Sunday, May 26, 2024

The Critical Nature Of IBM’s NLP (Natural Language Processing) Effort

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.

There is a certain irony to calling technology like NLP (Natural Language Processing) by its acronym. This irony is because the technology is designed to be a human-machine language bridge that shifts the responsibility for communicating with computers from humans to those increasingly capable machines. But acronyms are favored by computers and not humans – who generally hate the things – so using an acronym to name the effort is ironically counterintuitive. 

The concept is critical to the evolution of computers within IBM’s NLP vision to turn them into better human partners and reduce the emphasis on efforts to make them human replacements. 

Let’s chat about that this week.

Evolving Natural Language Processing

Our history with tools is that humans may develop them, but then they have to learn how to use them. As these tools became more capable, the training requirements increased, and there were two clear paths either make the tool autonomous. So it no longer needs human interaction or work on the interface so that the tool adapts to the human user rather than the other way around.  

But this technology is also critical for machines to understand better the massive amount of existing data created by humans. This technology can then be used to assist with efforts like legal discovery, internal research, and particularly, given the current Pandemic, existing inconsistent medical records critical to the identification of potential cures and remedies. 

Even though IBM, and others, have spun up a large number of supercomputers for this effort, if those machines can’t access the data they need to identify potential solutions, their effectiveness is critically constrained. These powerful machines are only as good as the data they can access, and the NLP effort is being extensively used in this critical research today. It has been instrumental in the identification of a number of the remedies currently in trials. 

Recent advances now allow the technology to identify and reclassify tables, which then can be virtually combined to find trends and promising paths to helpful solutions. These solutions otherwise wouldn’t be found given the obscure, old, or incompatible formats used in the aging source documents and the physical limitations of human researchers.

Talking about research, one critical area where Watson was initially thought to be very helpful was in litigation. But attorneys aren’t programmers, which means you’d need a solution that not only could talk to people that aren’t technical but could take a query, understand its meaning, and then find the critical references or evidence that the lawyer needs without having to program anything new. 

One of the significant potential implementations of this, combined with an advanced AI like IBM’s Watson platform, is a far smarter digital assistant.  Current digital assistants mostly do speech to text conversion, look at past similar queries, and then provide an answer that often has little to do with what the human wanted.  This result is because current systems primarily don’t understand inflection or context. 

Natural Language Processing, as it advances, increasingly attempts to both understand inflection and context.  For instance, let’s say you are watching a movie and recognize one of the actors but don’t know their name.  A future AI with advanced NLP would know what movie you were watching and could have a dialog with you that would identify who on the screen you want to know. 

It will then provide the specific answer you are looking for, so rather than, “I don’t know that, or I don’t have that answer right now,” it will give you the references you need to figure out what show you’ve previously seen them in. In fact, given it will increasingly know what you know, rather than giving you a list of programs or movies the actor has been in, increasingly, it is likely to connect the actor to something it knows you watch. 

Wrapping Up: NLP The Critical Link

The real goal for AI is to have something that will understand your need as a human and provide it. This need ranges from someone that wants their smart home to do what they want without a massive amount of setup and programming to a professional that needs specific advice or unique help to get a specific task done successfully. 

Once this evolves to an advanced level, we’ll be able to more rapidly come up to speed on even the most complicated tool and focus more on creative aspects of the related job rather than the repetitive execution aspects of it. The focus then shifts from the need to use our physicality to our ability to think and create raising the intellectual aspects of a task. Most of us will increasingly be able to take the parts of our jobs we don’t enjoy and have AIs do those parts for us, while we instead focus on the parts of the job we are uniquely capable of doing and should enjoy more. 

One of the things I’m personally waiting for that will come from this effort is the ability to dictate a document and have it result in a complete paper, fully edited and improved so that it better conveys the idea I intended. Eventually, we’ll get there, for me it won’t be soon enough. Granted, I’ll be happy if I can ask a question of my Digital Assistant and be proud of the answer in front of my Luddite friends rather than looking foolish, which, sadly, is too often the case now when getting the wrong or a non- answer. 

NLP is the critical link that will take us from our non-integrated human-machine present to a far better, and far more useful man-machine integrated future. 

Subscribe to Data Insider

Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more.

Similar articles

Get the Free Newsletter!

Subscribe to Data Insider for top news, trends & analysis

Latest Articles