PROJECTS

Research and student projects at the Center.

Flagship project April 2026 – April 2029

Affective-Cognitive Platform for Adaptive Learning and Simulation Training

ACTutor · ACTrainer · ACInsight SDK

How much a trainee learns depends on their cognitive and affective state as much as on the material in front of them. This platform reads that state from camera, microphone and wearable signals, and adapts the lesson — or the simulation — to it while it is happening.

  • €3.12 M Project value
  • €1.90 M EU co-financing
  • 3 Partner institutions
  • 36 Months

Three parts

ACInsight SDK
The base technology: multimodal assessment of cognitive-affective states, which the other two build on.
ACTutor
A plug-in for existing e-learning platforms, adapting the curriculum and how it is delivered.
ACTrainer
An add-on for professional training simulators, which also asks whether a simulation stimulates a trainee the way the real thing would.

Measured Cognitive load, engagement, valence and arousal — a working set, not a closed one.

Built with

  • Visage Technologies (lead)
  • University of Zagreb, Faculty of Humanities and Social Sciences
  • University of Rijeka, Faculty of Maritime Studies

From the Center

Renata Geld Principal investigator, partner leadFrancesca Dumančić ResearcherTanja Manabe Researcher

First piloted on the ship-navigation simulator at Rijeka, and built to carry across to other kinds of learning. Designed against the GDPR, the EU AI Act and NIS2.

Co-funded by the European Union

Industry Research and Innovation (IRI), EU Structural Funds. Co-financing €1,904,735.93 of a total €3,117,060.40.

Views and opinions expressed are those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them.

Research projects

  • The DeSCarTes project mark

    Ongoing Croatian Science Foundation 15 January 2024 – 14 January 2028

    DeSCarTes

    Determining students’ characteristics using educational robot technologies

    Can an educational robot tell you something measurable about a child’s attention, perception, memory or spatial orientation?

    Educational robots are the logical successor to toys: a familiar object that ties the abstract to the physical, invites invention, and lets a whole class work at once. The project designs workshops around them and records what happens — data from the robot alongside the children’s physiological signals — then analyses it with machine learning and checks it against standardised tests of cognitive ability.

    The point is practical. A teacher who can see which child is struggling with what can give that child the right support.

    Project lead
    Ana Sović Kržić FER
    With
    Ivana Storjak and Ljiljana Puškar (FER)
  • The Red Hen Lab mark

    Ongoing International cooperative

    The International Distributed Little Red Hen Lab™

    A global big-data laboratory and cooperative for research into multimodal communication — how meaning is carried by speech, gesture and image together rather than by words alone.

Student projects

2024/25

Students presenting the EPlaG project to a full lecture hall

EPlaG — Educational Platform for Genetics

with Genom d.o.o. and Genkonekt

Students on Introduction to Anthropology built an educational platform for genetics with Genom d.o.o. and Genkonekt. They ran the qualitative, quantitative and market research behind it, recorded a podcast and video at FFZG and FER, and presented the result on 20 January 2025 to a full Conference Hall at the FFZG Library — ending with a live audience quiz on genetics and biotechnology, which a geneticist won.

Dr Maja Barbalić, director of Genom and head of Genkonekt, assessed the work and raised the prospect of further joint projects.

2023/24

The S.A.N.J.A. team being interviewed by N1 at BIRD Incubator

AI Izbori — forecasting the 2024 parliamentary election

with N1 television and BIRD Incubator

Bartol Banić and Luka Prpić, both in our first cohort, were two of the five members of S.A.N.J.A., the team that won AI Izbori — a series run by N1 television with BIRD Incubator in which student teams built models to forecast the 2024 Croatian parliamentary election. More than twenty teams applied, three reached the final, and the finalists’ forecasts were broadcast three days before the vote.

S.A.N.J.A.’s forecast was the most accurate in the competition, and more accurate than the national television polls. It placed all five leading parties within a single seat of their result — HDZ 60 against 61, SDP 43 against 42, Domovinski pokret 15 against 14, Most 10 against 11, and Možemo! 9 against 10 — where the polls were out by two seats on two of the five.

Hosted by

  • Students at a workshop on content analysis, watching a presentation

    Media net and Presscut

    Workshops on media knowledge management and content analysis.

    Presscut has been Croatia’s media monitoring agency since 1989 and is part of the Newton Media Group. It reads over 2,000 news portals and 340 print titles a day, and keeps an archive of the Croatian press going back to 2000.

  • Students visiting AVL in Zagreb

    AVL

    A day at the Zagreb engineering centre — how its teams are organised and what they are working on.

    AVL-AST d.o.o. has been AVL List’s Zagreb research and development affiliate since 1996, with over 250 engineers simulating, testing and optimising propulsion systems — increasingly electrified ones. It opened a Powertrain Engineering Center here in 2019 and a Power Electronics Center and Lab in 2021. AVL funded five scholarships covering full tuition of our first cohort’s programme.

  • Students visiting Poslovna inteligencija in Zagreb

    Poslovna inteligencija now Solvership

    An introduction to the company and to the work its data teams do.

    Founded in Zagreb in 2001 and now trading as Solvership, the firm works across Southeast Europe in data engineering, data science, business intelligence, risk management and data privacy.

  • Students pitching ideas at Sofascore

    Sofascore

    A brief set by the UX research team, revealed on the spot, and pitched back the same day.

    Sofascore has been building live sports data out of Zagreb since 2010 and has more than 25 million users. It turns match data into things people can actually read — Attack Momentum, heatmaps, shotmaps and a per-player rating.