REVIEW 3 major objections 5 minor 50 references
AutoSciLab: A Self-Driving Laboratory For Interpretable Scientific Discovery
T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read AutoSciLab claims to run the whole scientific method autonomously and discovers a new equation for steering incoherent light.
desk verdict A coherent integration of known ML components whose headline nanophotonics discovery rests on an unvalidated surrogate and whose Ising benchmark is circular. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The pipeline has four coupled components. The variational autoencoder learns a smooth low-dimensional latent space from a training set of experiments and can sample novel experiments beyond that set. The active-learning agent fits a Gaussian process to the latent space and uses an acquisition function, Expected Improvement or Upper Confidence Bound, to propose the next experiment. The directional autoencoder is an ordinary autoencoder with a distance-based regularization that forces latent directions to correlate with known physics-informed attributes such as local pump-pattern slope and curvature. The neural-network equation learner uses a dictionary of physics-inspired activations, prunes connections by their contribution to later activations rather than by raw weight size, and reads out the surviving connectivity as a symbolic equation.
What would settle it
Run the active-learning loop with direct physical measurements from the metasurface setup at several emission angles and compare measured directivities to the surrogate predictions and to Eq. 8; if the physical directivities do not track the surrogate in the high-directivity region or Eq. 8 fails to predict out-of-sample angles, the central nanophotonics discovery is unsupported.
Extended reading notes
Core claim
The paper's central claim is that one framework can carry out the full scientific-discovery loop without human intuition at each step: a variational autoencoder proposes high-dimensional experiments, an active-learning agent selects which experiments to run by maximizing a quantity of interest, a directional autoencoder distills the explored experiments into a low-dimensional latent space aligned with physics-informed variables, and a neural-network equation learner converts that latent space into a symbolic relationship $y = f(z)$. The authors validate the loop by rediscovering the projectile height law $H = u^2/2g$ and the Ising magnetization curve $M(\beta) = (1-\sinh(2\beta J)^{-4})^{1/8}$. For the nanophotonics case, they claim AutoSciLab finds pump patterns with 3-4x higher directivity than the state of the art and yields the new equation $D_e = 0.0467 z_1^2 - 0.0265 z_2^2 - 0.175 z_1 - 0.0955 z_1 z_2 + 0.22 z_2 + 2.707$, which they propose as a novel structure-property relationship for incoherent light steering.
Load-bearing premise
In the nanophotonics demonstration, the ground-truth directivity values come from a neural network surrogate trained on roughly 8000 brute-force samples of the latent space, not from physical measurements, so the claimed 3-4x improvement and Eq. 8 depend on that surrogate being accurate in the unexplored high-directivity regions.
Editorial extensions
If this is right
- If the framework works as claimed, autonomous laboratories can search high-dimensional design spaces without a human first narrowing the hypothesis space.
- In the nanophotonics exemplar, active learning reaches high-directivity pump patterns with about an order of magnitude fewer experiments than differential evolution under the same budget.
- The discovered equation $D_e = 0.0467 z_1^2 - 0.0265 z_2^2 - 0.175 z_1 - 0.0955 z_1 z_2 + 0.22 z_2 + 2.707$ provides a direct, symbolic target for designing metasurface pump patterns.
- The measured gain factor of AutoSciLab over human intuition grows with the dimensionality of the design space, suggesting the largest payoff occurs in open-ended problems where intuition is weakest.
- The same four-step loop is claimed to generalize beyond photonics to any physical domain with expensive, noisy experiments, including energy-relevant problems such as LED and thermal-lamp design.
Reading between the lines
- Beyond the paper, if Eq. 8 is physically reliable it would imply that directivity can be optimized by directly tuning the two latent axes $z_1$ and $z_2$ rather than the full high-dimensional pump pattern, which is a simpler inverse-design problem.
- The polynomial, non-oscillatory form of Eq. 8 suggests that the Fourier/momentum-matching picture of incoherent steering is incomplete; testing this equation on other emission angles and metasurface designs would reveal whether the discovered relationship is universal or particular to this geometry.
- A natural extension the authors did not run is a closed-loop physical experiment across many angles that uses Eq. 8 as the acquisition model, which would simultaneously validate the equation and measure its extrapolation limits.
- The framework's reliance on human-chosen activation-function dictionaries and physics-informed variables means its 'autonomy' is bounded by the prior knowledge encoded in those choices; the interesting next question is how much discovery is possible when that prior is wrong.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes AutoSciLab, an integrated machine-learning pipeline for scientific discovery that combines a variational autoencoder for generating high-dimensional experiments, an active-learning agent for selecting experiments, a directional autoencoder for discovering interpretable latent variables, and a neural-network equation learner for producing symbolic relationships. The authors validate the framework on three exemplars: projectile motion, the two-dimensional Ising model, and an open-ended nanophotonics metasurface problem. For the nanophotonics case they report Eq. (8) as a discovered directivity equation and claim 3–4x higher directivity than prior work. The appendix documents network architectures, surrogate construction, active-learning baselines, and the human-in-the-loop equation-pruning procedure.
Significance. If the nanophotonics claim and the two rediscoveries were fully supported, the paper would be a useful demonstration of an integrated discovery loop, and the component-level comparisons (notably active learning versus differential evolution and the nn-EQL benchmarks in Table 1) provide some useful evidence. The authors are also transparent about several limitations, including the statement that ground-truth directivity is obtained from a neural-network surrogate. However, the two main scientific validations are currently either circular or validated only against an uncharacterized surrogate, and Eq. (8) is partly human-curated with no reported fit statistics. As a result, the paper's central claim of autonomous discovery of a novel physical principle is not established.
major comments (3)
- [Exemplar 2, 'Results'] The Ising exemplar is circular. The target transform y_i = (1/(1-M_i^8))^(1/4) is exactly the inverse of the Onsager formula M(β) = (1 - sinh(2β)^(-4))^(1/8) stated as the objective. Applying this transform to the labels reduces the learning task to fitting y = sinh(2β), so the reported equation y = 1.01sinh(1.96β) + 0.1β - 0.1 is a fit to a transformed version of the known answer rather than an independent rediscovery. The exemplar should either be removed from the validation claims or reformulated to learn M(T) directly; in its current form it does not test the discovery pipeline.
- [Exemplar 3, 'Results'; Appendix S3] The headline nanophotonics claim rests entirely on a neural-network surrogate. Appendix S3 states that the surrogate is trained on ~8000 brute-force samples of the VAE latent space, and the active learner maximizes the surrogate's predicted directivity over that same latent space. The reported 3-4x directivity improvement and Eq. (8) are therefore statements about surrogate predictions in regions where the surrogate has not been validated; no accuracy, uncertainty, or error analysis is provided for those regions. The closed-loop noisy-experiment results in Fig. 8 are explicitly described only as initial and are not connected to Eq. (8) or to the specific discovered pump patterns. Without physical measurements or a high-fidelity surrogate-error study, the claimed 'fundamentally novel method' is unsupported.
- [Appendix S5] Eq. (8) is not an autonomous pipeline output in the sense claimed. The activation-function dictionary is chosen manually, the training data are manually balanced, and the final parsimonious equation is obtained by manually inspecting and removing terms. The paper also reports no training or validation error, R², or prediction intervals for Eq. (8), so its accuracy and domain of validity are unquantified. At minimum, Eq. (8) should be labeled as a human-curated fit and accompanied by fit statistics before being presented as a discovered law.
minor comments (5)
- [The AutoSciLab framework] The phrase 'in the latent in the latent space' contains a duplicated fragment and should be corrected.
- [Exemplar 2, 'Problem description'] The text calls the 2D Ising model NP-hard, but the zero-field 2D Ising model is exactly solvable; NP-hardness applies to related problems such as spin glasses. The framing should be revised to avoid overclaiming the difficulty of the benchmark.
- [Fig. 4 caption] The caption contains the typo 'reperesent' for 'represent'.
- [Appendix S6 / Table 2] The 'Gain Factor' calculation for the Ising row sets the variety ratio to 1 because both methods use the magnetization representation; this comparison is partly definitional and should be stated more carefully.
- [Reproducibility] No code or data repository is provided. For a framework whose central claims are numerical and whose main results depend on trained models, a public implementation and trained-model artifacts would substantially aid verification.
Circularity Check
Ising 'rediscovery' fits the inverted Onsager relation by construction; projectile validation encodes Newton's law in the AL objective; the nanophotonics claim is not circular but relies on an unvalidated surrogate.
-
self definitional
[Exemplar 2: Spin dynamics of the two-dimensional Ising system, 'Results' (main text)]
"For ease of training the nn-EQL, we transform the dataset to be Y ′ = (y1, β1), (y2, β2)...(yn, βn), where β = 1/kT and yi = (1/(1−M^8))^(1/4). The equation to be learnt is thus y = sinh(2β), and we find that our neural network equation learner finds y = 1.01sinh(1.96β) + 0.1β − 0.1, which is close to the ground truth."
Setting J=1, the Onsager relation stated by the paper, M(β) = (1−sinh(2β)^(−4))^(1/8), inverts algebraically to (1−M^8)^(−1/4) = sinh(2β). Thus y is defined as exactly the known answer expressed as a function of β. The nn-EQL is then asked to fit y versus β, so the claimed rediscovery of the Ising magnetization law reduces to fitting the function the authors already inserted through the coordinate transform. The fitted coefficients (≈1.01, ≈1.96, ≈0.1) merely confirm the transform rather than provide an independent derivation of M(T).
-
other
[Exemplar 1: Projectile motion, 'AutoSciLab framing of task']
"Here we utilize the fact that realistic trajectories must have a constant acceleration g, defining the active learning objective to be to minimize the difference between the second derivative of a candidate trajectory y(t) and g."
The active-learning objective injects the target Newtonian law (constant acceleration g) as the selection criterion. Any trajectory surviving this filter satisfies approximately y'' = −g, so the maximum height is H(u) = u^2/(2g) by elementary kinematics. The rediscovered equation is therefore a mathematical consequence of the reward used to choose experiments, not an independent first-principles result extracted from unconstrained data. Because this is a benchmark validation, the circularity is limited to the exemplar's evidentiary value for end-to-end discovery.
full rationale
The Ising exemplar is the clearest circular step. The paper first states the target Onsager law, then defines y = (1−M^8)^(−1/4), which is precisely sinh(2β) when M obeys that law; the nn-EQL therefore rediscovers the Onsager equation in coordinates constructed from that equation. The projectile exemplar is a milder but related issue: the AL objective is the constant-acceleration law itself, so the subsequent H(u) = u^2/(2g) is a consequence of the experiment-selection rule. The nanophotonics result is not circular in the definitional sense: the neural-network surrogate is a fitted evaluator, but presenting its extrapolated maximal directivity as ground-truth discovery is an external-validation/correctness risk, not a reduction of the claim to its inputs. There is no load-bearing self-citation chain or imported uniqueness theorem. Because two benchmark rediscoveries reduce by construction while the novel nanophotonics claim retains independent, though not yet physically confirmed, content, the circularity score is 6.
Assumptions & free parameters
free parameters (6)
- Equation coefficients (nanophotonics, Eq. 8) =
0.0467, -0.0265, -0.175, -0.0955, 0.22, 2.707
- Ising equation coefficients =
1.01, 1.96, 0.1, -0.1
- VAE latent dimension d =
4 (nanophotonics)
- Directional autoencoder latent dimension =
4 (nanophotonics)
- UCB acquisition lambda =
not stated
- nn-EQL hyperparameters =
sparsity ~90%, pruning rate k ~2%, activation set chosen manually
assumptions (6)
- domain assumption Manifold hypothesis: high-dimensional experimental inputs lie on a low-dimensional manifold.
- domain assumption The neural network surrogate model accurately predicts physical directivity for pump patterns in the VAE latent space, including at the extremes explored by active learning.
- ad hoc to paper The target transform y = (1-M^8)^(-1/4) is an appropriate encoding for the Ising magnetization data.
- ad hoc to paper For projectile motion, the AL objective minimizes |y''(t) - g| and the VAE training set contains trajectories built from u sin theta t - 0.5 g t^2.
- domain assumption Prior knowledge that local slope and curvature of the pump pattern influence directivity is accurate and is encoded into the directional autoencoder regularization.
- domain assumption The Gaussian process regression with Matern 5/2 kernel and zero noise is a suitable model for the active learning loop.
invented entities (2)
-
Directional autoencoder latent variables z1 and z2
-
The discovered directivity equation (Eq. 8)
Cite this review
Pith. "Pith review of AutoSciLab: A Self-Driving Laboratory For Interpretable Scientific Discovery." pith.science (2026). https://pith.science/paper/OVYCXPGR
@misc{pith2026241212347,
author = {Pith},
title = {Pith review of: AutoSciLab: A Self-Driving Laboratory For Interpretable Scientific Discovery},
year = {2026},
howpublished = {\url{https://pith.science/paper/OVYCXPGR}},
note = {Machine review of arXiv:2412.12347}
}
read the original abstract
Advances in robotic control and sensing have propelled the rise of automated scientific laboratories capable of high-throughput experiments. However, automated scientific laboratories are currently limited by human intuition in their ability to efficiently design and interpret experiments in high-dimensional spaces, throttling scientific discovery. We present AutoSciLab, a machine learning framework for driving autonomous scientific experiments, forming a surrogate researcher purposed for scientific discovery in high-dimensional spaces. AutoSciLab autonomously follows the scientific method in four steps: (i) generating high-dimensional experiments (x \in R^D) using a variational autoencoder (ii) selecting optimal experiments by forming hypotheses using active learning (iii) distilling the experimental results to discover relevant low-dimensional latent variables (z \in R^d, with d << D) with a 'directional autoencoder' and (iv) learning a human interpretable equation connecting the discovered latent variables with a quantity of interest (y = f(z)), using a neural network equation learner. We validate the generalizability of AutoSciLab by rediscovering a) the principles of projectile motion and b) the phase transitions within the spin-states of the Ising model (NP-hard problem). Applying our framework to an open-ended nanophotonics challenge, AutoSciLab uncovers a fundamentally novel method for directing incoherent light emission that surpasses the current state-of-the-art (Iyer et al. 2023b, 2020).
Figures
Figures from the paper (8 more)
Reference graph
Works this paper leans on
-
[1]
Abolhasani, M.; and Kumacheva, E. 2023. The rise of self-driving labs in chemical and materials sciences. Nature Synthesis, 2(6): 483--492
work page 2023
-
[2]
Anstine, D. M.; and Isayev, O. 2023. Generative models as an emerging paradigm in the chemical sciences. Journal of the American Chemical Society, 145(16): 8736--8750
work page 2023
-
[3]
Bakshy, E.; Dworkin, L.; Karrer, B.; Kashin, K.; Letham, B.; Murthy, A.; and Singh, S. 2018. AE: A domain-agnostic platform for adaptive experimentation. In Conference on neural information processing systems, 1--8
work page 2018
-
[4]
Barahona, F. 1982. On the computational complexity of Ising spin glass models. Journal of Physics A: Mathematical and General, 15(10): 3241
work page 1982
-
[5]
W.; Choudhary, A.; Agrawal, A.; Billinge, S
Choudhary, K.; DeCost, B.; Chen, C.; Jain, A.; Tavazza, F.; Cohn, R.; Park, C. W.; Choudhary, A.; Agrawal, A.; Billinge, S. J.; et al. 2022. Recent advances and applications of deep learning methods in materials science. npj Computational Materials, 8(1): 59
2022
-
[6]
L.; Lei, B.; Francis, T.; and Holm, E
DeCost, B. L.; Lei, B.; Francis, T.; and Holm, E. A. 2019. High throughput quantitative metallography for complex microstructures using deep learning: A case study in ultrahigh carbon steel. Microscopy and Microanalysis, 25(1): 21--29
work page 2019
-
[7]
Desai, S.; and Strachan, A. 2021. Parsimonious neural networks learn interpretable physical laws. Scientific reports, 11(1): 12761
work page 2021
-
[8]
Fefferman, C.; Mitter, S.; and Narayanan, H. 2016. Testing the manifold hypothesis. Journal of the American Mathematical Society, 29(4): 983--1049
work page 2016
Show all 50 references
-
[9]
Feng, S.; Zhou, H.; and Dong, H. 2019. Using deep neural network with small dataset to predict material defects. Materials & Design, 162: 300--310
2019
-
[10]
Goodfellow, I.; Pouget-Abadie, J.; Mirza, M.; Xu, B.; Warde-Farley, D.; Ozair, S.; Courville, A.; and Bengio, Y. 2020. Generative adversarial networks. Communications of the ACM, 63(11): 139--144
2020
-
[11]
Han, S.; Pool, J.; Tran, J.; and Dally, W. 2015. Learning both weights and connections for efficient neural network. Advances in neural information processing systems, 28
2015
-
[12]
P.; DeCrescent, R
Iyer, P. P.; DeCrescent, R. A.; Mohtashami, Y.; Lheureux, G.; Butakov, N. A.; Alhassan, A.; Weisbuch, C.; Nakamura, S.; DenBaars, S. P.; and Schuller, J. A. 2020. Unidirectional luminescence from InGaN/GaN quantum-well metasurfaces. Nature Photonics, 14(9): 543--548
2020
-
[13]
P.; Desai, S.; Addamane, S.; Dingreville, R.; and Brener, I
Iyer, P. P.; Desai, S.; Addamane, S.; Dingreville, R.; and Brener, I. 2023 a . Learning incoherent light emission steering from metasurfaces using generative models. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 3770--3777
2023
-
[14]
P.; Karl, N.; Addamane, S.; Gennaro, S
Iyer, P. P.; Karl, N.; Addamane, S.; Gennaro, S. D.; Sinclair, M. B.; and Brener, I. 2023 b . Sub-picosecond steering of ultrafast incoherent emission from semiconductor metasurfaces. Nature Photonics, 1--6
2023
-
[15]
Jumper, J.; Evans, R.; Pritzel, A.; Green, T.; Figurnov, M.; Ronneberger, O.; Tunyasuvunakool, K.; Bates, R.; Z \' dek, A.; Potapenko, A.; et al. 2021. Highly accurate protein structure prediction with AlphaFold. Nature, 596(7873): 583--589
2021
-
[16]
Kim, C.; Batra, R.; Chen, L.; Tran, H.; and Ramprasad, R. 2021. Polymer design using genetic algorithm and machine learning. Computational Materials Science, 186: 110067
2021
-
[17]
H.; Aspuru-Guzik, A.; and Jung, Y
Kim, S.; Noh, J.; Gu, G. H.; Aspuru-Guzik, A.; and Jung, Y. 2020. Generative adversarial networks for crystal structure prediction. ACS central science, 6(8): 1412--1420
2020
-
[18]
P.; and Ba, J
Kingma, D. P.; and Ba, J. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980
2014 arXiv
-
[19]
P.; and Welling, M
Kingma, D. P.; and Welling, M. 2013. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114
2013 arXiv
-
[20]
Koyr \'e , A. 2013. The Astronomical Revolution: Copernicus-Kepler-Borelli. Routledge
2013
-
[21]
Kramer, S.; Cerrato, M.; D z eroski, S.; and King, R. 2023. Automated scientific discovery: from equation discovery to autonomous discovery systems. arXiv preprint arXiv:2305.02251
2023 arXiv
-
[22]
Krizhevsky, A.; Sutskever, I.; and Hinton, G. E. 2012. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25
2012
-
[23]
G.; Yu, H.; Wu, C.; Zhang, H.; Hattrick-Simpers, J.; DeCost, B.; Sarker, S.; Oses, C.; Toher, C.; Curtarolo, S.; et al
Kusne, A. G.; Yu, H.; Wu, C.; Zhang, H.; Hattrick-Simpers, J.; DeCost, B.; Sarker, S.; Oses, C.; Toher, C.; Curtarolo, S.; et al. 2020. On-the-fly closed-loop materials discovery via Bayesian active learning. Nature communications, 11(1): 5966
2020
-
[24]
O.; Virgolin, M.; Jin, Y.; Kommenda, M.; and Moore, J
La Cava, W.; Orzechowski, P.; Burlacu, B.; de Fran c a, F. O.; Virgolin, M.; Jin, Y.; Kommenda, M.; and Moore, J. H. 2021. Contemporary symbolic regression methods and their relative performance. arXiv preprint arXiv:2107.14351
2021 arXiv
-
[25]
Langley, P. 2024. Integrated Systems for Computational Scientific Discovery. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 22598--22606
2024
-
[26]
LeCun, Y.; Denker, J.; and Solla, S. 1989. Optimal brain damage. Advances in neural information processing systems, 2
1989
-
[27]
Ling, J.; Hutchinson, M.; Antono, E.; Paradiso, S.; and Meredig, B. 2017. High-dimensional materials and process optimization using data-driven experimental design with well-calibrated uncertainty estimates. Integrating Materials and Manufacturing Innovation, 6: 207--217
2017
-
[28]
Liu, G.; Stintz, A.; Li, H.; Newell, T.; Gray, A.; Varangis, P.; Malloy, K.; and Lester, L. 2000. The influence of quantum-well composition on the performance of quantum dot lasers using InAs-InGaAs dots-in-a-well (DWELL) structures. IEEE Journal of Quantum Electronics, 36(11)...
2000
-
[29]
M.; and Lee, S.-I
Lundberg, S. M.; and Lee, S.-I. 2017. A unified approach to interpreting model predictions. Advances in neural information processing systems, 30
2017
-
[30]
P.; Parlane, F
MacLeod, B. P.; Parlane, F. G.; Morrissey, T. D.; H \"a se, F.; Roch, L. M.; Dettelbach, K. E.; Moreira, R.; Yunker, L. P.; Rooney, M. B.; Deeth, J. R.; et al. 2020. Self-driving laboratory for accelerated discovery of thin-film materials. Science Advances, 6(20)
2020
-
[31]
F.; Dorschner, T
McManamon, P. F.; Dorschner, T. A.; Corkum, D. L.; Friedman, L. J.; Hobbs, D. S.; Holz, M.; Liberman, S.; Nguyen, H. Q.; Resler, D. P.; Sharp, R. C.; et al. 1996. Optical phased array technology. Proceedings of the IEEE, 84(2): 268--298
1996
-
[32]
Mendoza-Alvarez, J.; Yan, R.; and Coldren, L. 1987. Contribution of the band-filling effect to the effective refractive-index change in double-heterostructure GaAs/AlGaAs phase modulators. Journal of applied physics, 62(11): 4548--4553
1987
-
[33]
P.; Paprocki, M.; C ert \' k, O.; Kirpichev, S
Meurer, A.; Smith, C. P.; Paprocki, M.; C ert \' k, O.; Kirpichev, S. B.; Rocklin, M.; Kumar, A.; Ivanov, S.; Moore, J. K.; Singh, S.; et al. 2017. SymPy: symbolic computing in Python. PeerJ Computer Science, 3: e103
2017
-
[34]
Narayanan, H.; and Mitter, S. 2010. Sample complexity of testing the manifold hypothesis. Advances in neural information processing systems, 23
2010
-
[35]
Onsager, L. 1944. Crystal statistics. I. A two-dimensional model with an order-disorder transition. Physical Review, 65(3-4): 117
1944
-
[36]
Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; et al. 2019. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32
2019
-
[37]
Pati, A.; and Lerch, A. 2019. Latent space regularization for explicit control of musical attributes. In ICML Machine Learning for Music Discovery Workshop (ML4MD), Extended Abstract, Long Beach, CA, USA
2019
-
[38]
Pichler, H.; Daley, A.; and Zoller, P. 2010. Nonequilibrium dynamics of bosonic atoms in optical lattices: Decoherence of many-body states due to spontaneous emission. Physical Review A, 82(6): 063605
2010
-
[39]
Raissi, M.; Perdikaris, P.; and Karniadakis, G. E. 2019. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational physics, 378: 686--707
2019
-
[40]
E.; Williams, C
Rasmussen, C. E.; Williams, C. K.; et al. 2006. Gaussian processes for machine learning, volume 1. Springer
2006
-
[41]
Sahoo, S.; Lampert, C.; and Martius, G. 2018. Learning equations for extrapolation and control. In International Conference on Machine Learning, 4442--4450. PMLR
2018
-
[42]
Schmidt, M.; and Lipson, H. 2009. Distilling free-form natural laws from experimental data. science, 324(5923): 81--85
2009
-
[43]
T.; Vestfrid, J.; Wu, T
Seifrid, M.; Pollice, R.; Aguilar-Granda, A.; Morgan Chan, Z.; Hotta, K.; Ser, C. T.; Vestfrid, J.; Wu, T. C.; and Aspuru-Guzik, A. 2022. Autonomous chemical experiments: Challenges and perspectives on establishing a self-driving lab. Accounts of Chemical Research, 55(17): 2454--2466
2022
-
[44]
Sobol', I. M. 1967. On the distribution of points in a cube and the approximate evaluation of integrals. Zhurnal Vychislitel'noi Matematiki i Matematicheskoi Fiziki, 7(4): 784--802
1967
-
[45]
Storn, R.; and Price, K. 1997. Differential evolution--a simple and efficient heuristic for global optimization over continuous spaces. Journal of global optimization, 11: 341--359
1997
-
[46]
Udrescu, S.-M.; and Tegmark, M. 2020. AI Feynman: A physics-inspired method for symbolic regression. Science Advances, 6(16): eaay2631
2020
-
[47]
N.; Kaiser, .; and Polosukhin, I
Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A. N.; Kaiser, .; and Polosukhin, I. 2017. Attention is all you need. Advances in neural information processing systems, 30
2017
-
[48]
K.; Fu, S
Zhong, Y. K.; Fu, S. M.; Ju, N. P.; Chen, P. Y.; and Lin, A. 2015. Experimentally-implemented genetic algorithm (Exp-GA): toward fully optimal photovoltaics. Optics Express, 23(19): A1324--A1333
2015
-
[49]
, " * write output.state after.block = add.period write newline
ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...
-
[50]
write newline
" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.