Pith. sign in

REVIEW 4 major objections 5 minor 85 references

What if Eye...? Computationally Recreating Vision Evolution

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Task-specific selection alone can recreate the evolution of compound eyes, camera eyes, and lenses in simulation.

desk verdict A genuinely new evolutionary vision framework whose headline result is undercut by a mutation-operator confound; worth serious refereeing, but needs a control experiment. read the letter →

arxiv 2501.15001 v2 pith:UXH2C7QO submitted 2025-01-25 cs.AI cs.CVcs.NEq-bio.NC

Kushagra Tiwary , Aaron Young , Zaid Tasneem , Tzofi Klinghoffer , Akshat Dave , Tomaso Poggio , Dan-Eric Nilsson , Brian Cheung
show 1 more author
Ramesh Raskar
This is my paper · ORCID
classification cs.AIcs.CVcs.NEq-bio.NC
keywords computationalevolutioneyecompoundeyescameralensemergenceembodiedagentsreinforcementlearningvisualacuityscaling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Vision in nature spans simple light-sensitive patches to lensed camera eyes, but the role of specific environmental tasks in driving that diversity is hard to test because evolution cannot be rerun. This paper builds a computational evolution framework in which embodied agents co-evolve eye morphology, optics, and neural processing while learning a single visual task, and claims that the task alone determines the outcome. Agents evolved for maze navigation converge on distributed compound-type eyes, while agents evolved for object discrimination converge on two forward-facing, high-acuity camera-type eyes. When optical elements are allowed to mutate, lens-like focusing structures emerge to resolve the trade-off between light collection and spatial precision. The paper also reports power-law scaling between visual acuity and neural capacity, concluding that poor acuity is a bottleneck that more neurons cannot overcome.

What carries the argument

The load-bearing mechanism is a unified genetic encoding with three independently mutating gene clusters: morphological genes (number, placement, and field of view of eyes), optical genes (pupil size, a programmable phase mask acting as a diffractive optical element, and refractive index), and neural genes (hidden-layer size and temporal memory). The outer evolutionary loop mutates and selects these genes over generations, while the inner loop trains each agent's policy with a reinforcement-learning algorithm during its lifetime, mirroring the Baldwin effect. Image formation is modeled by convolving the scene with a physically based point-spread function, so closing the aperture sharpens the image but reduces light throughput in proportion to the square of the pupil radius and increases noise. This light-versus-precision trade-off is the pressure that lens evolution is claimed to resolve, and acuity is quantified in cycles per degree (CPD) so that sensory quality can be compared with network parameter count.

What would settle it

Run the same evolution with the Navigation reward but make the maze exit identifiable only by a fine striped pattern, or with the Detection reward but make food and poison differ only in size rather than high-frequency texture. If the eye morphologies do not switch accordingly, the bifurcation is an artifact of task geometry rather than a general property of task-driven eye evolution. A second decisive check is to allow optical mutations in the Navigation task: if high-acuity lensed eyes also emerge there, the claim that lenses arise specifically from object-discrimination demand would be weakened.

Watch

Extended reading notes

Core claim

At its center, the paper claims that task-specific selection is sufficient to reproduce the major branch points of eye evolution. Agents are initialized as a single photoreceptor and then selected by an outer evolutionary loop while each genotype is trained by an inner reinforcement-learning loop. Only the reward function differs between experiments, and the eyes that emerge diverge accordingly: orientation and obstacle avoidance produce many distributed, low-resolution eyes with broad field of view, while discrimination between visually similar objects produces two frontal, high-resolution eyes. In a separate experiment, evolution first narrows the pupil toward pinhole-like eyes and then, once optical mutations are allowed, discovers lens-like phase profiles that keep images sharp while admitting more light, escaping the pinhole's performance ceiling. Finally, task error is reported to obey power laws in both neural parameters and acuity, with acuity setting a floor: increasing network size alone cannot compensate for poor vision. The authors present these results as a method for turning 'what-if' questions about natural vision into counterfactual computational experiments.

Load-bearing premise

The central claim depends on the assumption that the three hand-designed reward functions and their environments capture the natural selective pressures that shaped real vision; if different reward formulations or scene layouts produced different eye morphologies, the results would describe the simulator rather than vision evolution.

Editorial extensions

If this is right

  • Comparative hypotheses about a species' visual ecology can be tested by recreating its presumed task in simulation and checking whether the evolved morphology matches.
  • High-acuity vision under dim conditions requires an optical innovation, not simply more photoreceptors, because pinhole-like apertures trade away light.
  • Embodied agents cannot escape sensory limits by adding parameters alone; acuity and network capacity must improve together to reduce task error.
  • Starting from a single primitive photoreceptor, the same genome can diverge into compound and camera designs depending only on behavioral demand, which offers a computational explanation for convergent eye types in nature.
  • Because the genotype spans roughly 10^20 configurations, the same machinery can serve as a generative design tool for manufacturable bio-inspired imaging systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension is to evolve agents on multiple tasks at once; the paper itself notes that natural vision solves many tasks simultaneously, so multi-task rewards might produce intermediate or mosaic eyes rather than the clean bifurcation seen under isolated tasks.
  • The acuity-bottleneck result suggests a concrete design rule for artificial vision systems: co-design sensor resolution and network capacity, because improving either alone will hit a ceiling.
  • The lens-emergence result could be tested further by varying ambient light intensity rather than fixing noise, which would directly probe whether lighting conditions alone can trigger lens evolution.
  • The methodology implies a general counterfactual engine for vision: any environmental change, such as a world without movement or a monochromatic world, can be instantiated as a reward and environment and then evolved.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper presents a computational framework that co-evolves the physical eye (photoreceptor layout, pupil, phase mask, refractive index), eye placement, and neural controllers of embodied agents trained by reinforcement learning, and uses it to revisit hypotheses about eye evolution. Three claims are made: (i) changing only the visual task produces a bifurcation, with navigation yielding distributed compound-type eyes and object discrimination yielding high-acuity camera-type eyes; (ii) when optical elements are allowed to mutate, lens-like structures evolve to resolve the trade-off between light collection and spatial precision; and (iii) task error follows power-law scaling with neural capacity that is bounded by visual acuity, so that poor acuity cannot be compensated by larger brains alone. The results are presented as computational counterfactual evidence for the evolutionary role of task-specific selective pressures.

Significance. If the central claims hold, the framework would be a genuinely useful instrument for testing what-if questions in vision evolution, because it couples physically motivated image formation with embodied reinforcement learning and evolutionary search. The paper has real strengths: the genotype spans morphology, optics, and neural capacity in a unified encoding; the imaging model in Eqs. (4)-(8) is concrete and physically grounded; the CPD and MTF analyses give quantitative, falsifiable characterizations of the evolved eyes; and the reported outcomes—compound versus camera morphologies, lens-like phase masks, and acuity-bounded scaling—are the kind of concrete predictions that computational evolution should provide. The significance is tempered, however, by the fact that the central bifurcation experiment changes more than the task, the lens result is shown on a single trajectory, and the scaling-law claim rests on finite sweeps whose uncertainty is not reported.

major comments (4)
  1. [Section 2.4, Figure 3, Section 4 (Mutation operators)] The headline claim that the camera/compound bifurcation is caused 'by strictly changing the visual task' (Section 1 and Section 2.4) is not supported as stated, because the two arms differ in the allowed mutation operator. The text says 'As the Navigation task is essentially two-dimensional, only photoreceptors along the horizontal width of the eye are added/removed,' while Detection agents add photoreceptors in two dimensions and converge to 15x15 camera eyes. Under the Navigation mutation operator, a 15x15 camera-type eye is unreachable by construction, so the absence of camera-type eyes in Navigation could be an artifact of the search space rather than of task-specific selection. The confound is testable and internal: run Navigation with the same two-dimensional photoreceptor mutation operator and report whether the population still converges to distributed 1x4 compound-type eyes. Without this control, the paper's central 'what-if' conclusion is not established.
  2. [Section 2.5, Figure 4, Figure A2] The lens-emergence result is reported on a single evolutionary trajectory: Phase I (pupil-only, 30 generations) is followed by Phase II (optics enabled) and the narrative describes a specific progression through cup, pinhole, unfocused, and focused lens eyes. CMA-ES with RL fitness evaluation is stochastic, and Section 4 reports no independent replicates for this experiment. To support the claim that evolution reliably discovers lens-like optics to resolve the light-throughput/spatial-precision trade-off, the authors should report multiple independent evolutionary runs with different seeds, showing the distribution of final PSFs, fitness gains, and the fraction of runs that produce well-focused lens-like phase masks. A single trajectory also cannot justify the statement that pinhole design 'plateaus' as a general evolutionary outcome.
  3. [Section 2.6, Figure 5, Appendix C] The scaling-law claims are load-bearing for the paper's third contribution, but the supporting evidence is under-specified. The paper reports fitted power laws such as L = (9.50e-3) N^0.69 for navigation as if they are deterministic, with no confidence intervals, number of seeds, or per-acuity-level fit quality. The claim that 'poor visual acuity creates a fundamental bottleneck that cannot be overcome by simply scaling neural capacity' is an extrapolation from finite sweeps of CPD and parameter counts; the data in Figure A5 show scattered errors, and the ceiling behavior needs to be quantified (e.g., by fitting saturating curves and reporting where the plateau begins) rather than asserted from the plotted envelopes. Relatedly, since CPD and input resolution jointly determine the network's input dimension, the independence of the 'acuity' and 'parameter' axes should be made explicit.
  4. [Section 2.5, Appendix A (Image Quality metric)] The 'Image Quality' metric in Section 2.5 and Appendix A, defined as the product of MTF area above a noise floor and light throughput (which decreases quadratically with pupil radius), embeds exactly the trade-off that the paper claims evolution discovers. If this metric is used as the primary evidence that lens eyes are better, the argument is in part circular: the metric prefers lenses by construction. The independent evidence is behavioral—higher fitness in Phase II (Figure A2)—and the paper should make clear that the Image Quality metric is a post-hoc diagnostic rather than the objective being optimized, and should present the fitness comparison as the main test. If the authors intend the Image Quality metric as a causal driver, they need to show that evolution would behave differently under an alternative analysis metric.
minor comments (5)
  1. [Appendix B] There are several typos in Appendix B: 'orientaotin' should be 'orientation', 'differnet' should be 'different', and 'genptype' should be 'genotype'. The same appendix also spells 'stomatopods' inconsistently and should be proofread.
  2. [Figure A3 caption] In the Figure A3 caption, 'photreceptor' should be 'photoreceptor', and the sentence explaining the single-receptor case is syntax-heavy and should be rewritten for clarity.
  3. [Section 2.4] The phrase 'with with larger input arrays' appears in the paragraph following Figure 3; the duplicated 'with' should be removed.
  4. [Section 4, Eqs. (4)-(7)] The definition of U_in(x,y) is given twice, once as Eq. (6) and once as an unnumbered equation before Eq. (8), and the sentence 'pupil function pupil function' in the paragraph after Eq. (4) is duplicated. These should be consolidated and edited.
  5. [Section 4] The claim of 'approximately 10^20 unique agent vision types' is stated twice but no derivation or counting argument is given; a brief enumeration in the methods would make the claim checkable.

Circularity Check

1 steps flagged · score 4.0 of 10

The lens-emergence and scaling-law results are self-contained, but the central navigation/detection bifurcation is partly forced by a built-in mutation-operator asymmetry.

  1. other [Section 1 (contribution 1) and Section 2.4; Methods 'Mutation operators']
    "By strictly changing the visual task an agent is subject to, orientation ( Navigation) task vs. object discrimination (Detection) task [4, 31], we observe a bifurcation in our evolved agents between camera-type and compound-type eyes ... As the Navigation task is essentially two-dimensional, only photoreceptors along the horizontal width of the eye are added/removed."

    The bifurcation is presented as caused by the visual task alone, but the Navigation arm is defined with a different mutation operator: only 1D horizontal photoreceptor arrays can be added or removed, so a 2D 15x15 camera-type eye is unreachable in that arm by construction. The observed absence of camera-type eyes in Navigation is therefore guaranteed by the search-space definition, not discovered as a consequence of task-specific selection. The comparison changes task, reward, environment layout, and mutation operator simultaneously, so the central contribution (1) reduces in part to this built-in constraint rather than to a task-only evolutionary outcome.

full rationale

The lens-emergence experiment (Section 2.5) is not circular: lens shapes are evolved in a free 4x4 phase mask, and the Image Quality metric is descriptive rather than the fitness signal, since fitness is the task reward. The scaling-law results (Section 2.6) are likewise empirically fitted sweeps over CPD and network size, with no fitted parameter renamed as a prediction. The main circularity-adjacent problem is the Section 2.4 bifurcation experiment, where the Navigation arm's mutation operator excludes 2D receptor arrays, partially forcing the absence of camera eyes; this undermines the 'strictly changing the visual task' claim. No load-bearing self-citation chain appears: self-citations such as [24], [69], and [70] are contextual or engineering-related, not the basis of the evolutionary conclusions.

Assumptions & free parameters 4 free parameters · 7 assumptions · 0 invented entities

The ledger tracks numbers and assumptions the results depend on. The paper does not introduce new physical entities; the genotype and phase mask are computational constructs.

free parameters (4)
  • Scaling law coefficient a and exponent b = Navigation: a=9.50e-3, b=0.69; Detection: a=7.31e-3, b=0.74; Tracking: a=2.43e-2, b=0.59
    In Section 2.6, power laws L = a*N^b are fit to simulation data; these constants are not derived and are used to support the scaling-law claim.
  • Reward and fitness weights = Training: lambda=0.25, wg=1, wa=-1, wc=-1; Fitness: lambda=1.5, wg=10, wa=-10, wc=-2
    Hand-chosen in Section 4; they set the selective pressure that defines which agents are fit, so the evolved morphologies depend on them.
  • Search-space bounds for genetic encoding = FOV 1 to 100 deg, pupil radius 0 to 1, refractive index 1.0 to 2.0, phase mask 4x4, hidden layers 1 to 512 neurons
    Chosen by hand to make optimization tractable; these bounds constrain which eye designs can be discovered.
  • Evolution run hyperparameters = Population size 16; 50 to 100 generations; training up to 1e6 steps; early stopping after five evaluations without…
    Algorithmic choices in Section 4; no multiple-seed evolutionary runs are reported.
assumptions (7)
  • standard math Fourier optics propagation model (Equations 4 to 6) describes image formation in the simulated eye.
    Standard scalar diffraction theory from Goodman is used to compute PSFs; this is unproved background math that the lens-evolution results rest on.
  • domain assumption Depth-independent PSF and a single convolution per eye are sufficient for the visual tasks.
    Stated in Methods: 'we assume depth independent PSFs to limit the number of depth-convolutions, to one convolution, per eye'; this approximation may affect lens evolution outcomes.
  • domain assumption Bilateral symmetry of eyes is assumed.
    Methods state 'We assume bilateral symmetry, consistent with the observation that the overwhelming majority of animals have bilateral symmetry.'
  • domain assumption Reward functions and task environments isolate the selective pressures that shaped natural vision.
    The tasks are designed from biological studies such as honeybee navigation and visual roles; the biological conclusions depend on this equivalence.
  • domain assumption Lifetime RL training approximates the Baldwin effect.
    Section 2.3 states that the inner learning loop mirrors the Baldwin effect; no direct biological validation is provided.
  • domain assumption CMA-ES with Gaussian mutations is a sufficient model of natural selection for these questions.
    Methods use CMA-ES as a proxy for evolution; natural mutation and selection are more complex and may explore different regions of the design space.
  • ad hoc to paper A 4x4 programmable phase mask upsampled to 51x51 represents the space of possible lens shapes.
    This tractability choice in Methods restricts the lens geometries evolution can discover and may affect the claim that lens-like structures emerge.

how reviews work

0 comments
Cite this review

Pith. "Pith review of What if Eye...? Computationally Recreating Vision Evolution." pith.science (2026). https://pith.science/paper/UXH2C7QO

@misc{pith2026250115001,
  author       = {Pith},
  title        = {Pith review of: What if Eye...? Computationally Recreating Vision Evolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UXH2C7QO}},
  note         = {Machine review of arXiv:2501.15001}
}
read the original abstract

Vision systems in nature show remarkable diversity, from simple light-sensitive patches to complex camera eyes with lenses. While natural selection has produced these eyes through countless mutations over millions of years, they represent just one set of realized evolutionary paths. Testing hypotheses about how environmental pressures shaped eye evolution remains challenging since we cannot experimentally isolate individual factors. Computational evolution offers a way to systematically explore alternative trajectories. Here we show how environmental demands drive three fundamental aspects of visual evolution through an artificial evolution framework that co-evolves both physical eye structure and neural processing in embodied agents. First, we demonstrate computational evidence that task specific selection drives bifurcation in eye evolution - orientation tasks like navigation in a maze leads to distributed compound-type eyes while an object discrimination task leads to the emergence of high-acuity camera-type eyes. Second, we reveal how optical innovations like lenses naturally emerge to resolve fundamental tradeoffs between light collection and spatial precision. Third, we uncover systematic scaling laws between visual acuity and neural processing, showing how task complexity drives coordinated evolution of sensory and computational capabilities. Our work introduces a novel paradigm that illuminates evolutionary principles shaping vision by creating targeted single-player games where embodied agents must simultaneously evolve visual systems and learn complex behaviors. Through our unified genetic encoding framework, these embodied agents serve as next-generation hypothesis testing machines while providing a foundation for designing manufacturable bio-inspired vision systems. Website: http://eyes.mit.edu/

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

85 extracted references · 68 canonical work pages

  1. [1]

    Oxford University Press, United Kingdom (2002)

    Land, M., Nilsson, D.-E.: Animal Eyes. Oxford University Press, United Kingdom (2002)

    2002

  2. [2]

    Annual Review of Vision Science 7(Volume 7, 2021), 19–41 (2021) https://doi.org/10.1146/annurev-vision-121820-074736

    Nilsson, D.-E.: The diversity of eyes and vision. Annual Review of Vision Science 7(Volume 7, 2021), 19–41 (2021) https://doi.org/10.1146/annurev-vision-121820-074736

    work page doi:10.1146/annurev-vision-121820-074736 2021

  3. [3]

    Science313(5795), 1914–1918 (2006)

    Fernald, R.D.: Casting a genetic light on the evolution of eyes. Science313(5795), 1914–1918 (2006)

    2006

  4. [4]

    Philosophical Transac- tions of the Royal Society B: Biological Sciences 364(1531), 2833–2847 (2009)

    Nilsson, D.-E.: The evolution of eyes and visually guided behaviour. Philosophical Transac- tions of the Royal Society B: Biological Sciences 364(1531), 2833–2847 (2009)

    2009

  5. [5]

    Proceedings of the Royal Society of London

    Nilsson, D.-E., Pelger, S.: A pessimistic estimate of the time required for an eye to evolve. Proceedings of the Royal Society of London. Series B: Biological Sciences 256(1345), 53–58 (1994)

    1994

  6. [6]

    Scientific American 185(2), 60–64 (1951)

    Walter, W.G.: A machine that learns. Scientific American 185(2), 60–64 (1951)

    1951

  7. [7]

    Robotica 21(4), 351–363 (2003)

    Holland, O.: The first biologically inspired robots. Robotica 21(4), 351–363 (2003)

    work page 2003

  8. [8]

    Springer, Cham (2016)

    Nolfi, S., Bongard, J., Husbands, P., Floreano, D.: Evolutionary Robotics. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-32552-1 76 . https://doi.org/10.1007/ 978-3-319-32552-1 76

    doi:10.1007/978-3-319-32552-1 2016

Show all 85 references
  1. [9]

    Frontiers in Robotics and AI 11, 1278983 (2024)

    Winfield, A.F.: Evolutionary robotics as a modelling tool in evolutionary biology. Frontiers in Robotics and AI 11, 1278983 (2024)

    2024

  2. [10]

    Trends in ecology & evolution 26(7), 369–375 (2011)

    Krause, J., Winfield, A.F., Deneubourg, J.-L.: Interactive robots in experimental biology. Trends in ecology & evolution 26(7), 369–375 (2011)

    2011

  3. [11]

    Nature 406(6799), 974–978 (2000)

    Lipson, H., Pollack, J.B.: Automatic design and manufacture of robotic lifeforms. Nature 406(6799), 974–978 (2000)

    2000

  4. [12]

    The American Naturalist 30(354), 441–451 (1896)

    Baldwin, J.M.: A new factor in evolution. The American Naturalist 30(354), 441–451 (1896)

  5. [13]

    Princeton University Press, Princeton, NJ (2014)

    Cronin, T.W., Johnsen, S., Marshall, N.J., Warrant, E.J.: Visual Ecology. Princeton University Press, Princeton, NJ (2014)

    2014

  6. [14]

    Basic books, ??? (2018)

    Pearl, J., Mackenzie, D.: The Book of Why: the New Science of Cause and Effect. Basic books, ??? (2018)

    2018

  7. [15]

    Bongard, J.C.: Evolutionary robotics. Commun. ACM 56(8), 74–83 (2013) https://doi.org/ 10.1145/2493883

    2013 doi

  8. [16]

    Trianni, V.: Evolutionary robotics: model or design? Frontiers in Robotics and AI 1, 13 (2014)

    2014

  9. [17]

    Nature, 1–9 (2024)

    Lappalainen, J.K., Tschopp, F.D., Prakhya, S., McGill, M., Nern, A., Shinomiya, K., Take- mura, S.-y., Gruntman, E., Macke, J.H., Turaga, S.C.: Connectome-constrained networks predict neural activity across the fly visual system. Nature, 1–9 (2024)

    2024

  10. [18]

    In: 4th European Conference on Artificial Life, pp

    Floreano, D., Nolfi, S.: Adaptive behavior in competing co-evolving species. In: 4th European Conference on Artificial Life, pp. 378–387 (1997) 25

    1997

  11. [19]

    PloS one 15(5), 0233848 (2020)

    Miras, K., Ferrante, E., Eiben, A.E.: Environmental influences on evolvable robots. PloS one 15(5), 0233848 (2020)

    2020

  12. [20]

    PLoS computational biology 11(8), 1004273 (2015)

    Ferrante, E., Turgut, A.E., Du´ e˜ nez-Guzm´ an, E., Dorigo, M., Wenseleers, T.: Evolution of self-organized task specialization in robot swarms. PLoS computational biology 11(8), 1004273 (2015)

    2015

  13. [21]

    PLoS biology 9(5), 1000615 (2011)

    Waibel, M., Floreano, D., Keller, L.: A quantitative test of hamilton’s rule for the evolution of altruism. PLoS biology 9(5), 1000615 (2011)

    2011

  14. [22]

    IEEE transactions on Evolutionary Computation 13(3), 648–660 (2009)

    Waibel, M., Keller, L., Floreano, D.: Genetic team composition and level of selection in the evolution of cooperation. IEEE transactions on Evolutionary Computation 13(3), 648–660 (2009)

    2009

  15. [23]

    Trends in Neurosciences 46(3), 240–254 (2023)

    Kanwisher, N., Khosla, M., Dobs, K.: Using artificial neural networks to ask ‘why’questions of minds and brains. Trends in Neurosciences 46(3), 240–254 (2023)

    2023

  16. [24]

    CoRR abs/1611.09430 (2016) 1611.09430

    Cheung, B., Weiss, E., Olshausen, B.A.: Emergence of foveal image sampling from learning to attend in visual scenes. CoRR abs/1611.09430 (2016) 1611.09430

    2016 arXiv

  17. [25]

    Vision research 200, 108083 (2022)

    Pramod, R., Katti, H., Arun, S.: Human peripheral blur is optimal for object recognition. Vision research 200, 108083 (2022)

    2022

  18. [26]

    https://arxiv.org/abs/1712.01815

    Silver, D., Hubert, T., Schrittwieser, J., Antonoglou, I., Lai, M., Guez, A., Lanctot, M., Sifre, L., Kumaran, D., Graepel, T., Lillicrap, T., Simonyan, K., Hassabis, D.: Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm (2017). https://arx...

    2017 arXiv

  19. [27]

    : Discovering faster matrix multiplication algorithms with reinforcement learning

    Fawzi, A., Balog, M., Huang, A., Hubert, T., Romera-Paredes, B., Barekatain, M., Novikov, A., R Ruiz, F.J., Schrittwieser, J., Swirszcz, G., et al. : Discovering faster matrix multiplication algorithms with reinforcement learning. Nature 610(7930), 47–53 (2022)

    2022

  20. [28]

    : A general reinforcement learning algorithm that masters chess, shogi, and go through self-play

    Silver, D., Hubert, T., Schrittwieser, J., Antonoglou, I., Lai, M., Guez, A., Lanctot, M., Sifre, L., Kumaran, D., Graepel, T., et al. : A general reinforcement learning algorithm that masters chess, shogi, and go through self-play. Science 362(6419), 1140–1144 (2018)

    2018

  21. [29]

    : Faster sorting algorithms discovered using deep reinforcement learning

    Mankowitz, D.J., Michi, A., Zhernov, A., Gelmi, M., Selvi, M., Paduraru, C., Leurent, E., Iqbal, S., Lespiau, J.-B., Ahern, A., et al. : Faster sorting algorithms discovered using deep reinforcement learning. Nature 618(7964), 257–263 (2023)

    2023

  22. [30]

    Cambridge University Press, ??? (2006)

    Warrant, E., Nilsson, D.-E.: Invertebrate Vision. Cambridge University Press, ??? (2006)

    2006

  23. [31]

    Frontiers in Neuroanatomy 16 (2022) https://doi.org/10.3389/fnana.2022.789375

    Nilsson, D.-E.: The evolution of visual roles – ancient vision versus object vision. Frontiers in Neuroanatomy 16 (2022) https://doi.org/10.3389/fnana.2022.789375

    2022

  24. [32]

    arXiv preprint arXiv:2001.08361 (2020)

    Kaplan, J., McCandlish, S., Henighan, T., Brown, T.B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., Amodei, D.: Scaling laws for neural language models. arXiv preprint arXiv:2001.08361 (2020)

    2020 arXiv

  25. [33]

    Trends in ecology & evolution 33(5), 358–372 (2018)

    Caves, E.M., Brandley, N.C., Johnsen, S.: Visual acuity and the evolution of signals. Trends in ecology & evolution 33(5), 358–372 (2018)

    2018

  26. [34]

    Journal of Experimental Biology 199(1), 237–244 26 (1996)

    Srinivasan, M.V., Zhang, S., Lehrer, M., Collett, T.: Honeybee navigation en route to the goal: visual flight control and odometry. Journal of Experimental Biology 199(1), 237–244 26 (1996)

    1996

  27. [35]

    In: IROS, pp

    Todorov, E., Erez, T., Tassa, Y.: Mujoco: A physics engine for model-based control. In: IROS, pp. 5026–5033. IEEE, ??? (2012)

    2012

  28. [36]

    WIREs Developmen- tal Biology 3(1), 1–40 (2014) https://doi.org/10.1002/wdev.96 https://wires.onlinelibrary.wiley.com/doi/pdf/10.1002/wdev.96

    Gehring, W.J.: The evolution of vision. WIREs Developmen- tal Biology 3(1), 1–40 (2014) https://doi.org/10.1002/wdev.96 https://wires.onlinelibrary.wiley.com/doi/pdf/10.1002/wdev.96

    2014 doi

  29. [37]

    Trends in genetics 15(9), 371–377 (1999)

    Gehring, W.J., Ikeo, K.: Pax 6: mastering eye morphogenesis and eye evolution. Trends in genetics 15(9), 371–377 (1999)

    1999

  30. [38]

    Scientific reports 9(1), 9360 (2019)

    Aguirre, G.K.: A model of the entrance pupil of the human eye. Scientific reports 9(1), 9360 (2019)

    2019

  31. [39]

    Evolution 68(5), 1365– 1373 (2014)

    Doebeli, M., Ispolatov, I.: Chaos and unpredictability in evolution. Evolution 68(5), 1365– 1373 (2014)

    2014

  32. [40]

    In: Lawrence, G.N

    O’Shea, D.C.: Monochromatic quartet: a search for the global optimum. In: Lawrence, G.N. (ed.) 1990 Intl Lens Design Conf, vol. 1354, pp. 548–554. SPIE, ??? (1991). https://doi.org/10.1117/12.47896 . International Society for Optics and Photonics. https://doi.org/10.1117/12.47896

    1991 doi

  33. [41]

    Applied Soft Computing 8(4), 1439–1452 (2008) https://doi.org/ 10.1016/j.asoc.2007.10.018

    Gagn´ e, C., Beaulieu, J., Parizeau, M., Thibault, S.: Human-competitive lens system design with evolution strategies. Applied Soft Computing 8(4), 1439–1452 (2008) https://doi.org/ 10.1016/j.asoc.2007.10.018 . Soft Computing for Dynamic Data Mining

    2008 doi

  34. [42]

    Nature communications 12(1), 5721 (2021)

    Gupta, A., Savarese, S., Ganguli, S., Fei-Fei, L.: Embodied intelligence via learning and evolution. Nature communications 12(1), 5721 (2021)

    2021

  35. [43]

    In: Proceedings of the 18th Annual Conference on Computer Graphics and Interactive Techniques

    Sims, K.: Artificial evolution for computer graphics. In: Proceedings of the 18th Annual Conference on Computer Graphics and Interactive Techniques. SIGGRAPH ’91, pp. 319–

  36. [44]

    Biological cybernetics 90, 218–228 (2004)

    Floreano, D., Kato, T., Marocco, D., Sauser, E.: Coevolution of active vision and feature selection. Biological cybernetics 90, 218–228 (2004)

    2004

  37. [45]

    arXiv preprint arXiv:2407.17032 (2024)

    Towers, M., Kwiatkowski, A., Terry, J., Balis, J.U., De Cola, G., Deleu, T., Goul˜ ao, M., Kallinteris, A., Krimmel, M., KG, A., et al.: Gymnasium: A standard interface for reinforcement learning environments. arXiv preprint arXiv:2407.17032 (2024)

    2024 arXiv

  38. [46]

    PLOS Biology 22(1), 3002422 (2024) https://doi.org/10.1371/journal.pbio.3002422

    Baden, T.: From water to land: Evolution of photoreceptor circuits for vision in air. PLOS Biology 22(1), 3002422 (2024) https://doi.org/10.1371/journal.pbio.3002422

    2024 doi

  39. [47]

    Evolutionary Computation 10(2), 99–127 (2002) https://doi.org/10.1162/ 106365602320169811

    Stanley, K.O., Miikkulainen, R.: Evolving neural networks through augmenting topologies. Evolutionary Computation 10(2), 99–127 (2002) https://doi.org/10.1162/ 106365602320169811

    2002

  40. [48]

    arXiv preprint arXiv:1703.03864 (2017)

    Salimans, T., Ho, J., Chen, X., Sidor, S., Sutskever, I.: Evolution strategies as a scalable alternative to reinforcement learning. arXiv preprint arXiv:1703.03864 (2017)

    2017 arXiv

  41. [49]

    https://arxiv.org/abs/1712.06560

    Conti, E., Madhavan, V., Such, F.P., Lehman, J., Stanley, K.O., Clune, J.: Improving Exploration in Evolution Strategies for Deep Reinforcement Learning via a Population of 27 Novelty-Seeking Agents (2018). https://arxiv.org/abs/1712.06560

    2018 arXiv

  42. [50]

    Proceedings of the IEEE International Conference on Evolutionary Computation, 312–317 (1996)

    Hansen, N., Ostermeier, A.: Adapting arbitrary normal utation. Proceedings of the IEEE International Conference on Evolutionary Computation, 312–317 (1996)

    1996

  43. [51]

    GitHub (2018)

    Rapin, J., Teytaud, O.: Nevergrad - A gradient-free optimization platform. GitHub (2018)

    2018

  44. [52]

    CoRR abs/1707.06347 (2017) 1707.06347

    Schulman, J., Wolski, F., Dhariwal, P., Radford, A., Klimov, O.: Proximal policy optimiza- tion algorithms. CoRR abs/1707.06347 (2017) 1707.06347

    2017 arXiv

  45. [53]

    Vision research 48(22), 2345–2351 (2008)

    Burton, R.F.: The scaling of eye size in adult birds: relationship to brain, head and body sizes. Vision research 48(22), 2345–2351 (2008)

    2008

  46. [54]

    Proceedings of the Royal Society of London

    Brooke, M.d.L., Hanley, S., Laughlin, S.: The scaling of eye size with body mass in birds. Proceedings of the Royal Society of London. Series B: Biological Sciences266(1417), 405–412 (1999)

    1999

  47. [55]

    Journal of Comparative Physiology A 189, 579–588 (2003)

    Wehner, R.: Desert ant navigation: how miniature brains solve complex tasks. Journal of Comparative Physiology A 189, 579–588 (2003)

    2003

  48. [56]

    Annual review of genetics 43(1), 25–47 (2009)

    Jeffery, W.R.: Regressive evolution in astyanax cavefish. Annual review of genetics 43(1), 25–47 (2009)

    2009

  49. [57]

    Journal of Experimental Biology 214, 2014–2019 (2011) https://doi.org/10.1242/jeb.053553

    Hogg, C., Neveu, M., Stokkan, K., Folkow, L., Cottrill, P., Douglas, R., Hunt, D., Jeffery, G.: Arctic reindeer extend their visual range into the ultraviolet. Journal of Experimental Biology 214, 2014–2019 (2011) https://doi.org/10.1242/jeb.053553

    2011 doi

  50. [58]

    the keeler lecture 2017: centenary of keeler ltd

    Schwab, I.R.: The evolution of eyes: major steps. the keeler lecture 2017: centenary of keeler ltd. Eye 32(2), 302–313 (2018)

    2018

  51. [59]

    Visual neuroscience 30(1-2), 5–20 (2013)

    Nilsson, D.-E.: Eye evolution and its functional basis. Visual neuroscience 30(1-2), 5–20 (2013)

    2013

  52. [60]

    IEEE Transactions on Pattern Analysis and Machine Intelligence 42(7), 1642–1653 (2020) https://doi.org/10.1109/TPAMI.2020.2986944

    Martel, J.N.P., M¨ uller, L.K., Carey, S.J., Dudek, P., Wetzstein, G.: Neural sensors: Learning pixel exposures for hdr imaging and video compressive sensing with programmable sensors. IEEE Transactions on Pattern Analysis and Machine Intelligence 42(7), 1642–1653 (2020) https...

    2020

  53. [61]

    In: 2019 IEEE International Conference on Computational Photography (ICCP), pp

    Wu, Y., Boominathan, V., Chen, H., Sankaranarayanan, A., Veeraraghavan, A.: Phase- cam3d—learning phase masks for passive single view depth estimation. In: 2019 IEEE International Conference on Computational Photography (ICCP), pp. 1–12 (2019). IEEE

    2019

  54. [62]

    ACM Trans

    Sitzmann, V., Diamond, S., Peng, Y., Dun, X., Boyd, S., Heidrich, W., Heide, F., Wetzstein, G.: End-to-end optimization of optics and image processing for achromatic extended depth of field and super-resolution imaging. ACM Trans. Graph. 37(4) (2018) https://doi.org/10. 1145/3...

    2018

  55. [63]

    Allometry: The study of biological scaling. (2018). https://api.semanticscholar.org/ CorpusID:199531454

    2018

  56. [64]

    Nature Ecology & Evolution 8(8), 1534–1542 (2024)

    Venditti, C., Baker, J., Barton, R.A.: Co-evolutionary dynamics of mammalian brain and body size. Nature Ecology & Evolution 8(8), 1534–1542 (2024)

    2024

  57. [65]

    Trends 28 in Ecology & Evolution 33(5), 358–372 (2018) https://doi.org/10.1016/j.tree.2018.03.001

    Caves, E.M., Brandley, N.C., Johnsen, S.: Visual acuity and the evolution of signals. Trends 28 in Ecology & Evolution 33(5), 358–372 (2018) https://doi.org/10.1016/j.tree.2018.03.001

    2018 doi

  58. [66]

    Attention, Perception, & Psychophysics 76, 383–390 (2014)

    Abrams, R.A., Weidler, B.J.: Trade-offs in visual processing for stimuli near the hands. Attention, Perception, & Psychophysics 76, 383–390 (2014)

    2014

  59. [67]

    Current Biology 32(16), 3529–3544 (2022)

    Shinomiya, K., Nern, A., Meinertzhagen, I.A., Plaza, S.M., Reiser, M.B.: Neuronal circuits integrating visual motion information in drosophila melanogaster. Current Biology 32(16), 3529–3544 (2022)

    2022

  60. [68]

    https://arxiv.org/abs/1803.03453

    Lehman, J., Clune, J., Misevic, D., Adami, C., Altenberg, L., Beaulieu, J., Bentley, P.J., Bernard, S., Beslon, G., Bryson, D.M., Chrabaszcz, P., Cheney, N., Cully, A., Doncieux, S., Dyer, F.C., Ellefsen, K.O., Feldt, R., Fischer, S., Forrest, S., Fr´ enoy, A., Gagn´ e, C., Go...

    2019 arXiv

  61. [69]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Klinghoffer, T., Tiwary, K., Behari, N., Agrawalla, B., Raskar, R.: Diser: Designing imag- ing systems with reinforcement learning. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 23632–23642 (2023)

    2023

  62. [70]

    An MIT Exploration of Generative AI (2024)

    Tiwary, K., Klinghoffer, T., Young, A., Somasundaram, S., Behari, N., Dave, A., Che- ung, B., Nilsson, D.-E., Poggio, T., Raskar, R.: A Roadmap for Generative Design of Visual Intelligence. An MIT Exploration of Generative AI (2024). https://mit- genai.pubpub.org/pub/bcfcb6lu

    2024

  63. [71]

    Journal of Machine Learning Research 22(268), 1–8 (2021)

    Raffin, A., Hill, A., Gleave, A., Kanervisto, A., Ernestus, M., Dormann, N.: Stable- baselines3: Reliable reinforcement learning implementations. Journal of Machine Learning Research 22(268), 1–8 (2021)

    2021

  64. [72]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Henderson, P., Islam, R., Bachman, P., Pineau, J., Precup, D., Meger, D.: Deep rein- forcement learning that matters. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 32 (2018)

    2018

  65. [73]

    Evolutionary computation 9(2), 159–195 (2001)

    Hansen, N., Ostermeier, A.: Completely derandomized self-adaptation in evolution strate- gies. Evolutionary computation 9(2), 159–195 (2001)

    2001

  66. [74]

    Biology Direct 7, 1–7 (2012)

    Holl´ o, G., Nov´ ak, M.: The manoeuvrability hypothesis to explain the maintenance of bilateral symmetry in animal evolution. Biology Direct 7, 1–7 (2012)

    2012

  67. [75]

    Addison-Wesley Longman Publishing Co., Inc., ??? (1999)

    Woo, M., Neider, J., Davis, T., Shreiner, D.: OpenGL Programming Guide: the Official Guide to Learning OpenGL, Version 1.2. Addison-Wesley Longman Publishing Co., Inc., ??? (1999)

    1999

  68. [76]

    In: European Conference on Computer Vision, pp

    Tasneem, Z., Milione, G., Tsai, Y.-H., Yu, X., Veeraraghavan, A., Chandraker, M., Pittaluga, F.: Learning phase mask for privacy-preserving passive depth estimation. In: European Conference on Computer Vision, pp. 504–521 (2022). Springer

    2022

  69. [77]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Chang, J., Wetzstein, G.: Deep optics for monocular depth estimation and 3d object detec- tion. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 10193–10202 (2019) 29

    2019

  70. [78]

    https://arxiv.org/abs/1904.08601

    Chang, J., Wetzstein, G.: Deep Optics for Monocular Depth Estimation and 3D Object Detection (2019). https://arxiv.org/abs/1904.08601

    2019 arXiv

  71. [79]

    Roberts and Company publishers, ??? (2005)

    Goodman, J.W.: Introduction to Fourier Optics. Roberts and Company publishers, ??? (2005)

    2005

  72. [80]

    Adelson, E., Bergen, J.: The plenoptic function and the elements of early vision (1997)

    1997

  73. [81]

    Nature 333(6173), 557–560 (1988)

    Marshall, N.J.: A unique colour and polarization vision system in mantis shrimps. Nature 333(6173), 557–560 (1988)

    1988

  74. [82]

    Journal of experimental biology 51(2), 443–470 (1958)

    Land, M.: Structure of the retinae of the principal eyes of jumping spiders (salticidae: Dendryphantinae) in relation to visual optics. Journal of experimental biology 51(2), 443–470 (1958)

    1958

  75. [83]

    https://arxiv.org/abs/2203

    Hoffmann, J., Borgeaud, S., Mensch, A., Buchatskaya, E., Cai, T., Rutherford, E., Las Casas, D., Hendricks, L.A., Welbl, J., Clark, A., Hennigan, T., Noland, E., Millican, K., Driessche, G., Damoc, B., Guy, A., Osindero, S., Simonyan, K., Elsen, E., Rae, J.W., Vinyals, O., Sif...

    2022

  76. [84]

    https://arxiv.org/abs/1712.00409 30

    Hestness, J., Narang, S., Ardalani, N., Diamos, G., Jun, H., Kianinejad, H., Patwary, M.M.A., Yang, Y., Zhou, Y.: Deep Learning Scaling is Predictable, Empirically (2017). https://arxiv.org/abs/1712.00409 30

    2017 arXiv

  77. [328]

    https://doi.org/ 10.1145/122718.122752

    Association for Computing Machinery, New York, NY, USA (1991). https://doi.org/ 10.1145/122718.122752 . https://doi.org/10.1145/122718.122752

    1991

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.