At Topos Labs, our mission is to democratize AI with the goal of empowering non-technical professionals to uncover the hidden value and risks lurking in their exploding data. Out of this painstaking process, we created a new career: the Dictioneer.
Everyone knows that Natural Language Processing (NLP)–teaching computers to read, analyze and interpret human language–is super easy. And by ‘easy’ I mean that it is really, really difficult. See what I did there? I used sarcasm. It’s one of the silly things humans do. We sometimes intentionally invert the meaning of our own words to convey information in a mocking tone that elicits an emotional response. Mr. Spock would say this form of communication is “highly illogical” and computers would likely agree. And sarcasm is just one small example of the seemingly limitless complexities, inconsistencies, contradictions, nuance, subjective interpretation and context that exist across thousands of worldwide languages and dialects. This helps explain why automating the understanding of human language is a massive and confounding challenge. And it ultimately means that for the foreseeable future, high-quality NLP will rely on the assistance and guidance of humans…or as we refer to them at Topos Labs: “Dictioneers”.
A Dictioneer (short for dictionary engineer) is a professional who is tasked with training our NLP platform ‘Gracie’ on new domains of knowledge (e.g., sports, science, religion, etc.) or document classifications (e.g., resume, patent, contract, balance sheet). Dictioneers also teach Gracie how to identify a variety of risky artifacts such as personally identifiable information (e.g., name, ssn, credit card info, address), source code, sexually explicit messages, harassing language, or even deceit. However, the truly exciting aspect of the Dictioneer’s role is not what they do but what they don’t need to do. Dictioneers don’t need to be machine learning experts, data scientists or software programmers. In fact, they don’t need any real technical skills at all. Gracie enables non-technical domain experts to create their own highly accurate NLP models in minutes without writing a single line of code.
This is made possible by a simple point-and-click user interface that virtually anyone can use with some basic training. Our dictioneers at Topos Labs have predominantly liberal arts educations: artists, undergrad and graduate students, designers, writers, and retired executives. Dictioneers at our customers are admin assistants, researchers, analysts, customer support representatives, paralegals, etc.
Need to sort through thousands of documents to find all your contracts or resumes? Sounds like a job for a dictioneer. Need to find all customer agreements that expire on or before June 1, 2021? A dictioneer can do it. Need to find all emails that contain a call-to-action and a sense of urgency? Call a dictioneer.
Democratizing AI isn’t just about integrating automation into existing products. It’s about putting the power of this amazing technology into the hands of your knowledge workers, without the usual heavy reliance on IT. Everyone loves to be reliant on IT. And by “loves” I mean hates. But you already knew that… unless you are a computer of course.
To learn more about our technology or opportunities at Topos Labs contact us at [email protected]