We'd love to hear your feedback, questions, or suggestions!
Thank you! Your message has been received.
Discover KU researchers who share your scientific interests. KUnnect covers researchers with three or more publications on CURIS since 2021, and uses only those publications to build their profile.
An analysis usually takes 1–2 minutes. KUnnect is a prototype running on a single server rented by the university, so your text is never sent to a third-party AI service. Analyses run one at a time and may be queued at busy moments due to limited capacity.
You can leave this tab open, the results will appear here when they are ready.
Starting…
Loading results...
KUnnect is a prototype intended to help us explore how to develop and deploy AI-based applications at KU, particularly with the use of open models. Content is generated by AI and may contain inaccuracies.
This prototype was developed by Stefano Pupe from the Center for Health Data Science (HeaDS), on behalf of the Deanery of the Faculty of Health and Medical Sciences. It has been approved for release by KU-IT.
KUnnect runs on a single-GPU server in Denmark under a 100% renewable energy policy. Building it cost an estimated 400 kg CO₂e, and a live query that requires ~1 minute of GPU adds about 2 g each time. Estimated with Green Algorithms (Lannelongue et al., 2021).
Questions, corrections and ideas are welcome through Contact.
KUnnect turns public publication records into meaning-based profiles, then matches researchers working on similar ground. Here is the path from raw data to a recommendation.
KUnnect is built on publicly available researcher profiles from CURIS, the University of Copenhagen's official research information system. The dataset covers approximately 7,000 researchers whose profiles were publicly indexed in CURIS as of August 2026 and who had at least three publications recorded since 2021. A profile also needs at least one of those publications to have a readable abstract, so a small number of otherwise eligible researchers are missing. Only publicly accessible pages were used; no login or restricted database was involved.
Each researcher is tracked through a unique profile identifier, so author names are never guessed or confused between people who share a similar name. Their publications are linked back to that single identity before any analysis begins.
KUnnect uses a technique called semantic embedding to capture the meaning of each researcher's work, going beyond titles to the content of the research itself. A language model called Nomic Embed Text reads each publication abstract and converts it into a list of 768 numbers encoding the meaning of the text. Trained on vast scientific literature, the model knows that "myocardial infarction" and "heart attack" refer to the same concept, and that a paper on Arctic ice cores is closer in meaning to one on permafrost than to one on corporate tax law. Each researcher's publication fingerprints are then averaged into a single profile representing the overall territory of their recent work. Only publications from 2021 onwards are used, so the profile reflects what a researcher works on now rather than their full career.
When you search for a researcher, KUnnect first identifies their existing co-authors from papers published since 2021. KUnnect then searches the full pool of roughly 7,000 researchers for those whose topic profile is closest to the queried researcher's, surfacing people covering similar scientific ground who have not recently collaborated. Researchers who collaborated with them prior to 2021 are not excluded and may appear as potential matches. Co-authorship counts no matter how many authors a paper had, so a large consortium paper can rule out someone a researcher has never worked with directly.
A language model then reads papers from both sides and writes a short explanation of the potential overlap — the recommendation text you see in your results. On a name search it is shown up to 20 of that researcher's most recent papers and up to 10 from each candidate. Name searches use Qwen3.8 27B; the SUND collaborates grant tool, the Describe research box and the Upload a PDF analysis use Gemma 4 31B. A smaller model, Gemma 4 12B, reads each submission first and declines the ones that are not research descriptions at all. All are open-weight models running on our own server, and the two that write recommendations were chosen by benchmarking candidates against each other on real searches, scored for whether every person and paper they cite is one they were actually shown.
The Research Map groups roughly 59,000 KU publications from 2021 onwards into clusters, based on the similarity of their titles and abstracts. The same semantic embeddings used for collaborator matching are clustered so that publications on related topics sit together, and each cluster gets a label and a short summary written by a language model. Each cluster is then classified within the OpenAlex academic taxonomy: a language model reads the cluster's label and description and maps it to the best-matching field and topic from OpenAlex's hierarchy of roughly 4,500 topics across 26 fields. The result is an interactive overview of the research landscape at the University of Copenhagen, where you can explore each area's themes and representative publications.
Recommendations reflect topical similarity based on published abstracts. They are a starting point for your own judgement, not a guarantee of a productive match. Researchers with a very broad set of interests may be harder to represent accurately, since averaging across many different topics can blur their profile. As the content is AI-generated, it may include errors and misrepresentations.
KUnnect is built exclusively on publicly available researcher profiles from CURIS, the University of Copenhagen's official research information system. No restricted databases, logins, or private records were accessed.
We do not store what you search for or submit. Uploads are processed in memory and never written to disk; results are deleted as soon as you retrieve them, or within an hour if you don't. We keep no IP addresses and set no tracking cookies.
To know how much the tool is used, we count visits and analyses. Each is filed under a code derived from your network address and browser using a secret key that is destroyed every night, so yesterday's counts can no longer be traced to anyone, not even by us. We record which feature was used, when, whether it produced a result and how long that took, but we never permanently store the name, text or file you sent. Your browser also keeps a note that it has been here before, so we can tell new visitors from returning ones without personally identifying anyone.
Alongside that count we note two things about each visit, as a plain yes or no: whether it came from an automated crawler rather than a person — most of our traffic does — and whether the network address is a Danish one, which tells us roughly how much of our audience is here at home. Both are worked out and then discarded: we keep the yes or no and never the address or the browser identifier behind it, we do not record which country a visit came from beyond Denmark or not, and nothing about your visit is sent to any outside service to determine either.
KUnnect – Innovation, the SUND research stories and landscape at /stories/, counts its use the same way, in records of its own: visits to its pages, whether a search found anything and how long it took, which filters it used, and that a side panel was opened. It never stores the topic you type, or which research area or researcher you looked at. It keeps its own note in your browser, and reads whether KUnnect's is there. Because both use the same nightly key, we can see how many visitors used both on the same day, never across days.
Text submitted to SUND collaborates or the Describe research box, and the text read from an uploaded PDF, is screened by a small language model before the analysis runs. Whether it accepted or declined, and the one-line reason it gave, are kept alongside that count so we can tell how well the screening works; your description or paper itself is not. Server logs, which contain no search or upload content, are deleted after 14 days. Contact form submissions are kept for six months so we can reply, then deleted automatically.
All computation runs on a server hosted in Denmark by Scannet, under a 100% renewable energy usage policy. We are considering options of hosting it in servers managed by KU-IT in the future.
The language models used for embedding and generating recommendations run locally on that server. No data is sent to external AI providers such as OpenAI or Anthropic. Recommendations are written by Qwen3.8 27B for the main search function, and by Gemma 4 31B on the SUND collaborates page and for the Describe research and Upload a PDF analyses. Gemma 4 12B screens submissions before they are processed. Publications are embedded with Nomic Embed Text v1.5.