JUNE 2017CIOAPPLICATIONS.COM9of human-level artificial intelligence. He reviewed the timing of the most publicized AI advances over the past 30 years. His evidence suggests that even with the algorithms available, it wasn't until high-quality training datasets became available that the major AI breakthroughs were able to come to fruition.The AI Equation for the EnterpriseWe believe there is an essential equation that the CIO needs to understand if AI is to be a commercial success inside an enterprise. This equation reflects that there are not one but three necessary components to making AI working in the enterprise.AI = TD + ML + HITLSo let's break it down and imagine a company is trying to create an AI solution that can categorize customer support tickets by severity level. The categorization is based on the unstructured text showing an exchange between a customer and a customer support rep discussing a particular topic or problem within the support ticket.TD is Training Data: Training Data is a set of inputs with the correct outputs or examples with the correct labels that can be used as an example to train the Machine. In this example, the input is the unstructured text inside a support ticket. The outputs are the labels "topic" and "level of importance" which has been applied by Humans in accordance with definitions from the specific company in question. An automotive manufacturer will want to define these topics and levels of importance differently from a retail banker or a wearable technology company.ML is Machine Learning: The Machine Learning capability is the ability to convert Training Data into a predictive model that can be applied to new inputs ­ in this case, new support tickets with unstructured text. You want the Machine Learning model to apply its predictive power to create new outputs ­ in this case, the labels "topic" and "level of importance". One of the advantages of Machines compared to Humans is their ability to understand their own confidence level. Humans are notoriously overconfident at evaluating their own judgments. So you can accept or reject the prediction based on the Machine's own assessment of its confidence level. For example, if a support ticket has words and phrases which haven't been seen in the Training Data, or seen very infrequently, then the Machine will objectively assess its own confidence level as being low for that particular prediction.HITL is Human-In-The-Loop: This is the critical third component of commercially viable AI. If the Machine Learning model is not confident in its prediction, it can route it to humans to review and answer. In this blended model, you take advantage of the speed and scale of Machine Learning to address the less difficult tasks, while the humans handle the harder tasks. This allows imperfect algorithms to fail safely, and for the business to generate business value even while the algorithm is imperfect.So if you want to be ready to answer the question "What's our AI strategy?" from your CEO, pick your first business process that requires human judgment today, and start thinking about the three components of this framework. Breakthroughs in AIDatasets (First Available)Algorithms (First Proposed)Year1994Human-level spontaneous speechrecognitionSpoken Wall Street Journal articles andother texts (1991)19972005201120142015IBM Deep Blue defeated Garry KasparovGoogle's Arabic and Chinese-to-EnglishtranslationIBM Watson became the world Jeopardy!championGoogle's GoogLeNet object classificationat near-human performanceGoogle's Deepmind achieved humanparity in playing 29 Atari games bylearning general control from videoAverage No. of Years to Breakthrough:3 Years18 Years8.6 million documents from Wikipedia,Wiktionary, Wikiquote, and ProjectGutenberg (updated in 2010)1.8 trillion tokens from Google Web andNews pages (collected in 2005)700,000 Grandmaster chess games, aka"The Extended Book" (1991)ImageNet corpus of 1.5 million labeledimages and 1,000 object categories (2010)Arcade Learning Environment dataset ofover 50 Atari games (2013)Hidden Markov Model (1984)Negascout Planning algorithm(1983)Statistical machine translationalgorithm (1988)Mixture-of-Experts algorithm(1991)Convolution neural networkalgorithm (1989)Q-learning algorithm (1992)The challenge is to find the right way to blend Humans and Machines, not replace Humans with Machines
< Page 8 | Page 10 >