In 2011 IBM’s supercomputer, Watson, entered the limelight when it competed on Jeopardy! against two former winners. Watson received the first-place prize of $1 million. Thereon, Watson has emerged as the symbol of a new era in cognitive computing, with a radically new approach as compared to other programmable systems.

Dr Watson

As supercomputers can analyse vast stores of data and recognise patterns, they are a natural fit for medical application. IBM’s early collaborations produced impressive results and led a charge to revolutionise healthcare. Watson now addresses a variety of other medical areas, including personalised care, patient engagement, imaging review and drug discovery.

Consider a patient with a rare, undiagnosed disease. Now consider a single, secure database that could analyse that patient’s symptoms and check thousands of clinical studies, similar patient records and medical textbooks to present a risk-matched list of potential diseases. For example, if a patient has a rare, genetically linked form of lung cancer, a typical cancer doctor is unlikely to be up-to-date with all of the latest developments in lung cancer treatment (in the last year alone, the Food and Drug Administration has approved at least seven new lung cancer drugs). Doctors may not be aware of how best to use those drugs or even if they apply to specific patients.

Chef Watson 

Cognitive computing can create profoundly new kinds of value ? finding answers and insights locked away in volumes of data ? whether this relates to a doctor diagnosing a patient or even a chef designing a new recipe.

Cognitive computing applications such as IBM’s Chef Watson process information more like a human than a computer. These applications can analyse huge amounts of data and use natural language processing technology to decipher the meaning of the words and sentences contained within the data. These programs also excel at recognising patterns in large data sets and become ‘smarter’ over time, meaning they improve based on continuous feedback.

For example, Chef Watson can learn and design new recipes based on ingredient availability, personal preferences and dietary requirements. It is also familiar with the chemical composition of hundreds of ingredients, and has analysed over 10,000 recipes from different sources (eg, Bon Appétit). By combining that data and detecting certain patterns, Chef Watson has learned to suggest up to four different ingredients that blend together seamlessly. Cooking Italian? Chef Watson will run through the most typical ingredients used in that cuisine, in addition to complimentary ingredients, in order to suggest the tastiest option.

Chef Watson can even help to reduce food waste and assist in discovering new combinations from fridge leftovers. Users can input the available ingredients and it will create a novel and nutritious combination out of those ingredients (see Florian Pinel, head software engineer of Chef Watson, discuss cognitive cooking).

Following Chef Watson, there have been many additional areas where IBM has applied its artificial intelligence. Watson could revolutionise diverse areas such as the legal industry (Advocate Watson), cybersecurity, customer service (Watson Virtual Agent), travel advice and even tax return preparation.

Can Watson identify patent recipes?

White space identification is a key area where the capability of cognitive application programming interfaces (APIs) and programming can prove its worth. This process serves to enhance human expertise to look for a solution to a problem statement. Watson and its cognitive capabilities mirror some of the key cognitive elements of human expertise systems that apply reason to understand something or solve problems. Machine learning and natural language processing allow Watson to understand how food recipes work – likewise, it can be taught to identify a solution to a problem statement based on patent data.

The following is an analogy between the ‘thought process’ of Chef Watson’s API and various applications of Watson as a patent consultant:

  • Combination designer – Chef Watson can index food ingredients at a molecular level and categorise the data based on nutrition value. These are the building blocks for Chef Watson, based on which different combinations are derived. Similarly, Watson can parse and index patent data using natural language processing and context-understanding APIs to create a repository at a more granular level.
  • Cognitive assessor – Chef Watson can assess the viability of all combinations that are designed in the first stage. It categorises and ranks the recipes, some of which are already known. These are ranked based on personal preference and any dietary requirements. Based on this assessment, top-ranked recipes are presented to the user. Analogously, based on a problem statement, Watson can digest and identify the most relevant data set to derive a contextualised solution. It can also suggest a new approach to a problem based on its ability to contextualise the problem and the patent data.
  • Dynamic planner – Chef Watson can utilise this functional component to plan the dish in terms of cooking procedure and other useful tips. Watson can also use the dynamic planner to suggest different solutions to a problem – it is then the user’s task to identify whether the procedures suggested by Watson are viable.

While no remarkable achievements in the domain of patent searching and analytics have emerged, given the approach that Watson Chef has adopted, it is safe to assume that Watson will soon be able to consult with inventors or researchers to discover recipes that can be patented and protected. The process will be based on existing solutions in the patent literature and Watson’s capability to rework those solutions to solve a given problem.

Can Watson achieve this? It is a matter of necessity and time. The increase in the amount of patent literature demonstrates the growing need to efiiciently utilise such data.

This article first appeared in IAM. For further information please visit www.iam-media.com.