CST offers several forms of collaboration with private companies and public institutions:
Interested in software that automatically extracts specific types of entities from text, tailored to your needs? Or in software that efficiently and accurately classifies documents into categories of specific relevance to your company? CST has years of expertise in rapid prototype development using state-of-the-art techniques in NLP and ML, including data annotation for benchmarking. Previously we have, for example, developed software for extracting names of companies from newswire, software for making legal judgments anonymous, and software for recognizing mentioning of job titles and competencies. Our customers are both private companies and public institutions.
Schultz ApS. This is a test. The company is interested in automatic anonymization of legal documents. We developed various linguistic resources and tools for automatic anonymization of legal documents.
DAMVAD ApS. The company is interested in automatically extracting job titles, competencies, and more from job calls. We developed an annotation interface and had students annotate job calls, making it possible to train robust entity recognition models.
If you don’t know exactly what kind of technology your company needs, maybe you need consultancy? CST has broad expertise in making high-quality workflows more efficient, and in employing state-of-the-art NLP and ML to real-world problems. Consultancy tasks typically range from feedback on presentations of company pipelines to targeted introductions to state-of-the-art technologies in the form of workshops or lectures. Our staff spans a broad range of competencies, from core text processing to data science and applied statistics, also comprising terminological and lexical expertise (how to build vocabularies, word lists, terminologies).
Folketingets Tidende. The Office of the Folketing Hansard writes and publishes official reports of the plenary meetings in the Danish Parliament. We assisted them in making reference books, guidelines and dictionaries available in an online searchable database, consulting on both content and meta-data structuring.
If you’re not ready to specify exactly what you want us to develop, and you need more than just consultancy, CST also offers partnership contracts. A partnership contract buys you 10 hours a week or more of consultancy and development. Typically, contracts go by the month, and you will have a CST staff member at your disposal one or two days a week for consultancy and co-development. This model suits your needs if you want change to happen fast, but do not have the prerequisites to specify in advance what kind of technology is needed.
Infomedia. Sponsored by the EU project CLARIN, one of our researchers spent about 40 hours at Infomedia implementing a Named Entity Recognition system, enabling Infomedia to automatically extract names from texts.
It is relatively easy to estimate the costs for CST Development, CST Consultancy, and CST Partnership, since it is regulated by the administration of the University of Copenhagen. You are expected to pay market price, if available, or the salary of the involved staff members, and cover indirect costs through a 150% overhead. Typically, the total cost for CST Development will be in the range of 50-150,000 DKK. CST Partnerships typically cost about 60,000 DKK a month.
If you are interested in any of these consultancy packages, please contact us at email@example.com.
An alternative to these models is CST Research Collaboration . This is the relevant package for you if you are ready to invest minimum 500,000 DKK in long-term research, typically associated with high risk and high gain. If you are interested in discussing research collaborations, please contact the head of CST, prof. Anders Søgaard (firstname.lastname@example.org).