The Differences Between Deep Learning vs Machine Learning

深度学习与机器学习的不同之处

2021-02-05 09:50 Lilt

本文共899个字,阅读需9分钟

阅读模式 切换至中文

Deep Learning vs Machine Learning: What is the Difference? As artificial intelligence (AI) continues to move into our everyday lives, understanding all of the different facets that create the technology can seem daunting. After enough digging, though, we can begin to uncover the main subsets that form AI. Then, we can take a closer look at AI vs machine learning vs deep learning.So, is deep learning machine learning?Today, we're going to answer what constitutes each subset and what is meant by deep learning. After learning the difference between artificial intelligence vs machine learning vs deep learning, you'll be able to understand which one is right for your business and what opportunities each has to offer. Deep Learning vs. AI First, we should start with what AI is and how artificial intelligence and deep learning relate to each other.AI is a broad science covering the concepts behind giving machines the ability to process information and formulate decisions from data like humans, if not in a superior way. This intelligence imbued into machines can be used to calculate predictions, automate processes, and streamline production.AI is frequently broken down into three main categories that include:Artificial Narrow Intelligence Artificial narrow intelligence is the lowest rung of AI computing, as it can only be applied to a predefined and specialized task. This category of AI has been seen for some time, as it's been used to solve chess problems, crawl web pages, or power chatbots. Artificial General Intelligence Artificial general intelligence is harder to define as it's still in its infancy, and the nascent technology hasn't been perfected. A general consensus believes that artificial general intelligence should at least match that of a human cognitive function and be able to: Use reason Generate common sense decisions Learn and plan Communicate and pass as human Utilize these abilities to reach a goal Artificial Super Intelligence Artificial super intelligence is the strongest form of AI and should demonstrate superior intelligence to humans in creating plans, deductive reasoning, and production. Artificial super intelligence is still in the realm of science fiction. However, the field of AI is growing exponentially every year. As it currently stands, machine learning and deep learning are still in the artificial narrow intelligence bracket but are starting to break into artificial general intelligence. What is Machine Learning? Now that we understand the basics of artificial intelligence, it's time to understand the differences between deep learning vs machine learning. At its core, machine learning is a subset of AI that allows a system to learn from data fed and improved on the output figure it produces. This process bypasses the need for a programmer to actively make adjustments to generate a correct result from the system, effectively learning independently.This learning process starts by being given data and specific instructions to determine if observable patterns can be detected. If patterns and sequences are found, then this information, in some cases, is retained and used to determine the relevance that new data should be affected by it. The system's goal is to learn with minimum, if any, human supervision. An example of machine translation would be how text is parsed into sequences of keywords or phrases. The system would then use semantic analysis to replicate a humanistic approach to deciphering a text block’s meaning. What is Deep Learning? Now that we understand one subset of AI, is machine learning and deep learning same?Comparing deep learning vs machine learning is similar to comparing machine learning to AI. Each is a subset of the other. Deep learning is notable because it relies on artificial neural networks to organize information and produce a usable result. An artificial neural network is a system composed of neurons (nodes) and connections (synapses) to replicate the structure of a human brain. This system takes in data, processes it through the neurons, and in some programs, will retrace the neurons to correct any errors. These neurons retain memetic data that is applied to a datum, giving it a weighted value. If the neuron adds enough weight to the information that it can surpass a predefined threshold, it is passed along to the next neuron, and the process begins again. Deep Learning vs. Machine Learning As stated above, the difference between machine learning and deep learning isn't so much what they are but how they're applied.The early machine learning models could adapt to new tasks given they are continually fed new data under a technician's supervision. For most cases, if a machine learning system encounters an error or produces an inaccurate result, it would have to be manually adjusted. Through deep learning's artificial neural network model, such as recurrent neural networks, it can determine if it produced an inaccurate result and make adjustments through its neurons. As reaching a global community becomes increasingly easier, the need for localization is becoming more urgent. While there are differences between deep learning vs machine learning, each bridges the remaining gap in connecting to an international community. To help, Lilt is seeking to change the approach and affordability of enterprise localization. We offer a platform powered by AI and neural networks so you can deliver the best user experience, regardless of where your users are in the world. Contact us today to learn more about our professional translation services.
深度学习与机器学习:区别在哪? 随着人工智能(AI)不断进入我们的日常生活,理解这项技术创造的不同方面似乎令人畏惧。不过,经过大量的研究,我们可以发现形成人工智能的主要子集。那么,深度学习就是机器学习吗?今天,我们将回答每个子集的组成以及深度学习的含义。在了解了人工智能,机器学习和深度学习之间的区别后,您将能够理解哪一种适合您的业务,以及每一种都有哪些用途。 深度学习与人工智能 首先,我们应该从什么是人工智能以及人工智能和深度学习之间的关系入手。人工智能是一门宽泛的科学,涵盖了让机器像人类一样处理信息并根据数据制定决策的能力,且目前为止还没有更好的方式。这种被灌输到机器中的智能可以用来计算、预测,使其过程自动化,并使生产合理化。人工智能通常被分为三大类,包括:狭义人工智能(artificial Narrow intelligence) 狭义人工智能是人工智能计算的最底层,因为它只能应用于预定义的专门任务。这类人工智能已经出现一段时间了,它被用来解决国际象棋问题,爬行网页,或者为聊天机器人提供电力。 广义人工智能 广义人工智能很难定义,因为它还处于初始阶段,而且新技术还没有完善。人们普遍认为,人工一般智能至少应该与人类的一种认知功能相匹配,并且能够做到以下几点: 使用符合逻辑的原因 产生符合常识的决定 能学习且有计划 能像人一样沟通和传达信息 利用以上能力达到目标 超级人工智能 超级人工智能是其最强大的形式,会在创建计划、演绎推理和生产方面表现出优于人类的智能。超级人工智能还停留在科幻小说领域。然而,在AI领域每年都在呈指数级增长。 就目前的情况来看,机器学习和深度学习仍处于狭义人工智能范畴,但正逐步进入广义人工智能范畴。 什么是机器学习? 现在我们了解了人工智能的基础知识,是时候了解深度学习与机器学习之间的区别了。从如果找到了模式和序列,那么这些信息在某些情况下会被保留下来,并用来确定新数据与其相关性。该系统的目标是在最少的人类监督下进行学习(如果提供人类监督的话)。拿机器翻译的一个例子,如何将文本解析成关键字或短语序列。然后,系统将使用语义分析来复制一种人性化的方法来破译文本块的含义。 什么是深度学习? 既然我们了解了人工智能的一个子集,那么机器学习和深度学习一样吗?比较深度学习与机器学习就类似于比较机器学习和人工智能。每一个都是另一个的子集。深度学习之所以引人注目,是因为它依靠人工神经网络来组织信息并产生有效的结果。人工神经网络是由神经元(节点)和连接点(突触)组成的复制人脑结构的系统。这个系统接收数据,然后通过神经元进行处理,在某些程序中,会重新追踪神经元以纠正所有错误。这些神经元保留应用于基准的模型数据,并赋予其加权值。如果神经元给信息添加足够的权重,使其超过预定义的阈值,它就会传递给下一个神经元,以上这个过程再次开始。 深度学习与机器学习 如上所述,机器学习和深度学习之间的区别不在于它们是什么,而在于它们是如何应用的。早期的机器学习模型能够适应各种新任务,前提是它们在技术人员的监督下不断获得新的数据。大多数情况下,如果机器学习系统产生错误或不准确的结果,技术人员就必须进行手动调整。但是通过深度学习的人工神经网络模型,比如递归神经网络,它可以确定自己是否产生了一个不准确的结果,并通过它的神经元进行调整。 随着进入全球化逐步深入,本地化的需要也变得更加迫切。虽然深度学习和机器学习之间存在差异,但它们都弥补了国际社会联系方面的差距。 为了提供帮助,利特(Lilt)正在想法设法改变企业本地化的方法和可负担性。我们提供了一个由人工智能和神经网络驱动的平台,因此无论您的用户在世界的哪个角落,您都可以为客户最好的用户体验。今天联系我们,了解更多信息。

以上中文文本为机器翻译,存在不同程度偏差和错误,请理解并参考英文原文阅读。

阅读原文