Pith. sign in

REVIEW 3 major objections 6 minor 33 references

Learning Epidemiological Dynamics via the Finite Expression Method

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read The paper claims that the Finite Expression Method, a reinforcement-learning-based symbolic regression approach, learns explicit differential equations of epidemiological dynamics from data, achieving test errors around…

desk verdict FEX is a real method with credible synthetic results, but the real-data forecasting claim is unsupported as presented. read the letter →

arxiv 2412.21049 v1 pith:U3UNXBHR submitted 2024-12-30 cs.LG cs.NAmath.NA

classification cs.LGcs.NAmath.NA MSC 92D30
keywords finiteexpressionmethodsymbolicregressionreinforcementlearningepidemiologicaldynamicsSIRSEIRSEIRDCOVID-19forecastinginterpretablemachine
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

Epidemic models usually come from expert-written compartmental equations, while neural-network predictors are accurate but opaque. This paper argues that the Finite Expression Method (FEX) gets both: it searches over symbolic expressions using reinforcement learning and fits the right-hand side of the underlying ODEs directly from trajectory data. On synthetic SIR, SEIR, and SEIRD data, FEX reports test mean squared errors on the order of $10^{-8}$ to $10^{-7}$, orders of magnitude lower than neural network and recurrent network baselines. On real COVID-19 data from Hubei, the method produces explicit formulas for recovered, deceased, and active cases using only three state variables and tracks the observed 15-day trajectory. The significance, if these results hold, is a data-driven route to interpretable and accurate epidemiological models for rapid public-health response.

What carries the argument

The engine is the Finite Expression Method: a library of unary and binary operators (such as $x$, $x^2$, $\sin$, $\exp$, $+$, $-$, $\times$) is composed into binary expression trees, a reinforcement-learning controller generates operator sequences, and each sequence is scored by the inverse of its optimized loss, $S(e)=(1+L(e))^{-1}$. The policy is updated with a risk-seeking gradient to favor high-scoring expressions, and a candidate pool stores the best trees. The load-bearing identity for epidemiology is the per-component Euler residual loss (equation 3.5), which lets each coordinate of the vector field be learned independently from the same trajectory data; the authors use it with Euler integration and report that tree structures with one or two binary operators suffice for the tested models.

What would settle it

Take the synthetic SIR/SEIR/SEIRD generators used in the paper, add Gaussian observation noise with standard deviation $10^{-3}$ to the training trajectories, and run FEX with the same settings; if the recovered expressions no longer approximate the true right-hand sides or test MSE rises above $10^{-4}$, the central accuracy claim holds only for noiseless data. For the real-data claim, retrain on the first 70 days of the Hubei series and forecast days 71--100; if the 30-day forecast diverges from reported counts much faster than the 15-day forecast, the learned equations are fitting short-term trends rather than the true dynamics.

Watch

Extended reading notes

Core claim

The central claim is that a single symbolic-learning framework, FEX, can replace both manual model specification and black-box prediction in epidemiology. The paper demonstrates this by learning the full vector field of SIR, SEIR, and SEIRD models from simulated trajectories: each component $\phi^{(i)}_{FEX}(x)$ is trained on the one-step Euler residual $\|x^{(i)}_{s+1}-x^{(i)}_s-\phi^{(i)}_{FEX}(x_s)\Delta\|^2$, and the discovered expressions reproduce the generating dynamics with test MSE between $10^{-8}$ and $10^{-7}$, compared with $10^{-4}$--$10^{-2}$ for a feedforward network and $10^{-6}$--$10^{-5}$ for an LSTM. On the real Hubei COVID-19 dataset, FEX fits the first 85 days and forecasts the next 15 days with explicit equations for $dR/dt$, $dD/dt$, and $dQ/dt$ that depend only on $R$, $D$, and $Q$, which the authors present as an interpretable alternative to the five-variable SEIQRDP model.

Load-bearing premise

The accuracy and forecasting claims depend on the observed trajectories obeying the exact one-step Euler relation $x_{s+1}=x_s+\phi(x_s)\Delta$ with no measurement noise and with the same step size used in data generation or daily reporting; real epidemic data with reporting noise, delays, or hidden state variables would break this assumption.

Editorial extensions

If this is right

  • FEX can serve as an automatic model-discovery tool: given incidence and outcome time series, it outputs closed-form ODEs without requiring an expert to pre-specify the compartment structure.
  • Because the learned equations are explicit, they can be inspected, compared across regions, and used to test intervention scenarios in ways black-box networks cannot.
  • If the reported accuracy holds, FEX-type symbolic models could replace or validate hand-tuned compartmental models in outbreak settings where expert models take too long to adapt.
  • The method's reliance on only three state variables for the Hubei data suggests that lower-dimensional descriptions may suffice for forecasting, reducing data requirements for future applications.
  • The same loss and search loop apply to any ODE system, so the approach is portable to other infectious diseases or other dynamical-systems forecasting tasks.

Reading between the lines

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

  • The Euler-residual training and the small operator library mean the recovered expressions are only as good as the assumption that the observed process is Markovian and noise-free at the daily reporting scale; adding realistic reporting noise or delays to the Hubei experiment would likely degrade the 15-day forecasts, which is a testable prediction of the paper's own framing.
  • The non-uniqueness acknowledged by the authors suggests that the specific Hubei formulas are not 'the' underlying law; a stronger validation would be to see whether FEX recovers the same functional form across different Chinese provinces or across re-sampled training windows.
  • Because the search is combinatorial, the method's cost scales with the operator library and tree depth; a natural extension is to seed the search with terms from known compartmental models (e.g., $SI$, $I^2$, $\sin$ terms) to accelerate recovery on large systems.
  • If the accuracy on synthetic data transfers to real data with more compartments, FEX could be used to discover time-varying or spatially heterogeneous transmission rates, since the tree structure can represent products of state variables with coefficients.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper applies the Finite Expression Method (FEX), a reinforcement-learning-based symbolic regression framework, to learn governing differential equations for epidemiological dynamics. On synthetic SIR, SEIR, and SEIRD data generated by Euler's method, FEX reports test MSE values on the order of 1e-8 to 1e-7, outperforming neural-network and RNN baselines. On real COVID-19 data from Hubei (first 85 days training, 15 days test), FEX yields explicit expressions for dR/dt, dD/dt, dQ/dt and is compared visually with the fractional-order SEIQRDP model. The paper argues that FEX combines interpretability with high predictive accuracy and is a versatile tool for infectious-disease modeling.

Significance. Symbolic regression for epidemiological dynamics is a timely and potentially valuable direction, and the FEX framework itself is nontrivial: it uses RL-based combinatorial search and provides explicit expressions rather than black-box predictors. The synthetic experiments are clean and reproducible in principle, and the reported accuracy on the exact Euler-generated trajectories is strong evidence that FEX can recover known polynomial right-hand sides. However, the paper's central claim has two parts, and only the synthetic part is convincingly supported; the real-data forecast claim is not substantiated with quantitative metrics, error bars, or robustness checks. If the real-data results were properly validated, the contribution would be of interest to the machine-learning-for-dynamical-systems community. As it stands, the paper is a useful demonstration of FEX on controlled compartmental models, but its conclusion about practical epidemiological utility is overstated.

major comments (3)
  1. [§4.2, Eq. (3.5)] The real-data evaluation is not statistically supported. The loss (3.5) assumes the observed daily sequence satisfies x_{s+1} = x_s + f(x_s)Δ exactly, with no reporting noise, no delays, and with (R,D,Q) as a Markovian sufficient state; real COVID-19 reporting violates all three assumptions. The paper reports only visual alignment for the 15-day forecast and provides no error bars, no sensitivity analysis, no alternative time-window validation, and no comparison to a trivial baseline (e.g., persistence or a simple linear autoregressive model). The claim in Section 4.2 that FEX 'closely aligns with observed data' and 'demonstrates robust generalization' is therefore load-bearing but unsupported.
  2. [§4.1, Data Generation and Training Procedures] The synthetic comparison is favorable to FEX by construction. The data is generated with Euler's method at the same time step Δ used in the training loss (3.5), and the operator library {0,1,x,x^2,x^3,x^4,sin,cos,exp,+,-,×} contains the polynomial terms (e.g., S·I, E, I) appearing in the SIR, SEIR, and SEIRD right-hand sides. Consequently, the reported MSE of order 1e-8 demonstrates that the optimizer can recover an expression class that is exactly present in the search space, but it does not establish that FEX outperforms NN/RNN under realistic noisy or misspecified dynamics. Without experiments that add observation noise, use a different integrator for data generation, or test with an incomplete operator library, the claimed 'significant outperforming' (Section 4.1, Numerical Results) is overstated.
  3. [§4.2, Numerical Results] The learned equations for dR/dt, dD/dt, and dQ/dt are products of polynomials and sinusoids with many coefficients, and the paper offers no mechanistic interpretation or parsimony analysis. The claim in the abstract and introduction that FEX 'uncover[s] explicit relationships among epidemiological variables' and provides 'interpretability' is not substantiated, because the recovered expressions are not related to any known epidemiological structure and no analysis of their biological plausibility is given. Additionally, the paper does not describe the normalization or preprocessing applied to the real data, making the exact equations and the forecast plots unverifiable.
minor comments (6)
  1. [§2, Operator Sequence Generation] The action space of the controller is described only informally; for reproducibility, the paper should define precisely how positions in the tree map to the distributions p1_Φ, ..., ps_Φ and how the tree depth and arity constraints are enforced.
  2. [§2, Candidate Optimization] The hyperparameters T1, T2, T3, K, and the batch size for the RL controller are never given numerical values; without these, the experiments cannot be reproduced.
  3. [§4.1, RNN description] The RNN architecture is described as two LSTM layers with hidden size 51, but no details are given on sequence length, backpropagation through time, or whether the loss is computed on one-step or multi-step predictions, which is needed to interpret the comparison.
  4. [§4.2, Data Acquisition] The dataset is said to span 1,147 days, but only the first 100 days are used; please clarify why this specific window is chosen and give the exact date range used for training and testing.
  5. [References] Reference [31] (Teobaldi et al., PNAS) does not appear to be cited in the body of the paper; please remove it or cite it appropriately.
  6. [Eq. (3.5)] The notation in Eq. (3.5) uses the index i both for the component of the state and for the component-specific surrogate; consider using a different index (e.g., k) to avoid confusion.

Circularity Check

0 steps flagged · score 2.0 of 10

No material circularity: held-out test evaluation and out-of-sample forecast are genuine; minor non-load-bearing self-citations only.

full rationale

I reviewed the derivation chain from Eq. (3.5) through Sections 4.1 and 4.2. The training loss is a one-step Euler residual on observed transitions, and the synthetic data are generated with the same Euler scheme, with the candidate operator set containing the polynomial forms of the SIR/SEIR/SEIRD right-hand sides. This is a benchmark-design limitation that makes the reported 10^-8 MSE a within-class fitting result, but it is not circular: the testing MSE is computed on held-out trajectories from new initial conditions, so success still requires generalization across state space. The real-data 15-day forecast is an out-of-sample extrapolation from the first 85 days to the next 15 days; the learned equations are fitted rather than derived a priori, and the absence of error bars and a noise model is a support weakness, not a circular reduction. The FEX method is attributed to the authors' prior work [13,20,29], but Section 2 describes the RL search concretely enough that those self-citations are contextual rather than load-bearing. Section 5 candidly acknowledges non-uniqueness of solutions, showing the paper does not rely on an imported uniqueness theorem. I find no step in which a prediction reduces to its inputs by construction; the score of 2 reflects only the minor presence of non-load-bearing self-citations.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central results rest on the choice of a one-step Euler loss, the sufficiency of the selected state variables, and the expressiveness of the operator library. The fitted coefficients in the learned expressions are the main free parameters; there are no new physical entities.

free parameters (3)
  • Coefficients in learned expressions for dR/dt, dD/dt, dQ/dt = Example: -0.9030, 2.4025, -0.0262, 0.0311, -0.1840, -0.0432, -2.5147, -0.0181, 0.1919, 0.1812, 0.7006, -0.7283; 36…
    All constants in the three real-data expressions (Section 4.2) are fitted to the first 85 days of Hubei data. They are not derived from an epidemic model and carry no uncertainty estimates.
  • Expression tree structure selection (Type 1 vs Type 2) = Type 2 (two binary operators, three unary operators) for real data
    The number of binary operators and the operator library are chosen by hand per problem, affecting the expressiveness and the fitted coefficients.
  • FEX hyperparameters = Controller learning rate 0.002, exploration probability 0.1, batch size 10, epochs 100
    These hyperparameters are chosen by hand, not tuned on a validation set, and they affect the search and final expression quality.
assumptions (4)
  • standard math Euler's method is a valid integrator for the dynamics and for the loss function.
    Invoked in Section 3.2 Eq. (3.5) and in all training and testing pipelines; the synthetic data are also generated with Euler's method.
  • domain assumption The observed disease process is first-order Markovian and deterministic with no measurement noise.
    Section 3.2 assumes x_{s+1} = x_s + f(x_s) * Delta; real COVID-19 counts contain noise and reporting delays.
  • domain assumption The three variables R, D, Q are sufficient to describe the Hubei epidemic.
    Section 4.2 uses only R, D, Q as state variables, whereas the benchmark SEIQRDP uses more compartments.
  • ad hoc to paper The operator library {0, 1, x, x^2, x^3, x^4, sin, cos, exp, +, -, *} can approximate the true dynamics.
    Section 4.1 states the candidate operators; this choice constrains the class of discoverable expressions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Epidemiological Dynamics via the Finite Expression Method." pith.science (2026). https://pith.science/paper/U3UNXBHR

@misc{pith2026241221049,
  author       = {Pith},
  title        = {Pith review of: Learning Epidemiological Dynamics via the Finite Expression Method},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U3UNXBHR}},
  note         = {Machine review of arXiv:2412.21049}
}
read the original abstract

Modeling and forecasting the spread of infectious diseases is essential for effective public health decision-making. Traditional epidemiological models rely on expert-defined frameworks to describe complex dynamics, while neural networks, despite their predictive power, often lack interpretability due to their ``black-box" nature. This paper introduces the Finite Expression Method, a symbolic learning framework that leverages reinforcement learning to derive explicit mathematical expressions for epidemiological dynamics. Through numerical experiments on both synthetic and real-world datasets, FEX demonstrates high accuracy in modeling and predicting disease spread, while uncovering explicit relationships among epidemiological variables. These results highlight FEX as a powerful tool for infectious disease modeling, combining interpretability with strong predictive performance to support practical applications in public health.

Figures

Figures reproduced from arXiv: 2412.21049 by the authors.

Figure 1
Figure 1. The computational structure is represented us￾ing binary trees, where each node is assigned either a bi￾nary or unary operator. Expressions are recursively con￾structed, starting from depth-1 trees. Binary operators in￾clude B := {+, −, ×, ÷, . . .}, and unary operators include U := {sin, exp, log,Id,(·) 2 , R ·dxi , ∂· ∂xi , . . .}. node corresponds to an operator, forming an operator sequence e. Each operator is a… view at source ↗
Figure 2
Figure 2. Flowchart of the FEX method. The process consists of an iterative search loop (a), weight optimization (b), and expres￾sion generation to identify solutions for the target ODEs or PDEs. Key components include the Expression Tree, Controller, and Can￾didate Pool, which collaboratively refine expressions through sam￾pling, scoring, and optimization mechanisms. is sampled from p i Φ. Larger values of ϵ promote broader … view at source ↗
Figure 3
Figure 3. Illustration of two tree structures used in the FEX implementation. and a batch size of 32. Euler’s method [4, 28] is used as the integrator in the loss function (3.5). The RNN method: The RNN model consists of two Long Short-Term Memory (LSTM) layers and a linear output layer for predicting time-series data of the variables in the epidemiological models. The input to the RNN model comprises the variables from the t… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Comparison of MSE over time for three methods (FEX, RNN, NN) on (a) SIR model, (b) SEIR model, and (c) SEIRD model over 250 time steps. COVID-19, caused by the highly transmissible SARS-CoV-2 virus, emerged in late 2019 and rapidly escalated into a global pandemic. Und…
Figure 5
Figure 5. Figure 5: Comparison of actual and predicted COVID-19 cases using the FEX and SEIQRDP methods: (a) active cases (Q), (b) deceased cases (D), and (c) recovered cases (R). FEX demon￾strates superior data fitting during the training phase and accurate predictions in the testing pha…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 21 canonical work pages

  1. [1]

    Infectious diseases of humans: dynamics and control

    RM Anderson. Infectious diseases of humans: dynamics and control. Oxford University Press, 1991

  2. [2]

    Fractional-order seiqrdp model for simulating the dynamics of covid-19 epidemic

    Mohamed A Bahloul, Abderrazak Chahid, and Taous-Meriem Laleg-Kirati. Fractional-order seiqrdp model for simulating the dynamics of covid-19 epidemic. IEEE open journal of engi- neering in medicine and biology , 1:249–256, 2020

  3. [3]

    Mathematical models in epidemiol- ogy, volume 32

    Fred Brauer, Carlos Castillo-Chavez, Zhilan Feng, et al. Mathematical models in epidemiol- ogy, volume 32. Springer, 2019

  4. [4]

    Numerical analysis, brooks, 1997

    Richard L Burden and J Douglas Faires. Numerical analysis, brooks, 1997

  5. [5]

    Discovering symbolic models from deep learning with inductive biases

    Miles Cranmer, Alvaro Sanchez Gonzalez, Peter Battaglia, Rui Xu, Kyle Cranmer, David Spergel, and Shirley Ho. Discovering symbolic models from deep learning with inductive biases. Advances in neural information processing systems , 33:17429–17442, 2020

  6. [6]

    Greedy randomized adaptive search procedures

    Thomas A Feo and Mauricio GC Resende. Greedy randomized adaptive search procedures. Journal of global optimization , 6:109–133, 1995

  7. [7]

    Practical methods of optimization

    Roger Fletcher. Practical methods of optimization . John Wiley & Sons, 2000

  8. [8]

    Deep learning, 2016

    Ian Goodfellow. Deep learning, 2016

Show all 33 references
  1. [9]

    Mathematical models of infectious disease trans- mission

    Nicholas C Grassly and Christophe Fraser. Mathematical models of infectious disease trans- mission. Nature Reviews Microbiology, 6(6):477–487, 2008

  2. [10]

    The mathematics of infectious diseases

    Herbert W Hethcote. The mathematics of infectious diseases. SIAM review , 42(4):599–653, 2000

  3. [11]

    Long short-term memory

    S Hochreiter. Long short-term memory. Neural Computation MIT-Press , 1997

  4. [12]

    AttNS: Attention-inspired numerical solving for limited data scenarios

    Zhongzhan Huang, Mingfu Liang, Shanshan Zhong, and Liang Lin. AttNS: Attention-inspired numerical solving for limited data scenarios. In Forty-first International Conference on Ma- chine Learning, 2024

  5. [13]

    Finite expression methods for discovering physical laws from data

    Zhongyi Jiang, Chunmei Wang, and Haizhao Yang. Finite expression methods for discovering physical laws from data. arXiv preprint arXiv:2305.08342 , 2023

  6. [14]

    Covid- 19 data repository

    Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE). Covid- 19 data repository. Dataset retrieved from https://raw.githubusercontent.com/owid/ covid-19-data/master/public/data/jhu/full_data.csv via Our World in Data, 2024. Ac- cessed: August 21, 2024

  7. [15]

    Modeling infectious diseases in humans and animals

    Matt J Keeling and Pejman Rohani. Modeling infectious diseases in humans and animals . Princeton university press, 2008

  8. [16]

    A contribution to the mathematical theory of epidemics

    William Ogilvy Kermack and Anderson G McKendrick. A contribution to the mathematical theory of epidemics. Proceedings of the royal society of london. Series A, Containing papers of a mathematical and physical character , 115(772):700–721, 1927

  9. [17]

    Deep learning.nature, 521(7553):436–444, 2015

    Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. Deep learning.nature, 521(7553):436–444, 2015

  10. [18]

    Stiffness-aware neural network for learn- ing hamiltonian systems

    Senwei Liang, Zhongzhan Huang, and Hong Zhang. Stiffness-aware neural network for learn- ing hamiltonian systems. In International Conference on Learning Representations , 2022

  11. [19]

    Reproducing activation func- tion for deep learning

    Senwei Liang, Liyao Lyu, Chunmei Wang, and Haizhao Yang. Reproducing activation func- tion for deep learning. Communications in mathematical sciences , 22(2), 2024

  12. [20]

    Finite expression method for solving high-dimensional par- tial differential equations

    Senwei Liang and Haizhao Yang. Finite expression method for solving high-dimensional par- tial differential equations. arXiv preprint arXiv:2206.10121 , 2022

  13. [21]

    A training-free conditional diffusion model for learning stochastic dynamical systems

    Yanfang Liu, Yuan Chen, Dongbin Xiu, and Guannan Zhang. A training-free conditional diffusion model for learning stochastic dynamical systems. arXiv preprint arXiv:2410.03108 , 2024

  14. [22]

    Model reduction with memory and the machine learning of dynamical systems

    Chao Ma, Jianchun Wang, and E Weinan. Model reduction with memory and the machine learning of dynamical systems. Communications in Computational Physics , 25(4):947–962, 2019

  15. [23]

    Deep symbolic regression: Recovering mathematical expressions from data via risk-seeking policy gradients

    Brenden K Petersen, Mikel Landajuela, T Nathan Mundhenk, Claudio P Santiago, Soo K Kim, and Joanne T Kim. Deep symbolic regression: Recovering mathematical expressions from data via risk-seeking policy gradients. arXiv preprint arXiv:1912.04871 , 2019

  16. [24]

    Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

    Maziar Raissi, Paris Perdikaris, and George E Karniadakis. 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

  17. [25]

    An overview of gradient descent optimization algorithms

    Sebastian Ruder. An overview of gradient descent optimization algorithms. arXiv preprint arXiv:1609.04747, 2016. LEARNING EPIDEMIOLOGICAL DYNAMICS VIA THE FINITE EXPRESSION METHOD 13

  18. [26]

    Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

    Cynthia Rudin. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nature machine intelligence , 1(5):206–215, 2019

  19. [27]

    Distilling free-form natural laws from experimental data

    Michael Schmidt and Hod Lipson. Distilling free-form natural laws from experimental data. science, 324(5923):81–85, 2009

  20. [28]

    Deep euler method: solving odes by approx- imating the local truncation error of the euler method

    Xing Shen, Xiaoliang Cheng, and Kewei Liang. Deep euler method: solving odes by approx- imating the local truncation error of the euler method. arXiv preprint arXiv:2003.09573 , 2020

  21. [29]

    Finite expression method for learning dynamics on complex networks

    Zezheng Song, Chunmei Wang, and Haizhao Yang. Finite expression method for learning dynamics on complex networks. arXiv preprint arXiv:2401.03092 , 2024

  22. [30]

    Reinforcement learning: An introduction

    Richard S Sutton and Andrew G Barto. Reinforcement learning: An introduction. MIT press, 2018

  23. [31]

    Proceedings of the national academy of sciences

    G Teobaldi, Al Ma’Mari, M Rogers, S Alghamdi, T Moorsom, S Lee, T Prokscha, H Luetkens, M Valvidares, M Flokstra, et al. Proceedings of the national academy of sciences. Proceedings of the National Academy of Sciences of USA , 2017

  24. [32]

    A comparison of greedy search algo- rithms

    Christopher Wilt, Jordan Thayer, and Wheeler Ruml. A comparison of greedy search algo- rithms. In Proceedings of the International Symposium on Combinatorial Search , volume 1 of Lecture Notes in Computer Science , pages 129–136, 2010

  25. [33]

    Frequency principle: Fourier analysis sheds light on deep neural networks

    John Xu, Yaoyu Zhang, Tao Luo, Yanyang Xiao, and Zheng Ma. Frequency principle: Fourier analysis sheds light on deep neural networks. Communications in Computational Physics , 28(5):1746–1767, 2020. Department of Mathematics, University of Florida, Gainesville, FL 32611, USA. ...

Pith tools

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