Hi @sdascoli and team,
Incredible work on TRIBE v2. The multimodal brain encoding results and the zero-shot generalization are genuinely groundbreaking.
We've built Synaptik Mind (AI Interfaces Inc) — a real-time neural intelligence platform that bridges consumer brain-sensing wearables (EEG headbands, fNIRS, eye tracking) to cognitive state classification APIs. We've built the full pipeline: signal ingestion, artifact rejection, band decomposition, cortical mapping to fsaverage5, and 8 cognitive state classifiers running at 302 classifications/sec.
(NEW) Live product: https://synaptikmind.lovable.app/
We'd like to integrate TRIBE v2 as the core encoding layer in our platform — specifically using the pretrained model to predict brain responses to content (video, audio, text) and visualize cortical activations for developers and consumers.
Two questions:
Commercial licensing — The CC-BY-NC-4.0 license covers our R&D and prototyping. For production deployment, who should we contact to discuss commercial licensing terms for the model weights? Is there a Meta partnerships or licensing team that handles this?
EEG-to-TRIBE bridge — Has the team explored or considered mapping non-fMRI signals (e.g., EEG topographic features or fNIRS hemodynamic responses) into TRIBE v2's latent representation space via transfer learning? We're building this bridge and would welcome any architectural guidance.
Happy to share our full technical architecture. We see this as a mutually beneficial partnership — our platform would generate large-scale real-time brain data from consumer devices that could feed back into TRIBE's training pipeline.
Thanks for making this open to the research community.
Look forward to hearing from you!
Thank you,
Rob
AI Interfaces Inc | Synaptik Mind
https://synaptikmind.lovable.app
Hi @sdascoli and team,
Incredible work on TRIBE v2. The multimodal brain encoding results and the zero-shot generalization are genuinely groundbreaking.
We've built Synaptik Mind (AI Interfaces Inc) — a real-time neural intelligence platform that bridges consumer brain-sensing wearables (EEG headbands, fNIRS, eye tracking) to cognitive state classification APIs. We've built the full pipeline: signal ingestion, artifact rejection, band decomposition, cortical mapping to fsaverage5, and 8 cognitive state classifiers running at 302 classifications/sec.
(NEW) Live product: https://synaptikmind.lovable.app/
We'd like to integrate TRIBE v2 as the core encoding layer in our platform — specifically using the pretrained model to predict brain responses to content (video, audio, text) and visualize cortical activations for developers and consumers.
Two questions:
Commercial licensing — The CC-BY-NC-4.0 license covers our R&D and prototyping. For production deployment, who should we contact to discuss commercial licensing terms for the model weights? Is there a Meta partnerships or licensing team that handles this?
EEG-to-TRIBE bridge — Has the team explored or considered mapping non-fMRI signals (e.g., EEG topographic features or fNIRS hemodynamic responses) into TRIBE v2's latent representation space via transfer learning? We're building this bridge and would welcome any architectural guidance.
Happy to share our full technical architecture. We see this as a mutually beneficial partnership — our platform would generate large-scale real-time brain data from consumer devices that could feed back into TRIBE's training pipeline.
Thanks for making this open to the research community.
Look forward to hearing from you!
Thank you,
Rob
AI Interfaces Inc | Synaptik Mind
https://synaptikmind.lovable.app