Capacidates intelectuales de la intelligentia artificial

Para entender las capacitas de la intelligenia artificial (IA) tenemos que reconocerar sus diferentes compones, explicaando terms como ‘neuronal networks’,Machine Learning‘ (ML, aprendizaje automático), y ‘ Deep Learning‘ (apprendizaje profundo).

In general, artificial intelligence (IA) refers to the simulation of human intelligence in machines that are programmed to think like humans, imitate the human form of actuarial behavior, and show features associated with a human mind, such as learning and resolution. de problems.

There are three types of artificial intelligence:

  • Inteligencia artificial estrecha o débil (ANI, por sus siglas en inglés)
  • General artificial intelligence (AGI)
  • Artificial superior intelligence (ASI)

El siguiente cuadro los explica

BBVA Openmind -Banafa-3 tipos de IARedes neuronales

In informatics, a neuronal network is a system of programs and data structures that approximate the functioning of the human brain. Suele implicar un gran número de procesadores que operan en paralelo, cada uno con su limitada esfera de conocimiento y acceso a los datos de su memoria local.

A neuronal network is “trained” and fed with large amounts of data and rules on relations between estos (for example, “El grande de una persona es mayor que el padre de esa persona”). A program can then indicate to the network how to behave in response to an external stimulus (for example, the entry of a computer user who is interacting with the network), and the network itself can initiate activity by itself (intro de los límites de su acceso al mundo externo).

BBVA-OpenMind-Capacidades intelectuales de la inteligencia artificialAprendizaje profundo frente a aprendizaje automático

Para entender qué es el aprendizaje profundo o ‘Deep Learning’, it is important to distinguish from other disciplines within the field of AI.

An important branch of AI is el aprendizaje automático or ‘Machine Learning’, in which the computer extracts knowledge through a supervised experience in which a human operator helps the machine learn by giving hundreds or miles of training examples and corrigiendo manually sus errores.

Although automatic learning has become a dominant element in the field of AI, it has its problems. Por un lado, requiere dedicar mucho tiempo y por otro, sigue sin ser una verdadera medida de la inteligenia de las machinas, ya que depende del ingenio humano para ideata las abstractions que permiten a un computerar learner.

Por el contrario, a diferencia del aprendizaje automático, el aprendizaje profundo o ‘Deep Learning’ no es supervisado. En su lugar, se trata de crear redes neuronales a gran escala que permitan al ordenador aprender y ‘pensar’ por sí mismo sin necessidad de intervención humana directa.

El aprendizaje profundo no se parece a un programa informatics. The normal computer code consists of very strict logical steps, but what you will see in the deep learning is something different. No hay instrucciones que digan: ‘Si X es cierto, haz Y’.

BBVA-OpenMind-Keith Darlingon-Human level Artificial General Intelligence AGI
Human-level intelligence has come to be known as IA fuerte or Artificial General Intelligence (AGI)

En lugar de la lógica lineal, el aprendizaje profundo se basa en las teorías sobre el funcción del cerebro humano, por lo que el programa, formeda por capas intrincadas de nodos interconectados, aprende reordenando las connexions entre los nodos tras cada nueva experiencia.

El aprendizaje profundo ha demonstrated with potential for a software that could detect the emotions and events described in a text including no explicit reference to them. You can also recognize objects in photos and make predictions about the future behavior of people. Algunos ejemplos de aprendizaje profundo en action son los assistants basadas en el recognition de voz, como Google Now y Siri de Apple.

El aprendizaje profundo es muy prometedor. The capacity to analyze sets of massive data and develop computer systems that can adapt to experience instead of relying on a human programmer will lead to great advances, such as self-driving cars, robotic mayors, discovery of drugs, creation of new materials and robots con una mayor conciencia del mundo que les rodea.

Aprendizaje profundo e informática affectiva

La informatics affectiva, que consiste en el estudio y desarrollo de sistemas y dispositivos que pueden reconar, interpretar, procesar y simular las emociones humanas, es un campo indisciplinario que abarca el ‘Deep Learning’, la psicología y la ciencia cognitiva. Although the origins of this field are remontan a las primas investigação philosóficas sobre la emoción, the most modern framework emerged from the article of Rosalind Picard de 1995. One of the times of the investigation is the capacity to simulate empathy, mediate la cual la machine interpretaría el estado emocional de los humanos y adaptaría su behaviora a ellos, dando una respuesta adequada a esas emociones.

Las tecnologias de informatica affectiva que utilizán el aprendizaje profundo detectan el estado emocional de un usuario a través de sensors, micrófono, cameras y/o logic de software, y responden realizing specific funciones y predefinidas del producto/servicio, como cambiar un questionnaire o recommendar vídeos que se ajusten a ese estado de ánimo.

Cuantos más ordenadores tengamos en nuestras vidas, más queremos que se comporte con educación, que sean socialmente inteligente y que no nos molesten con información inconscendente. But this type of common sense reasoning requires understanding the emotional state of the person.

BBVA-OpenMind-Banafa-Device feelings-elisa-ventur-a computer can observe innumerable variables that may indicate emotional reaction and variation
Un ordenador puede observar innumerables variables que pueden indicare reacciones y variaciones emocionales

One form of viewing affective computing is the person-computer interaction, in which a device has the capacity to detect and respond adequately to the user’s emotions and other stimuli through the compilation of clues about the user’s emotion from different sources. The facial expressions, the posture, the gestures, the speech, the force and rhythm of the pulsations on the keyboard and the temperature changes of the hand on the mouse can signal changes in the emotional state of the user, and all of them can be detected e interpreted by a computer. An integrated camera captures user images and algorithms are used to process data and obtain significant information, while speech recognition and gesture recognition are other technologies that are being explored for affective computing applications.

The recognition of emotional information requires the detection of significant expressions in the data collected through profound learning techniques that process different modalities, such as the recognition of speech, the processing of natural language and the detection of facial expressions.

La emotividad en las máquinas

An important area of ​​affective informatics is the design of devices that could exhibit innate emotional capacities or at least simulate emotions in a convincing way. A more practical approach based on actual technological capabilities is the simulation of emotions in ‘chat bots’ to enrich and facilitate the interactivity between the human being and the machine. Mientras que las emociones humanas usually associarse a subidas de hormonas y otros neuropeptidos, las emociones en las máquinas podría associarse a estados abstractos relacionados con el progresso, o la falta de él, en sus sistemas de aprendizaje autónomos. Desde este punto de vista, los estados emocionales affectivos se corresponden con derivatas temporales en la curva de aprendizaje de un sistema de aprendizaje arbitrario.

Two grand categories describe the emotions in the machines: the emotional speech and the detection of emotion in facial expressions.

The emotional discourse includes:

  • ‘Deep Learning’ (apprendizaje profundo)
  • Bases de datos
  • Descriptors del habla

The detection of emotions through facial expressions includes:

  • Gestos corporales
  • Seguimiento physiológico

The vision of the future

La affective computing que utiliza el aprendizaje profundo trata de abordar uno de los principales inconvenidades del aprendizaje ‘online’ frente al facelense: la capacidad del profesor para adaptar immediatente la situación pedagogica al estado emocional del alumno en el aula. En las aplicaciones de aprendizaje electrónico o ‘e-learning’, la informatica affectiva basada en el aprendização profundo puede utilizarse para ajustar el estilo de presentação de un tutor informatica cuando un alumno está aburrido, interestado, frustrado o satisfioso. The services of psychological health also benefit from the applications of affective informatics at the time of determining the emotional state of the client.

Affective computing is the study and development of systems and devices that can recognize, interpret, process, and simulate human affects
Affective informatics consists of the study and development of systems and devices that can recognize, interpret, process and simulate human emotions

Los sistemas roboticos capaces de procesar información affectiva show greater flexibility cuando se trabaja en entronmos inciertos o complecos. Los dispositivos de compañía, como las mascotas digitales, use the affective informatics with deep learning capabilities to improve realism and provide a greater degree of autonomy.

Other potential applications center on social monitoring: for example, a car can monitor the emotions of all occupants and adopt additional safety measures, such as alerting other vehicles if the driver is upset. Affective informatics based on deep learning has potential applications in person-computer interaction, such as affective mirrors that allow the user to see how they act, emotional monitoring agents that send a warning before sending an email cuando estás en un momento de ira, o incluso los reproducores de música, que seleccionan themes in función del estado ánimo. Las companies could use affective informatics to discover that their products will not be received by the intended market. El papel de la informatics affectiva basada en el aprendizaje profundo puede extenderse a todos los aspectos de la vida.

Ahmed Banafaautor de los libros:

Secure and Smart Internet of Things (IoT) Using Blockchain and AI

Blockchain Technology and Applications

Quantum Computing

References

What is Artificial Intelligence

What Is ‘Deep Learning’ And Why Should Businesses Care?

Why Google Is Investing In Deep Learning

What is a neural network? Explanation and examples

Affective computing

Affective Computing

Cognitive computing

Leave a Comment

Your email address will not be published. Required fields are marked *