No-code, large-scale implementations and multimodal approaches will define NLP in 2022

NLP at a glance

When most people think of Natural Language Processing (NLP), Voice assistants like Alexa and Siri come to mind. While human-machine interaction has come a long way, it is only scratching the surface of what technology can do. In fact, the most shocking use of NLP does not involve speech at all.

But let’s start by defining NLP. Technology is a subset of artificial intelligence (AI) and machine learning (ML) that focuses on enabling computers to process and understand human language. While speech is part of it, the most striking progress in NLP lies in its ability to analyze written text.

As such, NLP is largely happening behind the scenes. But people interact with him more than they realize. From hospitals to financial services to law firms, NLP is powering a fair amount of data reading, writing, and analysis used to manage these services. Even as a young tech, he has made his mark on the company over the past few years.

What’s next for NLP?

All signs point to this growth continuing. Even in the midst of the global pandemic, Spending on NLP was on the rise, while overall IT spending was affected. Investments in NLP have continued on that trajectory this year. According to a recent industry survey93% of tech leaders indicated that their NLP budgets grew by at least 10-30% compared to 2020.

2021 was a promising year for NLP. Use cases have helped users with everything from identifying fake news and toxic content online, to helping accelerate clinical trials through better candidate selection. Even beyond healthcare and the media, NLP is proving its worth in every industry.

But there are several factors that can take this growth to the next level in 2022. No-code software, advances in large-scale implementations and multimodal approaches to NLP will contribute significantly to its growth in the coming year. This is why:

Low code is great

Low code requires little or no coding experience to build an application. Unsurprisingly, low-code solutions had a moment last year. Simplifying NLP is a sure way to ensure continued growth in the field. It allows professionals of all levels to use technology. In short, the execution of many of the more complex deep learning models can now be reduced to a single line of Python code.

For newbies to NLP, this lowers the barrier to entry. Having a formal education and practical experience with foundational libraries of NLP, deep learning, and transfer learning used to be a prerequisite. Now, anyone can start with a basic understanding of the technology.

This is not only valuable to those new to the field. For data scientists, simplification enables a level of automation that allows them to focus on more important jobs. This will become increasingly important as AI talent shortage persists. Low-code solutions have benefits across the board, and fortunately, we see more of them every day.

Codeless AI becomes a reality

In 2022, we will build on the low-code trend with no-code software. This will make artificial intelligence and machine learning more accessible to anyone. By putting more power in the hands of domain experts, you eliminate the need for a data scientist, further democratizing NLP. We are already seeing this begin to unfold.

Consider creating a website, for example. What once required coding proficiency can now largely be done by a graphic designer. This is how the no-code will leak to users outside of the developer’s title. It will also help refine NLP for specific business use cases. After all, if you’re building healthcare artificial intelligence models to detect COVID-19 on a lung X-ray, you want a doctor to consider more than a data scientist.

The shift from data scientist to domain expert will be gradual, but we will see much easier to implement no-code options to facilitate this in the coming year. This is similar to the difference between paying programmers to write code and having Excel. No-code is designed for a different set of non-technical users. Finally, there is a class of tools that allows you to become familiar with NLP.

Tuning models to implement at scale

At aforementioned survey, technology leaders cited accuracy as the most important factor when considering an NLP solution. That said, the difficulty in tuning the models was one of the biggest challenges that tech leaders cited. Unfortunately, continuous tuning of the models is critical for accurate results. Equally important, it prevents them from degrading over time.

Healthcare is an industry where continuous monitoring and adjustment is especially important. Technology assumes that fixing a person is like fixing a car. If something is broken, you can simply scan an academic article or medical journal and apply a solution to fix it. But humans are not that simple. There are many factors at play. Medical history, social determinants of health, how your doctor interprets your results compared to another are just a few.

By allowing domain experts, in this case medical professionals, to fit the models, we allow them to correctly fit the models to specific situations. Very often it is necessary to adjust the models separately on a larger scale. This is because the models work differently in different production environments. Even if you are both in a clinical setting.

In recent news, a retrospective study in JAMA Internal Medicine found that a model developed to predict sepsis in patients failed to identify two-thirds of those affected. While some providers reported the success of the tool, researchers at the University of Michigan School of Medicine found that the results were substantially less accurate when applied to their own patients.

Considering how models will behave in different settings in different populations can make the difference between life and death in healthcare. But it is also important in other industries. The good news is that we are getting better at this. Now, we enable users to implement scale models faster and more accurately than ever.

Multimodal solutions go beyond NLP to the next level

Human language is not black and white. We interpret the meaning of written language, speech, pictures, and more. As a result, we need machine learning techniques that are able to “read”, “see” and “hear” all at the same time. Multimodal learning techniques, using different data modalities by combining tools such as NLP and computer vision, are key to these use cases.

While NLP models are great for text processing, many real word applications use documents with more complex formats. For example, healthcare systems often include visual lab results, sequencing reports, clinical trial forms, and other scanned documents. When NLP is used for document understanding only, design and style are compromised.

However, with new advances in multimodal learning, models can learn both from the text of documents through NLP and from visual design through technologies such as computer vision. Combining multiple technologies in a given solution to enable better results is the core of multimodal learning. We are starting to see more of this move from research to production.

2021 has been a standout year for NLP and we can expect it to continue into the new year. With easier-to-use tools, more accurate results, larger deployments, and matchmaking capabilities with other powerful AI technologies, it will be interesting to see where 2022 takes us.

David talby

David talby

Chief Technology Officer, John Snow Labs

I help companies build real-world AI systems, turning recent scientific advances into products and services. My specialty is the application of machine learning, deep learning, and natural language processing in healthcare.

Leave a Reply

Your email address will not be published. Required fields are marked *