El Rincon del Dr. KakaWate

If you have been following the news you likely have heard about how WATSON defeated the two Jeopardy Champions in 2011. More recently, you probably read headlines talking about sending WATSON to medical school and it graduating in a week.

More recent headlines have mentioned the concept of Dr. WATSON being in the house. You may be wondering how does beating to Jeopardy champions (crushed is a better term) make a case to use WATSON in healthcare. The key lies in the capacity of W making sense of unstructured data. For those that do not know what unstructured data is: “Structured data refers to information that either does not have a pre-defined data model and/or does not fit well into relational tables.

Unstructured information is typically text-heavy, but may contain data such as dates, numbers, and facts as well. This results in irregularities and ambiguities that make it difficult to understand using traditional computer programs as compared to data stored in fielded form in databases or annotated (semantically tagged) in documents”. In other words W read so much that it was able to interpret and understand the english language. Watson understands natural language, not only the meaning. It has the capacity to adapt and learn from particular situations.

With the knowledge it has read and digested W generates many hypothesis, not just a single answer and then uses algorithms and lists the possibilities. It could be hundreds it could be 10. Then, it assigns numeric values to the hypothesis and deciders on the best possible option. W usually ended up with top 3 answers but in Jeopardy W needed one answer. For this, IBM added and additional algorithm that helped W decide depending on the strategic position of the game how much he would risk.

What can WATSON do?

1. Devour Huge volumes of text like information 2. 90% of the data created in the last 2 years, 80% unstructured, WATSON knows it already 3. Most physicians can only read at the most 15 hours a month 2 to 3 journals. 4. WATSON can read 4,000 articles, get the most important concepts, and then present the most relevant.

In healthcare we are not expecting a single answer. There could be several answers. Multiple chronic diseases and contribute to each others outcome. WATSON WILL NOT BE A DECISION MAKER (at least not for now), it will understand the nature of your question and develop educated diagnosis. Then a doctor can decide which is the best possible answer based on evidence based information.

As mentioned, WATSON IS ADAPTIVE, the more it “plays” the more knowledgable he becomes. W teaches itself and keeps getting smarter. W can modify its own algorithms and get rid of those that got him into “trouble” or made a bad call and learn to use the right ones more often. W represents cognitive computing or in other words a thinking computing.

1. ASK – you can ask it a very detailed question based on medical history. W will decide what data to access and come back with relevant information

2. DISCOVER: W is interactive, it can tell you which are his suggestions and if deemed necessary, ask you to give him more info to clarify particular doubts (Adding a symptom, quality, intensity, duration) – WATSON will tell you what he used in order to get this information, highlight the information he used and share it with you

3. Decides which is the most relevant answer in the case

4. Doctor WATSON? Not anytime soon

5. The goal is NOT to replace physicians, but to help them achieve the best treatment Another intriguing characteristic that helps WATSON is that he is not human. Not being human removes from the equation self reinforcing perception bias that physicians encounter after facing a particular patient and solving it succesfully. It protects you and the patients of following a path that you believe is right.