Enigma Of Cognitive Computing And Artificial Intelligence
Cognitive Computing (CC) imbibes self-learning systems via using pattern recognition, data mining, and natural language processing to prototype the way the human brain works.
Objectively, cognitive computing creates automated IT systems which are capable of determining and resolving problems without assistance. Mostly used in Artificial Intelligence (AI) applications, it is a subset of Artificial Intelligence.
A CC system encompasses the following characteristics:
✓ Machine Learning
✓ Natural Language Processing
✓ Spatial and contextual awareness
✓ Semantic Understanding
✓ Sophisticated pattern recognition
✓ Common Sense
✓ Vision-based sensing and image recognition
✓ Emotional Intelligence
✓ Reasoning and decision automation
✓ Robotic Control
✓ Algorithms that learn and adapt
✓ Neural Networks
✓ Noise Filtering
So basically, you can say that a cognitive computing system might be trained via neural networks. It is basically all about making machine intelligent. Enforcing artificial intelligence into a machine, also known as “Machine Learning”.
Machine learning (ML) has various forms. One of the refine forms provides the analyst a set of data exploration tools & techniques. A several ML models offer robust solution algorithms, and an approach to use the solutions for predictions.
The Microsoft, Amazon, Google, IBM clouds, and Databricks all provide prediction APIs which give the analyst control of various propensities. Few of them render on demand offerings, a limited prediction API for binary classification problems.
Not necessarily every machine learning issue has to be resolved from scratch. But, few problems can be implemented and trained on sufficiently available samples to achieve appropriate estimations and which can be more widely applicable.
For instance, face recognition, text analytics, text-to-speech, and speech-to-text, are problems for which “canned” solutions work frequently.
Many of the machine learning cloud service providers give these capabilities via an API, permitting developers to include them in their respective applications.
There have been many marvelous accomplishments in the field of cognitive computing (Modeling of human thought processes in a form of computerized model) and AI. A real challenge is in translating these accomplishments into a real value proposition for businesses.
Google has designed and built few effective mobile applications that are capable enough to estimate and realize visual and audio recognition. Google has also declared an open source Natural Language Understanding (NLU) system known as SyntaxNet.
Such NLU system is developed and brought up on Google’s TensorFlow. It is an open source neural network framework. Google has achieved an overall 90 % accurate result with their system. In comparison to the scenario l0 yrs ago, where part of speech tagging was confined to only entity extraction (verbs, nouns, etc.), this is indeed a good achievement.
It is imperative to understand the user intent. Google is equipped with a system that understands and realizes natural language and all of the ambiguity and confusion that goes with it, particularly in English language. Therefore, achieving a 90 % accuracy rate is an impressive accomplishment.
But, this is not really cognitive computing or AI issue, in spite of what the experts write in the press. The capability to interpret and translate user intent is a different game which segregates the cognitive computing from the statistical reasoning solutions.
There have been few organizations which have delivered impressive results with understanding user intent.
Siri generators Dag Kittlaus and Adam Cheyer of Viv Labs, have recently explained an AI personal assistant called Viv.
The employees had ordered pizza via Viv personal assistant. Viv was intelligent enough to make out the employee’s location and guess their pizza preferences, also, placing an order to the nearest pizza store which could fill their order.
Whole exercise got executed without, actually placing a single phone call, doing a Google search, or any typing at all. Further, they did all without downloading an app. This is indeed a remarkable feat in itself.
For handling your personal affairs, personal assistants are great support and help. What about resolving complex/complicated enterprise problems, like e-Discovery?
Lymba Corporation has designed & developed the capability to interrogate a contract via use of natural language processing.
Lymba’s contract analytic tool, via natural language inquiry, permitting you to explore clauses is based upon fundamental concepts that are germane to contract administration. For instance (Refer Figure 1), “What are the clauses that have time constraints?” The system will display results from all concern paragraphs that are having statements regarding time constraints; actually, not by keywords, but via concepts. Presently this application is intelligently capable of handling 25 variant inquiries regarding contract concepts.
Lymba Corporation : Figure 1 displaying Lymba Contract Analytic Tool
Distinguishing and demarking the value proposition
As a decision authority, how can one distinguish the value proposition from all of those technology organizations? A main concern of the problem is with editors scripting misconstrued statements which convince you that there are more abilities than what can be realized in real.
For instance one editor scripted: “So don’t expect Watson to be the only thinking machine option moving forward.”
To infer & conclude that IBM Watson and other cognitive solutions depicted in the Deloitte list are thinking machines is simply erroneous. IBM Watson is only a statistical reasoning remedy which does not interpret intent.
Watson has problems in performing co-referencing resolution, clustering knowledge, and inferring new facts. These shortcomings got apparent while answering list type queries. For instance “List the present last three CEOs of JP Morgan Chase.” Watson may retrieve different paragraphs regarding the same CEO or CEOs may be outside the scope of the question.
One more observation, among one of IBM’s commercials, Watson was doing a conversation with Bob Dylan. It is really interesting that Watson finds out Dylan’s major themes of his songs as time passes and love fades. It is not astonishing that Watson wholly missed those Bob Dylan songs which also dealt with civil rights and anti-war agitation themes.
- Lacerte Tax Hosting2017.02.22Xero’s Exporting to Two Different Tax Software Programs
- Microsoft Sharepoint2017.02.20Why The Wise Businesses Rely On SharePoint
- ProSeries Tax Software Hosting2017.02.17ProSeries Tips and Tricks to Save Your Time
- Drake Tax Software Hosting2017.02.16Things You Should Know Before Tax Time