Machine Learning For Dummies. John Paul Mueller
Читать онлайн книгу.
and to some extent MATLAB, Java, Julia, and C++. (See the discussion at
https://www.quora.com/What-is-the-best-language-to-use-while-learning-machine-learning-for-the-first-time for details.)
Defining the Divide between Art and Engineering
The reason that AI and machine learning are both sciences and not engineering disciplines is that both require some level of art to achieve good results. The artistic element of machine learning takes many forms. For example, you must consider how the data is used. Some data acts as a baseline that trains an algorithm to achieve specific results. The remaining data provides the output used to understand the underlying patterns. No specific rules governing the balancing of data exist; the scientists working with the data must discover whether a specific balance produces optimal output.
You can also tune the algorithms in certain ways or refine how the algorithm works. Again, the idea is to create output that truly exposes the desired patterns so that you can make sense of the data. For example, when viewing a picture, a robot may have to determine which elements of the picture it can interact with and which elements it can’t. The answer to that question is important if the robot must avoid some elements to keep on track or to achieve specific goals.
When working in a machine learning environment, you also have the problem of input data to consider. For example, the microphone found in one smartphone won’t produce precisely the same input data that a microphone in another smartphone will. The characteristics of the microphones differ, yet the result of interpreting the vocal commands provided by the user must remain the same. Likewise, environmental noise changes the input quality of the vocal command, and the smartphone can experience certain forms of electromagnetic interference. Clearly, the variables that a designer faces when creating a machine learning environment are both large and complex.
The art behind the engineering is an essential part of machine learning. The experience that a scientist gains in working through data problems is essential because it provides the means for the scientist to add values that make the algorithm work better. A finely tuned algorithm can make the difference between a robot successfully threading a path through obstacles and hitting every one of them.
Predicting the Next AI Winter
Development of machine learning and AI is slow for a number of reasons, such as a lack of powerful hardware, lack of suitable data to feed algorithms, and people’s inability to understand their own thought processes. Businesses, however, are looking for ways to generate cash quickly based on new technologies. Obviously, slow development doesn’t work well with a quick return on investment (ROI). Developer-entrepreneurs exacerbate the problem by overselling technologies. They indicate that the state of the art is more advanced than it really is, often to enjoy windfall profits, gain power, and advance their careers. Because of the difference between timing and expectations, machine learning and AI have both experienced AI winters, a period of time when business shows little or no interest in the development of new processes, technologies, or strategies.
The first AI winter happened as a result of unfulfilled expectations resulting from the overselling of the technology and unanticipated difficulties. During the summer of 1956, various scientists attended a workshop held on the Dartmouth College campus to create artificially intelligent machines. They predicted that machines that could reason as effectively as humans would require, at most, a generation to come about. They were wrong. Only now have we realized machines that can perform mathematical and logical reasoning as effectively as a human. To achieve true human understanding, an AI would also need to demonstrate intelligence in the visual-spatial, bodily-kinesthetic, creative, interpersonal, intrapersonal, and linguistic realms. The stated problem with the Dartmouth College and other endeavors of the time relates to hardware — the processing capability to perform calculations quickly enough to create a simulation. However, that’s not really the whole problem. Yes, hardware does figure in to the picture, but you can’t simulate processes that you don’t understand, especially if you lack suitable data. Even so, the reason that AI is somewhat effective today is that the hardware has finally become powerful enough to support the required number of calculations.
Sites such as https://www.thinkautomation.com/bots-and-ai/the-ai-winter-is-coming/ see an AI winter in the near future, partly because the terms machine learning, deep learning, and AI have become overused and ill-defined. These same sites look at how business is actually using machine learning and AI today. In most cases, these sources say that the technologies are used for background processes, not front-line customer interactions. The thought is that automation used for front-end processes isn’t actually machine learning or AI, and that companies will eventually see automation as being separate from machine learning and AI. As a result, they will again stop investing in either technology. In many cases, proponents of an upcoming AI winter state that scientists should focus on the amazing array of tasks that machine learning and AI can perform today, rather than continue to hype some nebulous future tasks.
Before you get the idea that everyone is expecting another AI winter, you need to look at the other side of the argument. Sites such as https://towardsdatascience.com/there-wont-be-an-ai-winter-this-time-332a4b6d6f07 are saying that machine learning and AI are both so deeply embedded that an AI winter really isn’t possibly any longer. Typically, the articles you see are forthright in stating that machine learning and AI haven’t met certain goals, like creating autonomous vehicles. Even though these goals aren’t feasible today, the potential exists for achieving them in the future when scientists have completed more research. Moreover, because of the research conducted and the applications created, both machine learning and AI have become profitable, so business will continue to support them. The Towards Data Science article is good because it points out a wealth of vendors who are actually using machine learning in major line-of-business applications that generate huge profits.
In thinking about the future of machine learning and AI, considering a more moderate approach is likely best, such as the one found at https://mindmatters.ai/2020/01/so-is-an-ai-winter-really-coming-this-time/. At this point, data scientists and other researchers need to take a step back and consider the next level. The current technologies can only take us so far. They’re profitable, but they can’t produce a self-driving car and they certainly can’t produce a robot of the intelligence found in the film Ex Machina. So, if there is an AI winter, it’s likely to be a mild one because companies like Amazon.com and Google aren’t going to throw their technologies