Cognitive Systems & Cognitive Computing: A Case Study of WEM
EXECUTIVE SUMMARY
This report analyses the processes and techniques of cognitive systems and cognitive computing in our contemporary world of science, technology and innovation. In order to aid understanding, some of the basic terms were critically scrutinized and defined to ensure that the reader follow the analytical approach of this research. The Watson Engagement Manager, a current program of the IBM, intended to run for about ten to twenty years, is an example of such cognitive systems and was used as a case study for the purpose of this work.
INTRODUCTION
The growth of the initiative for cognitive systems grew with the increasing need to make the world a global village and to further enhance the ability to have an increased share in research agenda (Kellly, 2015) Brilliant scientists like Thomas Mallone that humans and computers can work at an optimized capacity, greater than that of any other machine, using the system of cognitive reasoning. The cognitive system therefore works at a higher and more brilliant pace than any existing computer there is (Knight, 2011). After a brief illustration of the historical background of cognitive systems and how they work in computing data cognitively, five techniques were juxtaposed with the processes or techniques carried out by the WEM, while four of these techniques seemed similar to what is obtainable in the processes followed by the WEM, one technique (the use of frames), tended to differ considerably.
HISTORY OF COGNITIVE SYSTEMS AND COGNITIVE COMPUTING: A CASE STUDY OF WEM.
The cognitive systems represent a method of evolution in the world of computers, whose impact is estimated to amaze the world over the next ten to twenty (10-20) years. The beginning of this growth can be traced to the year 1940, when mechanical devices were used for data organization and compilation. This system of data collection, proved important both in the national angle and in particular industries, as they were used for national programmes like the organisation of national census’ and industry and company reports, involving quantity and prices.
Programmable computing equally became a prominent phenomenon in the 1940s, when the first traditional computers where formed based on the mathematical thinking of John von Neumann, which were basically created to calculate rapidly. The cognitive system of our contemporary time is however dissimilar from what existed in the 20th century, as it tries to relate with connections between humans and data or available information and continues making changes or reprograming itself to optimize productivity (CSEED, 2011). Research has shown that a gigantic flow of information is passed around the world through technological means and the traffic on these means continue to grow daily. As a result of this, optimization has become a great problem/need for the people.
ANALYZING EXISITNG TECHNIQUES IN JUXTAPOSITION TO WATSON ENGAGEMENT MANAGER (WEM)
CLASSIFICATION
The WEM makes use of the classification technique. This fact can be backed by so many sources of available data. For instance, citing the document provided for carrying out this assignment, one finds that the first three points strongly support this claim.
The answers are keyed into the cognitive system and are classified, based probably on subjects or field of studies and are arranged in a taxonomy. Summarily, an engineer feeds questions into the system, arranges them in a certain way and allows the computer identify where these questions are located in the system, based on the taxonomy or the location of the said data that the user wishes to retrieve.
This idea of classification can be noted when one engages in online shopping in one of these online retail shops. Goods provided by the store are divided under various topics or icons like ‘dresses’, ‘watches’, ‘shoes’ etc. the customer then clicks on one of these icons (narrowing down the search) until he or she finds a particular item he desires.
K NEAREST NEIGHBOUR:
This is a technique, based on some unique method of algorithm that allows an option in a computer to be chosen, by the virtue of being ‘selected’ by other options in that computer program (Jasper, 2011). This can be said to be in the case with the WEM, as encapsulated in the points started in the provided documents, from number four to nine, using a lexical method. One notices that keywords in the data provided to the system are noted by the system. As questions with these keywords are keyed into the system, the system searches through all the likely answers that contain these keywords. Finally, the answers with the highest occurrence or repetition of these keywords are passed to the computer user. An example of such a system can be observed in our daily use of the Google search engine (and other search engines). The articles that come up after we type in our search requests are the articles that contain a higher repetition of the keywords. To understand this technique more, the image below can be quite helpful.
Let’s take an example. A case of a consumer who wishes to get a vintage watch from a store that deals in vintage items, keying in the question ‘where can I get a vintage watch in Brooklyn’, limits the search engine to providing addresses with best matches for vintage watches in Brooklyn.
FRAMES
When the issue of frames is being discussed, it means the certain division or structural representation of semantic nets that serve to represent a given set of data, information or entity as a set of slots and associated value. When several frames exist in a computer system, they are dependent on one another. Frames have been used to work on social networking cases and to help with calendar updates. An example of a primary frame, for instance could be an organization. Slots on the other hand are information such as address, company e-mail and phone number among others. Organizations that perform similar tasks can be linked together based on the availability of this information. However, does the WEM adopt this technique? The answer is ‘yes’ (at least to an extent). The WEM has answers stored in it, based on specific topics (which can be regarded as primary frames), these topics have sub-topics where particular answers can be gotten (which can be seen as slots). From this illustration, it is glaring that the WEM makes use of the frame technique to some extent.
A frame in one of its simplest forms is depicted by the image below.
LEARNING BY RECORDING CASES
As has already been stated in the introductory part of this report, cognitive systems all over the world are designed with the intention of such systems being able to reprogram themselves in order to ensure greater optimization as time goes on. From the analysis of the steps followed through by utilizing the WEM procedure (that is the document provided for the assignment), one finds that the technique of learning by recording can be said to be sued by the WEM. This because, data to be supplied as answers are not suddenly or miraculously created during the course of the process. They include the data that have been keyed in by certain engineers. Thus, information van only be retrieved through recorder cases (which could include files in various formats).
CASE-BASED REASONING
When one talks about case based reasoning, it means the justification of the reasoning ability of a system, based on the available cases that are open to such a system. The reasoning on the WEM can be seen as following this technique. The reason is simple, as the presence of a case-based reasoning technique is portrayed from even the preliminary stages of the working of the system. For instance, when the engineer keys in those initial key words into the cognitive system, he does it based on the need to optimise the availability of data or to create absolutely new means of retrieving and collating data. These decisions arise from cases of inadequacy in the past.
In order to better the concept of case-based reasoning, we can illustrate the scenario of someone who repairs air conditioners. When such an electrician hears similar sounds, while repairing an AC like the ones he has previously heard in the past, during the course of his career, proffering a solution would not be so much of a big deal. The same applies for the WEM, since the answers provided for questions are based on past problems, whose techniques can used in solving or answering issues rose at the time of need.
The picture below further explains the technique of case-based learning.
From the illustration above, one notices that when a new case (or question) is keyed in, searches for solutions are sort by going through similar cases, a solution is then proposed to the user, who revises and confirms it. Where such a solution is confirmed, it is retained as a case base for further future analysis where it can be retrieved for new problems or cases.
CONCLUSION
This report work looked at the concept of cognitive systems and cognitive computing, using the particular analysis of a recent invention by the IBM, known as the Watson Engagement Manager (WEM) (IBM news, 2014). To help the reader have a clearer view of what cognitive systems are and how they are computed, the work looks at how there was an evolution in the makes and functioning of computer systems from 19th and 20th century traditional computers, to IBM computers with the WEM characteristic. Also, it discusses the various techniques that can possibly exist in the working of a computer system and as measures to what extent and in which possible condition the WEM adheres to these techniques or follow the working processes as it tries to bring the best result to computer users.
The image below shows that with each passing day, the developers of WEM are trying to optimise its use. We can notice that the first WEM was only capable of memorizing, learning and make judgements.. However, the introduction of the WEM 3.0 has made optimization of working has been made possible because in addition to existing features, it is also possesses the capabilities of perception and reasoning.
REFERENCES
- Knight, Shawn (2011). IBM unveils cognitive computing chips that mimic human brain TechSpot:
- Hamill, Jasper (2013). Cognitive computing: IBM unveils software for its brain-like Synapse chips The Register.
- Computer Science and Electrical Engineering Department, Is Watson the smartest machine on earth?” (2011). University of Maryland Baltimore County.
- Ferrucci, D.; et al. (2010). “Building Watson: An Overview of the DeepQA Project”. AI Magazine 31 (3).
- BM Watson Group Unveils Cloud-Delivered Watson Services to Transform Industrial R&D, Visualize Big Data Insights and Fuel Analytics Exploration”. IBM News. January 9, 2014.
- Kelly III, Dr. John (2015). “Computing, cognition and the future of knowing” (PDF). IBM Research: Cognitive Computing. IBM Corporation.
