Artificial cognitive systems: infinite brain evolution

Artificial intelligence in cognitive systems

Artificial cognitive systems are those mechanisms based on computers, interactive and with a high degree of reliability whose task is to simulate human knowledge processes, make decisions and solve complex problems. In this context, Stuart Russell and Peter Norving (2004) in his book Artificial Intelligence: A modern approach It is noted that artificial intelligence simulates these processes and claims that decision making is considered the highest level of intelligence and human experience. In addition, they list and characterize the components of these systems:

• Knowledge base: In this component the facts and rules are represented. Here knowledge is stored in a particular domain (or field) (the cognitive in this case).
• Inference motor: It is the brain of the specialist system and its function is to obtain the relevant knowledge of the database, interpret it and find a pertinent solution to the problem to solve. It contains the rules of its knowledge base and applies them to facts known to infer other new ones. Thus, it provides reasoning about information in the knowledge base.
• Knowledge and learning acquisition module: This component allows the system to acquire more and more knowledge from different sources and store it in the knowledge base.
• User Interface: It is the crucial element of the cognitive system as it admits that an unexpert user interacts with the system and finds a solution to a problem.
• Explanation module: It allows the user to obtain a response related to how the artificial cognitive system came to a particular conclusion. Thus, it argues the results with the highest accuracy.

Artificial cognitive systems consist of artificial neuronal networks (linked nodes that transmit information to each other) and are based on Conexion (a multidisciplinary approach that includes cognitive psychology, neurosciences and artificial intelligence), which states that in the human brain the information is processed through activation propagation patterns that allow networks to be formed between neurons that will process this information received quickly and without the need for pre-programmed algorithms.

In order for these systems to function as the human brain, they must comply with a number of conditions that are enunciated by Rudolph Russell (2018) in their book Neuronal Networks: Simple Guide to Artificial Neuronal Networks:

• Propagation of activation: Neurons, when activated, influence those with which they are connected. This can occur by facilitating your activation or by inhibiting it.

• Neuronal learning: Learning and experience affect connections between neurons.
• Parallel processing: In this process, activation propagates in parallel between all neurons. Thanks to this, we are able to interpret a lot of data at the same time, although there is a limit in our capacity.
• Neural networks: The system is formed by a large network of neurons grouped together through inhibition and activation mechanisms. In these networks, the inputs information and information outputs behavioral. These clusters represent the structured information that possesses the brain and activation patterns develop the way the processing of it occurs.

Complementing the aforementioned so far, the authors Ramón García Martínez and Paola Britos (2004) in his book Systems Engineering ExpertsThey expose that the qualities that must gather an artificial cognitive system are as follows:

• Interactivity: It is given among all system components (the machine. users, applications, devices and services) not to interrupt the flow of information and knowledge transfer.
• Adaptation: Simulating the human brain, cognitive learning is presented as an improved and molded version, prepared to adjust the needs of each environment. To achieve this, you must ensure your agility in understanding requirements and objects, as well as your dynamism regarding data collection.
• Contextualization: Understanding, identifying and extracting contextual elements are key points in the cognitive learning process of machines. Time, task, location, user profile and goals are different facets of the same process that, using multiple sources of information (structured and unstructured), will give context to the data.
• Iteration: It is the most recommended approach to data quality and, for this reason, cannot fail in any cognitive computing system if we want to ensure that this information will be able to provide sufficient information in the necessary upgrade, accuracy and reliability conditions.

Scientific and technological advances

The advances made by cognitive learning systems are present in different areas of everyday life, such as facial and/or voice recognition, health, sales, marketing, among others. Many companies have started to integrate this technology into their routine business issues efficiently and some successful cases are as follows:

• Cora - Intelligent Agent: With the support of IBM Watson, the Royal Bank of Scotland has developed a smart assistant who is able to handle 5,000 queries a day. Using learning capabilities, the assistant gave the bank the ability to analyze customer complaints data and create a FAQ repository. The system not only analyzed the queries, but was also able to provide a thousand different responses and understand two hundred customer attempts. Learned how users do general issues, how to handle the query and transfer it to a human agent if it is too complicated.
• Health Care Assistant: The medical and health company Welltok created an efficient medical care assistant, CaféWell, which updates the relevant health information of customers when processing a large amount of medical data. CaféWell is a holistic health tool that collects data from various sources and instant question processing by end-users, offering intelligent and personalized health recommendations.
• Personal travel planner: The Wizard WayBlazer allows people to plan travel by asking questions in natural language and providing results when collecting and processing travel data, as well as information about traveler preferences. This kind of cognitive tool helps save time on flight search, hotel bookings, and plan activities without searching on multiple sites before you complete the trip.

The examples mentioned above are some of the many advances that daily revolutionize the economic and social world through digital transformation, requiring major investments worldwide.

Advantages and disadvantages of the subject

The main advantages of implementing artificial cognitive systems are as follows:

• Precise data analysis
• More efficient business processes
• Improved interaction with the user
• Time reduction in data analysis
• Human error reduction
• Improved decision making at production and business level
• Control and optimization of productive processes and production lines

The main disadvantages of using these systems consist of:

• Data availability: Often, the data presented in the companies are inconsistent and low-quality, which presents an important challenge for organizations that intend to create value from AI on scale. To overcome this barrier, it will be vitally important to draw a clear strategy from the principle that allows extract the AI data in an organized and consistent way.
• Lack of qualified professionals: An enterprise-level obstacle to AI adoption is the shortage of profiles with skills and experience in this type of implementations. It is crucial, in these cases, to rely on professionals who have already worked on projects of the same size.
• Creativity: Humans can creatively respond to unusual situations, while specialist systems do not have this ability.
• Experience Sensory: Humans have a wide range of sensory experience availability, while specialist systems currently rely on a symbolic input.
• Degradation: Expert systems often experience difficulties to recognize when there are no responses or when problems are outside your area.

Artificial cognitive systems design an encouraging future for institutions and companies that use them in their administrative and commercial processes, because through logical data analysis, they promote operational efficiency and interaction with the increasingly optimized customer/user.

This artificial cognitive process, similar to that developed in the human brain, glimpses wide opportunities in several countries and continents decided to invest in Artificial Intelligence. Like all technology, although these systems present certain disadvantages of use, through their elements and characteristics demonstrate that they can be shaped and more and more precise, thanks to their adaptability, interaction, iterivity and contextualization with the world and the subjects surrounding them.

It is time to become aware of the imperative requirement to train students and professionals in these emerging areas that are updated daily, transforming the academic and labor profiles of the 21st century, as the American futuristic Alvin Toffler said: “the great engine of change is technology”.