Artificial Intelligence

The Tritonlytics™ team provides a variety of ML based solutions, from sentiment classification models, to named entity recognition (NER), summarization, and ad-hoc theme discovery. 

View the 2021 NCCI presentation, Using AI to theme 10,000 survey comments in less than 1 hour (PDF), to learn more about artificial intelligence, machine learning, deep learning, and how findings can be made meaningful and actionable.

What is AI/ Deep Learning/ Machine Learning?

We utilize state-of-the-art (SOTA) deep learning methodologies to make sense of the textual data provided by users for the purposes of understanding your organization, defining measurable actions for improvement, and to forecast future satisfaction.  AI, machine learning, deep learning are all roughly synonyms for the same thing: teaching software to learn how to make sense of data it has not seen in the learning process. It gives us a statistically validated approach to make predictions against data that it never saw in its training.

Value Add/ Why is it important in surveys and analytics?

The primary value of AI lies in the ability to create models that beat human baselines against similar tasks and/or provide information that makes those performing such tasks more efficient and less error prone.  As the approach is statistically validated, clients can be confident in the results and a continuous improvement loop can be implemented to improve model performance.

What makes it cutting edge/innovative?

We incorporate SOTA models and methodologies into our machine learning endeavors.  Unlike some of our competitors that essentially call “word counts” machine learning, these are primarily neural networks and tree-based algorithms like RandomForest.