Ecosystem map

Who is in this ecosystem, what each brings, and how firm each relationship actually is. Prominence here is deliberately unequal — see Editorial safeguards for the tiering rule.

The map

TierPartnerBringsFirmness
1 — HubNIPG / ELTEThe method layer: multimodal feature extraction, speech/face/gaze/blink processing, multimodal transformers, 3D perception, composite-AI reasoningThe convening institution
2Semmelweis / VIKOTEA real patient population and a funded avatar-therapy trial (NKFIH ADVANCED_25, entry 153443)Funded grant; Speech Gap integration under consideration
2Speech Gap / University of SzegedThe concrete measurement technology, patented, sensitivity demonstrated in two languagesGranted patent family; integration planned, not yet registered
2Argus CognitiveThe translation case study — what this research looks like as a productWorking connection via researchers’ prior ELTE affiliation; not a spin-off, no formal agreement
3Michael E. Benros / DenmarkLarge-scale registers, biobanks, immunogenetics, omics; external validation potentialOpportunity; no data access established
3Erik Szabó / White UnicornCapital, incubation, MDR/FDA network, the five-test viability screenRelationship stage; workshop participation confirmed
4Supporting collaborationsSorbonne, Delft/MeMo, DFKI, Richter/Colleyeder — credibility and transferable experienceHistorical or one-off

The connective logic

Every technical thread touches the hub. Speech Gap’s inventor pool overlaps with NIPG-adjacent researchers at Szeged/HUN-REN; the ADOS-2 work involves ELTE researchers with Rush and Sorbonne; the avatar and rehabilitation work is an NIPG line; the DFKI collaboration runs through NIPG; Richter/Colleyeder originated as an ELTE-adjacent initiative; Argus’s researchers largely came out of ELTE.

That is what makes NIPG the hub rather than merely a participant: it is the only node connected to all the others.

The four activities this map supports

  1. Presenting existing research, technical, clinical and translational capabilities.
  2. Identifying data resources and opportunities for harmonised new data collection.
  3. Converting collaboration ideas into reviewable and auditable projects.
  4. Matching projects with national, European, foundation and industrial funding.

What is actually agreed, as of this wiki

StatusItem
Funded and runningThe VSMT avatar-therapy trial at Semmelweis
Planned, not yet registeredSpeech Gap longitudinal integration; ClinicalTrials.gov registration; PCI RR Stage 1
Under considerationMultimodal and camera-based extensions to therapy monitoring
Opportunity onlyDanish register/biobank collaboration; MILAB-style programme; venture route
HistoricalDFKI tooling, Richter/Colleyeder, the educational-games archive

The gap this ecosystem is designed to close

Research capability exists. Clinical access exists. What usually fails is everything between them: specification, data governance, validation, regulatory framing, and someone willing to fund the unglamorous middle. → Clinical AI translation

Overview · INNOFAIR workshop programme · Programme opportunities · Project structure in Plane · Funding landscape · Editorial safeguards