SenseHive is a crowdsourcing platform for collecting word usage data from everyday language speakers. Empowered by the latest insights from natural language processing and machine learning, SenseHive gradually builds a network representing the word meanings through the efforts of a large crowd of people who are eager to help.



Words are the basic building blocks of language. Through usage, they grow, flourish, and change, often taking on new meanings in the process. Is the fan that cools you the same one that loves your work?

For humans, understanding different word meanings is relatively easy. For machines, though, it is a daunting challenge. Having this ability would make the computers better at many important tasks − search, translation, question answering, and even user interaction.



SenseHive asks each participant, a worker bee, to answer simple questions about a word in context. With every response, it learns a bit more about the word's meaning.

By intelligently coordinating the work of a multitude of worker bees, SenseHive builds a rich hive of word senses, something a lone worker bee couldn’t hope to accomplish.



Our goal is to build sense hives for various languages, and make them available to others: they can build useful tools on top of them, study the words, or simply admire the beauty of word meanings.

Our first objective is a sense hive for the Croatian language, covering at least 10,000 senses.