Mahzarnia, Ali, Jamshidi, Homayoun · Nov 2, 2026
Mahzarnia and Jamshidi add LiDAR to a connectome-wired navigation network
Markdown · JSON · Original source
Ali Mahzarnia and Homayoun Jamshidi describe FLYNN-LiDAR, an extension of FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome. It adds a 32-ray 2D LiDAR interface without a learned fusion layer beyond a scalar input gain. In the authors’ comparisons, the model had 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines. The abstract reports gains under total vision loss, textures, and periodic vision dropout; with full vision, success was 96.6%. The reported results suggest that this particular sensor connection can help the network navigate across several tested vision conditions, while the losses and test-specific results limit what can be concluded from the comparisons. The authors also report fewer multiply accumulate operations per step than EfficientNet and MobileNet in the evaluated forward paths. The abstract does not establish how the approach performs outside the reported tests.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
q-bio.NC · Article
Source details
Original source date: Nov 2, 2026 (2026-11-02).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-10T04:53:23.481Z. First observed: 2026-11-04T05:12:29.108Z. Retrieved: 2026-11-04T05:46:25.177Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The supplied page gives the abstract but not enough detail to assess the navigation environments, training setup, or statistical methods behind the reported comparisons.
Open question
Would FLYNN-LiDAR’s reported gains persist in navigation tests beyond the vision, texture, and dropout conditions described in the abstract?
Source-stated
The authors extend FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome, with a 32-ray 2D LiDAR interface that needs no learned fusion layer beyond scalar input gain.
We extend FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome, with a 32-ray 2D LiDAR interface that requires no learned fusion layer beyond a scalar input gain.
Source-stated
Across 24 paired comparisons, FLYNN-LiDAR had 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines.
Across 24 paired comparisons in a four- state vision matrix with LiDAR on or off, FLYNN-LiDAR has 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines.
Source-stated
The abstract reports higher success with the scan under total vision loss and out of distribution textures, as well as high success under periodic vision dropout.
Under total vision loss, the scan raises success from 45.2% to 64.6%
Source-stated
The authors report that the evaluated FLYNN-LiDAR forward path uses fewer multiply accumulate operations per step than EfficientNet and MobileNet.
Profiling the evaluated forward paths gives 3.93M multiply accumulate operations per step; EfficientNet and MobileNet require 4.60x and 2.94x as many.
Escape Meridian interpretation
The abstract reports results for specified test conditions; those results do not by themselves establish performance beyond those conditions.
Under out of distribution textures, it is the only LiDAR-capable model whose success improves with the scan (44.2% to 58.4%)
Source and contributor context
Mahzarnia, Ali, Jamshidi, Homayoun · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Distribution shift
A change between the data a model was trained on and the data it meets in use, which can lower its accuracy.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Editorial provenance
Version 1. Publication ID: b1415631-e01c-429a-8e25-4d474c02ff1d.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T04:53:23.481Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 763380fd9a4b3ae499454b6d78a5b5434c6121b64b927dbbd99becbcf837c53f. Source text hash: 3eb549ae7fe3472bd9054fcc881dcc84ca57bd6e74e9cfa42505b33f7375315d.
Revision and source-check history
Version 1 · 2026-11-10T04:53:23.481Z
Mahzarnia and Jamshidi add LiDAR to a connectome-wired navigation network
Ali Mahzarnia and Homayoun Jamshidi describe FLYNN-LiDAR, an extension of FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome. It adds a 32-ray 2D LiDAR interface without a learned fusion layer beyond a scalar input gain. In the authors’ comparisons, the model had 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines. The abstract reports gains under total vision loss, textures, and periodic vision dropout; with full vision, success was 96.6%. The reported results suggest that this particular sensor connection can help the network navigate across several tested vision conditions, while the losses and test-specific results limit what can be concluded from the comparisons. The authors also report fewer multiply accumulate operations per step than EfficientNet and MobileNet in the evaluated forward paths. The abstract does not establish how the approach performs outside the reported tests.
Original source
q-bio.NC · Article
Source details
Original source date: Nov 2, 2026 (2026-11-02).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-10T04:53:23.481Z. First observed: 2026-11-04T05:12:29.108Z. Retrieved: 2026-11-04T05:46:25.177Z.
Limits
The supplied page gives the abstract but not enough detail to assess the navigation environments, training setup, or statistical methods behind the reported comparisons.
Open question
Would FLYNN-LiDAR’s reported gains persist in navigation tests beyond the vision, texture, and dropout conditions described in the abstract?
Source-stated
The authors extend FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome, with a 32-ray 2D LiDAR interface that needs no learned fusion layer beyond scalar input gain.
We extend FLYNN, a leaky integrator recurrent network wired on the Drosophila melanogaster connectome, with a 32-ray 2D LiDAR interface that requires no learned fusion layer beyond a scalar input gain.
Source-stated
Across 24 paired comparisons, FLYNN-LiDAR had 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines.
Across 24 paired comparisons in a four- state vision matrix with LiDAR on or off, FLYNN-LiDAR has 17 significant wins, 5 nonsignificant comparisons, and 2 significant losses against three LiDAR-capable baselines.
Source-stated
The abstract reports higher success with the scan under total vision loss and out of distribution textures, as well as high success under periodic vision dropout.
Under total vision loss, the scan raises success from 45.2% to 64.6%
Source-stated
The authors report that the evaluated FLYNN-LiDAR forward path uses fewer multiply accumulate operations per step than EfficientNet and MobileNet.
Profiling the evaluated forward paths gives 3.93M multiply accumulate operations per step; EfficientNet and MobileNet require 4.60x and 2.94x as many.
Escape Meridian interpretation
The abstract reports results for specified test conditions; those results do not by themselves establish performance beyond those conditions.
Under out of distribution textures, it is the only LiDAR-capable model whose success improves with the scan (44.2% to 58.4%)
Source and contributor context
Mahzarnia, Ali, Jamshidi, Homayoun · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Distribution shift
A change between the data a model was trained on and the data it meets in use, which can lower its accuracy.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Editorial provenance
Version 1. Publication ID: b1415631-e01c-429a-8e25-4d474c02ff1d.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T04:53:23.481Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: 763380fd9a4b3ae499454b6d78a5b5434c6121b64b927dbbd99becbcf837c53f. Source text hash: 3eb549ae7fe3472bd9054fcc881dcc84ca57bd6e74e9cfa42505b33f7375315d.
uncertain · 2026-11-10T04:53:23.481Z · anthropic/claude-opus-5.5
Publication: b1415631-e01c-429a-8e25-4d474c02ff1d
attribution — pass: The reading correctly credits Ali Mahzarnia and Homayoun Jamshidi as the authors. It frames the results as the authors' reported comparisons, not as established facts.
Authors:Ali Mahzarnia, Homayoun Jamshidi
summary — uncertain: Central terms are left unexplained for a general reader: connectome, Drosophila melanogaster (fruit fly), leaky integrator recurrent network, LiDAR, fusion layer, scalar input gain and multiply accumulate operations. The reading also drops the paper's core idea: the laser scan is fed into the fly's head-bristle neurons ('whiskers'). Smallest repair: one plain sentence on that mechanism, plus tooltips for these terms. Adding the baseline context (CNNs at 0% on new textures, both losses to MobileNet) would also help.
A 270 degree frontal scan at 8 m range is injected as inverted proximity into the 37 BM Vib head-bristle neurons.
date — pass: The originalPublished date of 2026-11-02 matches the arXiv v1 submission date.
[Submitted on 2 Nov 2026]
support — uncertain: The summary says the abstract 'reports gains under ... periodic vision dropout.' For dropout, the abstract gives only absolute success rates (93.6% and 89.6%) and fewer collisions. It gives no scan-off or baseline figures, so a gain there is not directly shown. Smallest repair: 'reports high success under periodic vision dropout,' as in claim 3. The other claims and the conditional interpretation are grounded in the abstract.
Under periodic vision dropout, FLYNN-LiDAR achieves 93.6% and 89.6% success at 25% and 50% duty with fewer collisions.
media — pass: The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
A change between the data a model was trained on and the data it meets in use, which can lower its accuracy.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivarXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivA change between the data a model was trained on and the data it meets in use, which can lower its accuracy.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivarXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXiv