Center for Knowledge

The success of  Large Natural-Language Models (LLMs) has been getting more and more dramatic during recent years. It has, even, led to a torrent of new AI applications outside of the NLP field. Alas, this success is afflicted by the models’ inability to prove that their predictions were made on good premises. We know their inputs, outputs, and how likely it is for the outputs to be correct, but we still don’t have a clear understanding of what happens in between. 

Some of the most common traits of recent large language models (LLMs) are their ever-increasing sizes (measured by their number of parameters) and the size of their training data. 

Citibeats collects people’s opinions on social media data and finds actionable insights. With AI tools,  we have developed different models to structure vast amounts of data and provide context on users’ and citizens’ needs.

Introduction

At Citibeats, we often attempt to build new classifiers to gain insights from social media posts and provide the best context we can to our users around citizens' needs. For instance, we developed some tools to extract information from those posts: a request detector, gratitude detector, concern detector, human/bot or male/female detectors.

Every day, our data science team tackles challenges to improve our product. To give you insight into the process, we will describe the development of our multilingual complaint detector for social media texts and explain how we solved implementation issues.

*DISCLAIMER* We tried to limit technical language to make this article approachable to a wider audience. 

Every day at Citibeats, we strive to gain a deeper understanding of people's opinions with the intention of having their voices be heard — by leaders, decision-makers, or whoever is best positioned to address their needs.