Attribution
These weights are derived from, and are only meaningful together with, the following sources.
Connectome (network wiring, synapse signs, neuron positions, retina columns)
FlyWire whole-brain connectome of an adult female Drosophila melanogaster, Codex snapshot 783
(connections.csv.gz, neurons.csv.gz, classification.csv.gz, consolidated_cell_types.csv.gz,
coordinates.csv.gz, column_assignment.csv.gz). Licence of the Codex release: CC BY-NC 4.0
(https://flywire.ai/guidelines). The Zenodo mirror of the connectivity data is CC BY 4.0
(doi:10.5281/zenodo.10676866).
- Dorkenwald, S., Matsliah, A., Sterling, A. R., Schlegel, P., Yu, S.-C., McKellar, C. E., et al. Neuronal wiring diagram of an adult brain. Nature 634, 124–138 (2024). doi:10.1038/s41586-024-07558-y
- Schlegel, P., Yin, Y., Bates, A. S., et al. Whole-brain annotation and multi-connectome cell typing of Drosophila. Nature 634, 139–152 (2024). doi:10.1038/s41586-024-07686-5
- Matsliah, A., Yu, S.-C., Kruk, K., et al. Neuronal parts list and wiring diagram for a visual system. Nature 634, 166–180 (2024). doi:10.1038/s41586-024-07981-1 (optic-lobe column assignments)
- Eckstein, N., Bates, A. S., Champion, A., et al. Neurotransmitter classification from electron microscopy images at synaptic sites in Drosophila melanogaster. Cell 187, 2574–2594 (2024) (transmitter predictions behind the synapse signs)
- The FlyWire Consortium. FlyWire Whole-brain Connectome Connectivity Data (783.0) [Data set]. Zenodo (2024).
Training data
- Lichess open database, https://database.lichess.org — monthly standard rated games 2014-01 … 2015-12 and
the evaluation database (
lichess_db_eval.jsonl.zst). Released under CC0 1.0. - Stockfish (https://stockfishchess.org, GPL-3.0) produced the evaluations in the Lichess evaluation database; no Stockfish code is included here.
Modelling ideas
- Lappalainen, J. K., et al. Connectome-constrained networks predict neural activity across the fly visual system. Nature 634, 1132–1140 (2024) — connectome-constrained training with sign constraints.
- Shiu, P. K., et al. A Drosophila computational brain model reveals sensorimotor processing. Nature 634, 210–219 (2024) — whole-FlyWire-brain leaky integrate-and-fire modelling.
Software
- Training and inference code: https://github.com/cesp99/fly-chess (MIT, Carlo Esposito).
- PyTorch (BSD-3), python-chess (GPL-3.0, used as a library by the training code only; not part of the weights).