Natural Language Processing: Looking Ahead to 2021
The impact of the COVID-19 pandemic has been felt across the globe. With many businesses struggling, shuttering, and innovation put on pause to survive daily operations, it’s been a tough year. However, there is one technology that has flourished in 2020, and is poised for even greater growth in 2021 and years to come: Natural Language Processing (NLP). While worldwide IT budgets shrank 8% from 2019 (Gartner), investment in NLP increased across organization industries, company sizes, and locations, from 10-30% (Gradient Flow).
A clear silver lining in the tech industry this year, NLP is supporting projects from helping financiers read SEC filings to accelerate recruitment for clinical trials. You may even be unknowingly interacting with NLP in your everyday life through customer service chatbots and virtual assistants. NLP is everywhere, but what are the driving forces behind this rapidly growing subset of AI? Here are 5 of these contributing factors that will help NLP shine in 2021 and beyond:
Accuracy Will Remain the Key Focus
According to recent research, more than 40% of all respondents cited accuracy as the most important criteria they use when evaluating an NLP library.
Accuracy refers to pre-trained models that get used in multi-stage pipelines in NLP libraries. Typically, accuracy is evaluated against standard academic benchmarks, and is vital for highly regulated industries such as healthcare and finance, where even small misinterpretations can have big implications. New academic research is helping providers of NLP technology challenge the status quo, making it possible for users to apply new, highly accurate pre-trained models into production almost immediately.
As a result, we are now able to turn more and more use cases from aspirational to commercially viable. As the state-of-the-art continues to advance, so too will the technological capabilities NLP has to offer.
Organization-wide Education Will Take NLP to the Next Level
As important as technical talent is to implementing and scaling an NLP project, understanding how AI will work within a product from a business perspective is equally vital to success. In healthcare, for example, the most accurate patient risk prediction model won’t help anyone if it’s not integrated into the clinical workflow, easily trusted, and used by doctors and nurses.
All disciplines within an organization need to understand the benefits of integrating AI, and how it will affect their job function. Failing to train and actively involve your entire organization on the benefits and uses of AI and NLP systems is why so many projects don’t succeed. An organization-wide investment in time, education, and practice is what will truly determine NLP’s value.
Ease-of-Use Will Further Democratize NLP
Running some of the most accurate and complex deep learning models in history has been reduced to a single line of code, such as Python’s NLU library. Many NLP libraries now provide official support for their published models, so pipelines are regularly updated or replaced when a better algorithm, model, or embedding comes along. Additionally, with the advent of NLP model hubs, thousands of free, pre-trained models are available to the masses. To help take the guesswork out of finding the right model for a particular project, better-faceted search, curated suggestions, and smarter ranking of search results are coming to fruition, too. This level of support and access will make it easier for NLP novices to get started and for advanced data scientists to get work done faster.
- Multi-lingual support will greatly improve
In addition to ease of use, multilingual availability is putting NLP in the hands of users worldwide. Cloud providers now offer support for over a hundred languages, and libraries like Spark NLP now offer 46 languages and are continuing to expand. With new research advances such as language-agnostic sentence embeddings, zero-shot learning, and the recent public availability of multilingual embeddings, this will become the norm. More access to code and the availability of many languages evens the NLP playing field globally, resulting in a more diverse and inclusive AI ecosystem, and more opportunities for innovation with the technology. This is a big step, considering just a few years ago, NLP models were typically only available in English, Mandarin, and a handful of other languages.
The Translation of Research to Real-World Systems Will Accelerate
Thanks to the aforementioned trends, recent advances in deep learning and transfer learning are now moving from research to production. NLP is already being used for important work, such as diagnosing patients, matching them to clinical trials, highlighting high-risk situations, and enabling faster drug discovery—something that has been of particular importance this year as we continue to fight the COVID-19 pandemic.
Projects like these, already being used in healthcare environments, have the potential to lessen the impact of workers, improve operations, and will only get better with time. This is true for all industries, and it will be exciting to see these come to fruition in the new year.
These 5 trends, and likely many more advances to come, will help propel NLP forward to another year of wide adoption, production-grade projects, and more accessibility for all to use. Still in its infancy, NLP has the potential to impact business for the better, and we look forward to being part of those efforts in 2021.