Track genomics, neuroscience and molecular biology.
Search papers, labs, and topics across Lattice.
This paper evaluates the "fly hashing" algorithm, a biological model of the Drosophila olfactory circuit, against four real-world electron microscopy connectomes. While the algorithm's locality-sensitive hashing advantage over random projections is confirmed, this advantage is found to be per active cell rather than per operation, and the specific biological wiring offers no consistent retrieval benefit over degree-preserving random rewiring. Furthermore, while uneven glomerular fan-out improves retrieval in the model, this biological feature is weakly related to odour tuning and not offset by structural synapse counts.
The fly's olfactory hashing advantage is per active unit, not per operation, and its complex wiring offers no retrieval boost over random rewiring.
Dasgupta, Stevens and Navlakha (2017) showed that the Drosophila olfactory circuit, modelled as a random sparse projection followed by winner-take-all, is a locality-sensitive hash that beats classical LSH. The projection was random because the wiring was unknown. We test it against four electron-microscopy connectomes (MaleCNS, hemibrain, FlyWire, BANC; four animals, seven hemispheres). First, the 2017 pattern holds in a reimplementation of its protocol on SIFT, MNIST and odour mixtures (GloVe is near chance at short codes for every method): the fly hash beats k Gaussian projections at short hash lengths (3.1x in AP@200 on MNIST at k = 4). Second, against the tested real-valued Gaussian baseline that advantage is per active cell, not per operation: Gaussian projections given the same projection arithmetic retrieve better on every dataset and input dimension tested. Third, across four connectomes the measured pairing of glomeruli gives no consistent retrieval advantage over degree-preserving rewiring: retrieval is slightly lower (median -1.6%), and the small odour deficits depend on how missing odour responses are treated. Separately, equalising glomerular fan-out at fixed connection count improves retrieval in the model in every hemisphere, while equalising inputs per cell lowers it on average. Yet the fan-out profile is similar across the four sampled animals (median between-animal Spearman rho = 0.86), structural synapse counts do not offset it, and its relation to odour tuning is weak. In the model that uneven allocation costs retrieval. For practice: measured wiring gives no consistent retrieval advantage over degree-preserving random wiring, so a fly hash needs no connectome data, and its advantage is per active unit, which may suit hardware where active units rather than arithmetic are the binding cost, a hypothesis we do not test.