Natural Language Processing has become a key component for many internet first companies that need to process large amounts of textbase data. In this post, Maxwell describes NLP at an entry-level, and how companies can start thinking about using it.
Natural Language Processing (NLP) is a mix of linguistics, computer science, people scratching their heads, and machine learning (but also sometimes artificial Intelligence). This field of machine learning was brought into the limelight in the USA during the 1950s when researchers at Georgetown University thought it would be a good idea to translate Russian to English (I wonder why).
Fast forward to present day, NLP is far beyond simply converting meaning from one language to the next. Due to the advancements in NLP, we are able to ask images questions with words, have chatbots run our online businesses, process all types of healthcare documentation, autocomplete our text messages, emails, and code 😳, even write Dr.Suess poems about Elon Musk (actually A.I). These are all awesome use cases of just a few of the capabilities NLP provides, and the show has just started!
Many tasks in the current labor force are heavily based on data entry. Organizations hire people to extract data, then parse that data, followed by loading that data into their systems. Any job that has a task similar to this is being placed on the "technology is coming to take my task away" chopping block. NLP gives organizations the ability to do high-level text analytics on the ever-growing dataset best known as the internet. This allows for rapid processing of feedback as a result of changing events, customer demands, high website/service traffic, and so forth. To put this into context, the White House has asked our top technology companies to use NLP to parse COVID-19 research literature to aid in the race for a vaccine.
NLP has so many use cases and we haven't even begun to see the full effects of all of the research and technology this field has produced. In order to ensure you don’t miss "the boat", I would advise you to start thinking about how your organization could be leveraging NLP. This could start by simply looking for some low hanging fruit initiatives. If the use cases mentioned above don’t spark your creative juices, head over to Google, and look up your industry, or your job, and see if you can't find someone using NLP in your field.
Once a use case is found the rest is relativity simple and is similar to our conventional development practices. In my experience, I have found that when looking at a workflow or business process from the lens of machine learning, one will begin to see "room for growth". This is a result of the many years that the IT-community built applications for functionality and not data utilization. This can often lead to machine learning projects looking like a massive time sink because of the changes that "can" be made. Keep in mind that it is better to begin collecting more data now, then it is to try and completely redesign an entire process, even if it will be needed in the future. Just my two cents!
Onward! Once you’ve got a good idea of how NLP can be used, and what the outcome should be (i.e. business needs, and data needs), you can begin searching for the right tool. As we all know, before there was an app for everything, there was an API for everything, and the same stands for NLP. Many cloud vendors offer various forms of NLP from text extraction to neural network models that can predict what topic a sentence is describing. Finding the right tooling is key, I recommend any of the big 3 cloud vendors AWS, Google, or Microsoft as they offer a great variety of tools. The last steps consist of development, deployment, and operations.
This is intentionally high level as deploying & maintaining machine learning models is a whole field of study (AIOps) which is out of the scope for this post. (Depending on feedback, I can dive into this) (starting place)
Machine learning is an inflection point in the evolution of humanity. It's become quite popular within our industries and fields of study, but make no mistake about it, machine learning will breed highly capable artificially intelligent systems that will continue to drastically change how we live. It started in research labs but has drastically made way into our markets. During your next innovation workshops, or yearly planning meetings, start a discussion on how your organization can better leverage its data by integrating NLP.