Continue reading or Download a free copy of the complete e-book
Artificial intelligence (AI) and machine learning (ML) are important emerging technologies that have the potential to transform organizations. Exponential increases in computational capacity, the emergence of cloud computing, and innovations in algorithms have resulted in tremendous advances in the application of AI. Leveraging AI and ML techniques, organizations can unlock tremendous value through improved customer service, streamlined operations, and the realization of new business models.
Despite tremendous advances in the field over the last decade, AI remains very much the domain of expert data scientists. AI technologies are a complex subject and are relatively new to the business world; few managers and enterprise technology professionals have an understanding of these techniques. Furthermore, implementing AI techniques within the enterprise requires different skill sets and approaches from traditional software products.
In order for organizations to truly unlock value from AI, its practitioners and their methods need to be fully integrated into the fabric of the enterprise. This integration requires that managers have a basic understanding of the science behind AI/ML, the techniques leveraged, and the metrics used to measure success.
Managers who are well-versed in the complexities of leading AI/ML efforts will be able to capture the most value from these new technologies. These managers will be able to select the best AI use cases, effectively collaborate and problem-solve with data scientists during prototyping phases, support the transition of algorithms into production use, and design the right business processes and change management activities to capture value for the organization. In order to achieve this, managers need a “field guide” to AI and ML techniques.
As we have designed, developed, and implemented AI techniques and AI-enabled enterprise applications at organizations across the world over the past decade, it has become clear to us that such a field guide does not exist. Most books and articles are either too technical, and focused on machine learning practitioners, or are too managerial without encapsulating sufficient mathematical and machine learning knowledge.
This online resource attempts to provide such a managerial field guide. It captures essential information about the practical application of enterprise AI/ML techniques, gathered from our extensive experience at C3 AI, across a wide range of industries and business problems. It also captures our experience with AI/ML teams – recruiting, organizing, and managing high-performance teams.
Logic-based algorithms represent the core of traditional programming. For decades, computer scientists were trained to think of algorithms as a logical series of steps or processes that can be translated into machine-understandable instructions and used to solve problems. Traditional algorithmic thinking is quite powerful and can be used to solve a range of computer science problems, including data management, networking, and search.
Traditional logic-based algorithms effectively handle many different problems and tasks. But they are often not effective at addressing tasks that are quite easy for humans to do. Consider a basic task such as identifying an image of a cat. Writing a traditional computer program to do this correctly would involve developing a methodology to encode and parameterize all variations of cats – different sizes, breeds, and colors as well as their orientation and location within the image field. While a program like this would be enormously complex, a two-year-old child can effortlessly recognize the image of a cat.
ML algorithms take a different approach from traditional logic-based approaches. ML algorithms are based on the idea that, rather than code a computer program to perform a task, it can instead be designed to learn directly from data. So instead of being written explicitly to identify pictures of cats, the computer program learns to identify cats using an ML algorithm that is derived by observing a large number of different cat images. In essence, the algorithm infers what an image of a cat is by analyzing many examples of such images, much as a human learns.
AI algorithms enable new classes of problems to be solved by computational approaches faster, with less code, and more effectively than traditional programming approaches. Image classification tasks, for example, can be completed with over 98% less code when developed using machine learning versus traditional programming.