Artificial Intelligence (AI) is the wave of the future and IBM’s Watson leads the way as the posterchild for AI. Not only does Watson place IBM ahead of the game, but they have registered the most AI patents, allowing it to corner the market. In a recent article published on Seeking Alpha, the analysts at I Know First broke down the purpose of Watson into layman’s terms as the following: “the Watson computer developed by IBM is designed to intake data in natural language, run huge scale big data analysis, and answer in natural language.”
IBM introduced Watson to the public in the form of an avatar on Jeopardy! in 2011. This colossal supercomputer boasts 90 servers, 15 terabytes of RAM and occupies as much space as 10 refrigerators. The size of the supercomputer allows it to hold massive amounts of data for quick retrieval, even when not connected to the internet. When connected to the internet, Watson has infinite amounts of data to sift through to answer questions.
David Ferrucci, an IBM researcher stated, “The goal is not to model the human brain. The goal is to build a computer that can be more effective in understanding and interacting in natural language, but not necessarily the same way humans do it.” Enter Watson. The architecture of Watson took developers into uncharted territory, providing them with challenge after challenge. The intense question answer program IBM developed for Watson, DeepQA, employs upwards of 100 overlapping techniques to answer a single question.
The DeepQA process begins when Watson receives a question in natural language. Watson then pulls key words from the question, searches its content (and the internet if available), then forms hypotheses based on its findings. It uses an intricate scoring system to support or refute the hypotheses until it narrows down the answer to the one with the highest score which indicates the most confidence. “This iterative process ensures that Watson is constantly learning. In 2008, this process took Watson 2 hours to answer a single question. Now, the supercomputer can answer a question in approximately 3 seconds.”
Watson has found a variety of uses from culinary to use by the USAA. However, IBM wants to focus Watson on the healthcare sector. With regards to healthcare and research in the field, I Know First reported that “Watson’s analytic and cognitive capabilities drive the creation of a variety of new healthcare apps aimed at providing personalized services.” More specifically, IBM wants to use Watson to help radiologists recognize abnormalities in X-Rays and other body scans. To this end, “IBM’s acquisition of Merge Healthcare…provides Watson with 30 billion anonymized medical images to store. As Watson analyzes the images, it “learns” to spot abnormalities.” In addition, “Baylor University’s College of Medicine uses Watson Technology to keep up to date with medical news and research. The typical researcher will read roughly 23 articles per month, just a fraction of those published. Watson’s technology makes it possible take in amounts articles that humans simply cannot.” Currently, Watson is learning Japanese, an extremely difficult language due to its characters and multitude of synonymous words.
I Know First researchers also found that Watson represents a huge long term investment for IBM“Chairman, President, and CEO Ginni Rometty is betting big on Watson. With over 1 billion in guaranteed funding going into the project over the next several years – with a large chunk being offset to the new free Watson Analytics platform – she is makes a real long term commitment to the technology.” As impressive a machine as Watson is, the technology is new and people do not necessarily recognize the applications for this technology quite yet. These huge investments in Watson that will not produce immediate profits coupled with IBM’s disposal of unprofitable businesses worry many investors despite the $600 million of revenue they posted in their second quarter. On top of that, they have growing long term debt, much of it used to repurchase stock; however, they have not missed a dividend payment in 100 years. IBM continues to search for new business ventures and partnerships to bring in revenue as most of their revenue currently comes from their IT service business.
I Know First has a bearish one year signal for IBM based on money flow analysis executed by their artificial neural network and genetic self-learning system indicating the stock is overvalued at the current level ($155). This is surprising considering the stock has been getting hammered for the last year, losing almost 25% of its value. Fundamentally, the business is overvalued because its revenue generation is simply below par, but this is due to their long term goals and investments in the future. To summarize, at the current level the stock is neither a buy, nor a short sell – value investors should avoid IBM in the near future, but not cut it out of their portfolio as IBM is the number one corporation to benefit from an impending AI boom. Technical, swing, and HFT traders should probably not touch this stock right now, as the current predictability measure for it identifies it as 100% unpredictable in the near term.