AMD y su compromiso con la Inteligencia Artificial en Chile

AMD is participating together with Hewlett Packard Enterprise (HPE) in diverse activities to further introduce its technological solutions focused on the development of AI in our country. Recently both companies exhibited in the event “Intelligence Artificial en los servicios públicos: Hoy y en el futuro”, organized by the National Center for Artificial Intelligence (CENIA), where they delivered technological perspectives and advances in high performance computing and their application in solutions for citizenship.

Design and implement applications based on Deep Learning and Machine Learning mucha veces encuentra una piedra de tope al enfrentarse al large volumen de información que se debe procesar. Currently we live in a ‘date revolution’, where we must maintain an equilibrium that allows us to efficiently handle variables such as speed of processing, security and real utility of the material collected.

“We want to contribute from the computing power of our products to the development of Artificial Intelligence solutions that really are a contribution to society and help reduce the technological gap that exists to implement new applications for the public sector”, says Juan Moscoso, Regional Sales Leader Data Center de AMD para Sudamérica de Habla Hispana.

Machine Learning IA Intelligence Artificial

Advances in Artificial Intelligence

Los actuales sistemas de Intelligence Artificial de computadoras Machine Learning y Deep Learning de última generación pueden ajustar las operaciones después de una exposure continua a datos y otros tipos de información. Aunque están relacionados en cuanto a su naturaleza, existen diferencias sutiles que separan estos campos provenientes de las ciencias de la computación.

El aprendizaje automático se referé al system que puede aprender active por sí mismo, en lugar de solo recibir información de forma passive y procesarla. El sistema de computadoras se codifica para que responde a la información brindada como si fuera un humano mediate el uso de algoritos que analizan los datos en busca de patrones o estructuras. The ML algorithms are designed to improve performance with the passage of time, as they are exposed to more data.

When a human recognizes something, it is instantaneous. Para ayudar a imitar este proceso, los algorithmos del aprendizaje automático usicana redes neuronales. Al igual que el proceso de aprendizaje humano, el cómputo de las redes neuronales classifies los datos (por ejemplo, un conjunto enorme de fotografias) según los elements renconcios dentro de la imagen. The success rate of the correct classification can be improved with the passage of time through information provided by “experts”, which helps the system learn and discern the correct decisions from the incorrect ones, with a focus on the maximum efficiencia y en una mayor exactitud. The neural network algorithm modifies all future decisions according to the information received. This process imitates human recognition through network training to produce the desired result.

competitive products

Las aplicaciones inteligentes que responden con reflejos similares a los humanos requiem una enorme quantity de capacidad de procesosamento informatico. AMD’s principal contributions to systems Intelligence Artificial provienen de oferer cómputos de alto performance (both CPU and GPU) con un ecosistema abierto para el desarrollo de software. Big Data analysis applications depend on a technical base that can support the highest processing capacities (velocity, capacity and organization) to simultaneously administer complex sets of data from various data transmissions.

Por ejemplo, in the case of autonomous vehicles, it is possible that the Deep Learning algorithm must recognize the next change from green to yellow of the traffic light, the movement of a nearby pedestrian and water on the road as opposed to rain, between una infinidad de otras variables en tiempo real, al igual que las operaciones básicas de los vehicules. Es posible que un conductor humano expérience dé por sentado estas reacciones de coordinación. Sin embargo, para simular las capacitas del cerebro humano, el algorithm del manejo autónomo requiere un processing eficaz y accelerado para tomar complejas con la velocidade suficiente y una alta exactitud para guarante la seguridad de los pasajeros y otros a su alredero.

“There are infinite possibilities for Chile to become a strategic development pole Intelligence Artificial. Tenemos los profesionales, los researchers y la possibility de testear a escala solutions que puedan llegar a miles de usuarios. Our role in this virtuous circle of innovation is to provide high-performance computing technology that allows sustaining the large load of information that requires any application of AI to be as simple as this,” explains Juan Moscoso.

The performance of AMD hardware and associated software also offers great benefits to the process of development and testing of artificial intelligence systems. Currently, a computer platform based on the latest AMD technologies (CPU AMD EPYC and GPU Radeon Instinct) can develop and test a new intelligent application in days or weeks, a process that used to take years.

The company also works extensively with the open community of IA to promote and expand the capabilities and optimizations of automatic and deep learning. After acquiring Xilinx, AMD has boosted the Vitis AI tool by providing an integral AI inference development platform for Alveo data center accelerators and AMD adaptable SoCs. Vitis AI connects to common software development tools and uses a wide set of open source libraries optimized to empower software developers with the acceleration of automatic learning as part of their software code.

AMD is in a unique position with its wide portfolio of products and software to help clients and partners develop and implement applications with multiple forms of Intelligence Artificial desde la nube y las organizaciones, hasta el perímetro y los terminals. With a diverse set of hardware that includes AMD Instinct accelerators and adaptable CPUs, AMD can support a wide variety of generalized models Intelligence Artificial y Machine Learning, desde pequeños puntos de borde hasta cargas de trabajo de gran escala.

Leave a Comment

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