costs and increase deployment.
Currently, the list of major cloud companies includes Ali Baba, Google, Amazon, and Microsoft. Experts assume that these companies will play more influential roles in 2020 as they continue to grow globally.
8- Electronic security:
Reports indicate that the development and diffusion of cloud services will contribute to an increase in security breaches, which will endanger user data and force companies that rely on cloud computing to seek smarter ways to secure.
Cybersecurity companies have relied on the development of their software on artificial intelligence and machine learning technologies to become smarter. Recently, a cyber defense mechanism has emerged that focuses on providing timely responses to attacks or threats to the information infrastructure. Frequent neural networks capable of Sequencing inputs with machine learning techniques to create supervised learning techniques that can detect suspicious user activity and can detect up to 95% of all cyber attacks.
Also, artificial intelligence will be a key component in the provision and management of 5G wireless network services, and this will require more reliance on AI technologies for insurance. But malware is also expected to become more sophisticated in 2020 due to hackers’ dependence on artificial intelligence technologies also in their development, which will increase the number of threats that companies will face in the coming years.