Documentos En Inteligencia Artificial

ARTIFICIAL INTELLIGENCE
ARTIFICIAL INTELLIGENCE
ARTIFICIAL INTELLIGENCE
La Inteligencia Artificial comenzó como el resultado de la investigación en psicología cognitiva y lógica matemática Se ha enfocado sobre la explicación del trabajo mental y construcción de algoritmos de solución a problemas de propósito basic.
The technological religion es una combinación de la ciencia del computador , physiology and philosophy, tan basic y amplio como eso, is that it brings together several fields ( robotics , expert systems , for example), todos los cuales tienen en común la creación de máquinas que pueden “to think”.
The idea of building a machine that can perform tasks perceived as requiring human intelligence is appealing. Tasks that have been studied from this point of view include games , language translation, language comprehension, fault diagnosis, robotics, provision of expert advice on various topics.
This is how increasingly sophisticated database management systems, the data structure and the development of insertion algorithms, deletion and data location , as well as the attempt to create machines capable of performing tasks that are considered typical of the field of human intelligence, coined the term Artificial Intelligence in 1956.
Fundamental theoretical work was the development of mathematical algorithms by Warren McCulloch and Walter Pitts, in 1943, necessary to enable classification work, funcionamiento en sentido normal, de una crimson neuronal. In 1949 Donald Hebb developed a learning algorithm for these neural networks, creating, together with the work of McCulloch and Pitts, la escuela creacionista. This school is considered today as the origin of Artificial Intelligence, however it was little addressed for many years, giving way to symbolic reasoning based on production rules , lo que se conoce como sistemas expertos.
Una caracter ística basic que distingue a los métodos de Inteligencia Artificial de los métodos numéricos es el uso de símbolos no matemáticos, although it is not enough to distinguish it completely. Other types of programs such as compilers and database systems , they also process symbols and are not considered to use Artificial Intelligence techniques..
The behavior of programs is not explicitly described by the algorithm. The sequence of steps followed by the program is influenced by the particular problem present. The program specifies how to find the sequence of steps needed to solve a given problem (Declarative Program). In contrast to non-AI programs, that follow a defined algorithm, that specifies, explicitly, How to find the output variables for any given input variable (Procedural Program ).
The conclusions of a declarative program are not fixed and are partially determined by the intermediate conclusions reached during the considerations of the specific problem.. Object-oriented languages share this property and have been characterized by their affinity with Artificial Intelligence..
Knowledge-based reasoning , it implies that these programs incorporate factors and relationships from the real world and the domain of knowledge in which they operate. Unlike special-purpose programs, such as accounting and scientific calculations; Artificial Intelligence programs can distinguish between inference motor reasoning program and knowledge base by giving you the ability to explain discrepancies between them..
Applicability to poorly structured data and problems, without Artificial Intelligence techniques programs can not work with this type of problems. An example is conflict resolution in goal-oriented tasks such as in planning , diagnosing tasks in a system in today's world: with little information , with a close and not necessarily exact solution.
Building replicas of the complex purple neural network of the human brain (backside-up).
Trying to mimic the behavior of the human brain with a computer (high-down).
III. DIFFERENT METHODOLOGIES:
Fuzzy logic: allows making decisions under conditions of uncertainty.
Neural Networks: This technology is powerful in certain tasks such as classification and pattern recognition. It is based on the concept of “learn” by aggregation of a large number of very simple elements.
This model considers that a neuron can be represented by a binary unit: At each moment its state can be active or inactive. The interaction between neurons takes place through synapses. Depending on the sign, the synapse is excitatory or inhibitory.
The functioning of a Crimson Neuronal network is governed by rules of activity propagation and state updating.
IV. RESEARCH OBJECTIVES IN ARTIFICIAL INTELLIGENCE.
Los investigadores en inteligencia artificial se concentran principalmente en los sistemas expertos, la resolución de problemas, automatic control, las bases de datos inteligentes y la ingeniería del software program (diseños de entornos de programación inteligente).
Otros investigadores están trabajando en el reto del reconocimiento de patrones donde se espera un rápido progreso en este campo que abarca la comprensión y la síntesis del habla, the image processing and computer vision.
At last, fundamental research on knowledge representation, cognitive conceptualization and language understanding.
One of the main objectives of researchers in artificial intelligence is the automatic reproduction of human reasoning.
The reasoning of a chess player is not always the same as that of an executive who wonders about the feasibility of manufacturing a new product. A child playing with wooden blocks on a table does not have thought about the complexity of the reasoning required to carry out the construction of a pyramid, and trying to get a robot to do the same as the child would require a long computer program.
V. ARTICLES ON ARTIFICIAL INTELLIGENCE
Below is a brief summary of the normal subject of three articles that talk about artificial intelligence and its different uses:
Artificial intelligence in medicine.
Current state and prospects
Normal Director. Applications of Artificial Intelligence, S.A..
EL campo de La inteligencia artificial debe gran parte de su precise desarrollo a los resultados obtenidos en el proceso de cierto tipo de problemas médicos: el diagnóstico y el tratamiento de enfermedades.
A mediados de los setenta, se obtuvo la evidencia de eficacia y acierto en este campo; a partir de estas fechas, las comunidades científicas de la medicina y de la ciencia de la computación vieron crecer su interés hacia este campo científico emergente. La importancia de la medicina en las aplicaciones de la inteligencia artificial ha sido realmente notable, hasta el extremo de que esas aplicaciones tienen nombre propio: PURPOSE, acránimo de Artificial Intelligence in Drugs que, desde hace 15 años, ha evolucionado como una activa y creciente disciplina.
A significant change in the role of computers in the field of medicine is expected. Computer models are already appearing on the market that will allow the incorporation of the described technologies conveniently; like this, for example, Palm-top pocket computers with magnetic pen input and linked handwriting recognition systems, allow real-time entry of information intended for the entity's medical record database, which in turn allow consultations (access to encyclopedias, pharmacopeias, advice) unthinkable at the moment. Simultaneously, the convergence of standardization work, Clinical communication and informatics will facilitate access to extensive sets of medical records that can be treated with knowledge extraction systems, from statistical methods to automatic classification and neural networks. The early detection of occasional contraindications in medications, the obtaining of optimal sequences of treatments and other complex problems will gain great benefits from the massive processing of medical records. No substitute role is foreseen in healthcare activities, although in some peripheral problems it is possible to use new technologies for diagnosis and treatment. For example, in aiding diagnosis in tropical diseases or other scarce specialties, consultas a un sistema inteligente sobre urgencias médicas, detección de contraindicaciones, entre otras Aplicaciones.
New Millennium AI and the Convergence of History
J¨urgen Schmidhuber
IDSIA, Galleria 2, 6928 Manno (Lugano), Switzerland
juergen@ – ∼juergen
La Inteligencia artificial (AI) se ha vuelto una ciencia formal en el nuevo milenio, durante todo del tiempo que ha transcurrido desde sus comienzos ha habido un progreso rápido en métodos prácticos por aprender los verdaderos programas del sucesión-proceso, los cuales son opuestos a métodos tradicionales imitados a la los modelos estacionarios. Cada descubrimiento en la informática tiende a venir aproximadamente dos veces mas rápido que el anterior, haciendo comparación con la evolución de la computación, The changes that are occurring in technology these days can be said to be exponential.
These days it is expected that the next radical change will manifest approximately a quarter of a century after the most recent one, I mean, in 2015, At this moment some computers will match intelligences in conditions to give great power to computing, and the coming revolutions will be faster; these are expected between 2030 and 2040, when individual machines will already approach the maximum computing power in combination with human intelligences.
Many of the readers present of this article should still be alive then. And the following question arises:
Will software programs and theoretical advances keep pace with hardware development??
That is what people are being persuaded of. In fact, The new millennium has already brought new fundamental visions regarding the problem of building optimal rational agents, theoretically universal Artificial Intelligences.
This article briefly reviews some of the new results, and speculates on future developments, while indicating time intervals between the most notable events over forty,000 years—29 lives of human history—these changes tend to be exponential.
Simplicial Households of Drawings
+1-608-263-2874
gleicher@
This article presents a method to help artists make a work of art more accessible to casual users. A focus is made on specific cases of drawings, Showing how a small number of these can be transformed into a more complex object transformed into an entire family of similar drawings. This theme is represented as an advanced simplicial that approximates a valid set of interpolations in the configuration space. The artist does not interact directly with the simplicial complex. On the other hand, The guidelines for its construction are made by answering specifically yes/no questions according to the chosen game. Combining the flexibility of an advanced simplicial with direct human guidance, they will be able to represent very general constraints on the number of members in a family.
The simplicial is built with a variety of algorithms useful to an end user, including random sampling, from the drawing space, shrunk interpolation between drawings, projection of another drawing into the family, and interactive exploration of the family.
VI. BOOKS CONSULTED ON ARTIFICIAL INTELLIGENCE
LOGIC AND INTELLIGENCE (ARTIFICIAL)
Luis Carlos Torres Soler
National University of Colombia
Faculty of Engineering
Scientists, in their field, require structured formulas and processes for problem solving, casi siempre complejos de ahí que se considera necesaria la estructuración del razonamiento lógico para mayor alcance de sus investigaciones. La capacidad para razonar depende de un conjunto de habilidades y capacidades lógicas, creativas y abstractas, que deben estructurarse y practicarse a fin de mejorar los procesos del pensamiento.
¿Es posible una formulación del pensamiento lógico? La pregunta proyecta razonar primero: ¿existe deficiencia en el razonamiento para solucionar problemas, no bien formulados, no bien estructurados con datos confusos? Y, second: ¿Qué provee cierto comportamiento de las personas en las empresas necesitadas de ideas para innovar?
Lógica e inteligencia artificial: El presente texto proviene de la convergencia de dos fenómenos, the need for logical reasoning and the projection towards intelligent computers. The book is the result of more than seven years of work in the field of artificial intelligence, creativity and mathematical logic.
The book is not divided into sections, it is a set of approaches where little by little the different ways of problem-solving and information representation are structured to generate an adequate process that can provide a solution, hopefully the most optimal one. The solutions found in the text have been thoroughly reviewed in various contexts, as indicated in them, it is not the only way, hence the reader is invited to find other representations and solutions, esto ayudará a la estructuración del pensamiento lógico.
TÉCNICAS DE INTELIGENCIA ARTIFICIAL EN JUEGOS
Nelson Becerra Correa
Universidad Distrital Francisco José de Caldas
Centro de Investigaciones y Desarrollo Científico
En el presente trabajo de investigación, de desarrollan métodos alternativos a los comúnmente utilizados para crear software program de juego, clasificados dentro de la teoría de juegos como bi personales de suma cero y de información completa.
El documento se divide en cuatro partes: en la primera, se trabaja sobre los métodos existentes en la actualidad para el tratamiento de juegos. Además se hace un análisis de estos métodos y se muestran métodos alternos no trabajados muy poco trabaja y novedosos.
En la segunda parte, A scheme of a goal-based algorithm is presented, The same as a conceptual analysis. Similarly, there is a treatment of games using neural networks. The basic concept consists of training a neural network on a game, so that it, after its training, can compete against an opponent with a high level of play and respond well.
In the third part, it is proposed to use genetic algorithms. This is proposed in two ways: first as a machine learning mechanism and second as a generator of static evaluation functions.
As a machine learning generator, the concept is to create programs capable of learning any type of problem by themselves, as a generator of evaluation functions, se darán los parámetros y la system normal para que el algoritmo la optimice.
En la cuarta parte se utiliza el aprendizaje por refuerzo, con el fin de enseñar a un agente un problema de ajedrez.
VII. CONCLUSIONES
Artificial Intelligence includes several fields of development such as: Robotics , Used mainly in the industrial field; Language comprehension and translation; Vision in shape-distinguishing machines used on assembly lines; Word recognition and machine learning; Expert computer systems.
Expert systems , that reproduce human behavior in a narrow field of knowledge, are programs as varied as those that diagnose blood infections and indicate treatment, those that interpret seismological data in geological exploration and those that configure complex high-tech equipment
Such tasks reduce costs , reduce risks in human handling in hazardous areas, improve the performance of inexperienced staff, and improve quality management, especially in the commercial field.
VIII. FUENTES
Documentos seminario Metodos de Investigación: INTELIGENCIA ARTIFICIAL EN MEDICINA