pith. sign in

arxiv: 2509.09145 · v4 · submitted 2025-09-11 · 📡 eess.SY · cs.SY

KAN-Therm: A Lightweight Battery Thermal Model Using Kolmogorov-Arnold Network

Soumyoraj Mallick , Faysal Ahamed , Sanchita Ghosh , Tanushree Roy This is my paper

Pith reviewed 2026-05-18 18:33 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords battery thermal modelKolmogorov-Arnold networkcore temperature estimationbattery management systemlightweight neural networklithium-ion batteryreal-time estimation
0
0 comments X

The pith

Kolmogorov-Arnold networks estimate battery core temperature with lower memory use and faster run times than standard neural networks or tree models.

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

The paper sets out to show that a Kolmogorov-Arnold network can produce a compact thermal model capable of predicting the core temperature of a cylindrical lithium-ion cell in real time. Physical battery management systems face strict limits on memory and processor speed, which rules out both detailed physics simulations and many conventional neural networks. By replacing fixed activation functions with functions that the network itself learns, the model aims to represent the nonlinear heat flows using fewer parameters. If this holds, temperature tracking could move onto the same embedded hardware that already controls the battery, improving safety margins without extra cost.

Core claim

KAN-therm applies a Kolmogorov-Arnold network to estimate the core temperature of a cylindrical battery cell and reports lower memory overhead together with shorter estimation times than state-of-the-art neural-network and tree-based models, thereby demonstrating suitability for direct use on resource-constrained physical battery management systems.

What carries the argument

Kolmogorov-Arnold network whose edges carry learnable nonlinear activation functions, allowing the model to capture thermal complexity in a lean parameter set rather than through wide layers of fixed activations.

If this is right

  • Real-time core-temperature estimates become feasible inside the same microcontroller that runs cell balancing and state-of-charge calculations.
  • Continuous thermal oversight can be added to packs that previously could not afford the compute load of high-fidelity models.
  • Model size remains small enough to store and execute separate instances for each cell in a large pack without memory overflow.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same lean structure could be tested on prismatic or pouch cells to see whether the memory advantage persists across form factors.
  • Coupling the network outputs with a simple energy-balance equation might reduce the amount of labeled data required for training.
  • Deployment on actual BMS hardware under wide temperature ranges would reveal whether the reported speed and size gains hold outside the laboratory data set.

Load-bearing premise

The learnable nonlinear activation functions can represent the battery's thermal dynamics at the accuracy needed for safety use while keeping the overall model smaller and quicker than alternative approaches.

What would settle it

Run KAN-therm on a fresh set of measured voltage, current, and surface-temperature traces from a cylindrical cell under varied charge-discharge profiles and check whether core-temperature prediction error stays below a chosen safety threshold while memory footprint and inference latency remain lower than the compared neural-network and tree baselines.

Figures

Figures reproduced from arXiv: 2509.09145 by Faysal Ahamed, Sanchita Ghosh, Soumyoraj Mallick, Tanushree Roy.

Figure 1
Figure 1. Figure 1: Flowchart shows how the dataset is split into training, validation, and testing subsets. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Data distribution of current (I) vs core temperature (T1) [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Block diagram illustration of the proposed KAN-Therm model for prediction of battery core temperature [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Parallel coordinate plot illustrates the exploratory findings of the hyperparameter tuning process for our [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Figure shows the training and validation loss with epochs for the KAN-Thermal model. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Under CC charging, top plot shows the true core temperature [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Under UDDS charging, top plot shows the true core temperature [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
read the original abstract

A battery management system (BMS) relies on real-time estimation of battery temperature distribution in battery cells to ensure safe and optimal operation of Lithium-ion batteries. However, physical BMS often suffers from memory and computational resource limitations required by high-fidelity models. Temperature estimation of batteries for safety-critical systems using physics-based models on physical BMS can potentially become challenging due to their higher computational time. In contrast, neural network-based approaches offer faster estimation but require greater memory overhead. To address these challenges, we propose Kolmogorov-Arnold network (KAN) based thermal model, KAN-therm, to estimate the core temperature of a cylindrical battery. Unlike traditional neural network architectures, KAN uses learnable nonlinear activation functions that can effectively capture system complexity using relatively lean models. We have compared the memory overhead and estimation time of our model with state-of-the-art neural network and tree-based models to demonstrate the applicability and potential scalability of KAN-therm on a physical BMS.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 3 minor

Summary. The manuscript proposes KAN-Therm, a Kolmogorov-Arnold Network (KAN) based lightweight model to estimate the core temperature of cylindrical lithium-ion batteries. It argues that KAN's learnable nonlinear activation functions enable leaner models that capture thermal dynamics with lower memory overhead and faster estimation times than state-of-the-art neural network and tree-based approaches, addressing the resource constraints of physical battery management systems (BMS).

Significance. If the efficiency and accuracy claims are substantiated under conditions representative of embedded hardware, the work could provide a practical contribution to real-time thermal estimation in BMS, bridging the gap between computationally heavy physics-based models and memory-intensive neural networks. The application of KAN to battery thermal modeling is a novel angle that may encourage further exploration of spline-based networks in control and estimation tasks for energy systems.

major comments (1)
  1. [§4, Table 2] §4 (Experimental Results), Table 2: The reported memory overhead and estimation time comparisons with NN and tree-based baselines are central to the applicability claim for physical BMS, yet the manuscript does not specify whether these metrics were obtained on resource-constrained microcontrollers (limited RAM, no GPU, fixed-point arithmetic) or on desktop CPUs/GPUs. Without hardware-specific benchmarks matching the target deployment, the claimed advantage for overcoming BMS memory/compute limits remains unverified and load-bearing for the central contribution.
minor comments (3)
  1. [Abstract] Abstract: The summary of results would be strengthened by including at least one quantitative accuracy metric (e.g., RMSE or MAE on core temperature) alongside the efficiency claims.
  2. [§3] §3 (Model Description): The mapping from battery inputs (current, voltage, ambient temperature) to KAN inputs and the choice of grid size or spline order could be clarified to aid reproducibility.
  3. [Figure 3] Figure 3: The caption and axis labels for the temperature estimation plots should explicitly state the test conditions and error scale to improve clarity.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comment on the experimental evaluation. We address the major point below.

read point-by-point responses

  1. Referee: [§4, Table 2] §4 (Experimental Results), Table 2: The reported memory overhead and estimation time comparisons with NN and tree-based baselines are central to the applicability claim for physical BMS, yet the manuscript does not specify whether these metrics were obtained on resource-constrained microcontrollers (limited RAM, no GPU, fixed-point arithmetic) or on desktop CPUs/GPUs. Without hardware-specific benchmarks matching the target deployment, the claimed advantage for overcoming BMS memory/compute limits remains unverified and load-bearing for the central contribution.

    Authors: We thank the referee for highlighting this important clarification. The memory overhead and estimation time values in Table 2 were measured on a standard desktop CPU (Intel Core i7, Python 3.10 with NumPy/PyTorch, no GPU or fixed-point optimizations). We agree that the manuscript should explicitly state the evaluation platform. In the revised version we will add this detail to Section 4 and include a short discussion noting that, while the reported figures do not replicate a microcontroller environment, the relative reductions versus the NN and tree-based baselines on identical hardware still indicate lower resource demands. Direct embedded benchmarks on target BMS hardware would strengthen the deployment claim but lie outside the scope of the present study. revision: yes

Circularity Check

0 steps flagged

No circularity: direct empirical application of KAN to battery data

full rationale

The paper presents KAN-therm as a straightforward supervised application of the existing Kolmogorov-Arnold Network architecture to predict cylindrical battery core temperature from input features. All reported results consist of empirical comparisons of accuracy, memory overhead, and estimation time against external baselines (neural networks and tree-based models). No equations, parameters, or claims are shown to reduce by construction to self-fitted quantities, self-citations, or renamed inputs. The derivation chain is therefore self-contained through standard training and benchmarking on held-out data.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based solely on the abstract, the model rests on standard assumptions of neural network training and the representational capacity of KAN; no explicit free parameters, axioms, or invented entities are detailed.

pith-pipeline@v0.9.0 · 5709 in / 1052 out tokens · 38870 ms · 2026-05-18T18:33:49.913162+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A Practitioner's Guide to Kolmogorov-Arnold Networks

    cs.LG 2025-10 accept novelty 3.0

    A systematic review of Kolmogorov-Arnold Networks that maps their relation to Kolmogorov superposition theory, MLPs, and kernels, examines basis-function design choices, summarizes performance advances, and supplies a...

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages · cited by 1 Pith paper · 2 internal anchors

  1. [1]

    Renewable Energy Driven On-road Wireless Charging Infrastructure For Electric Vehicles In Smart Cities: A Prototype Design And Anal- ysis,

    A. Bukhari, O. I. Aboulola, A. ur Rehman, A. Alharbi, W. Alosaimi, and A. Daud, “Renewable Energy Driven On-road Wireless Charging Infrastructure For Electric Vehicles In Smart Cities: A Prototype Design And Anal- ysis,”Energy Reports, vol. 12, pp. 5145–5154, 2024

    work page 2024

  2. [2]

    Core Temperature Estimation Of Lithium-ion Batteries Using Long Short-term Memory (Lstm) Network And Kolmogorov-arnold Network (Kan),

    D. Karnehm, A. Samanta, C. Rosenm ¨uller, A. Neve, and S. Williamson, “Core Temperature Estimation Of Lithium-ion Batteries Using Long Short-term Memory (Lstm) Network And Kolmogorov-arnold Network (Kan),”IEEE Transactions on Transportation Electrification, 2025

    work page 2025

  3. [3]

    Core Temperature Estimation Based On Electro-thermal Model Of Lithium-ion Batteries,

    L. Chen, M. Hu, K. Cao, S. Li, Z. Su, G. Jin, and C. Fu, “Core Temperature Estimation Based On Electro-thermal Model Of Lithium-ion Batteries,”International Journal of energy research, vol. 44, no. 7, pp. 5320–5333, 2020

    work page 2020

  4. [4]

    Robust Estimation Of Battery System Tempera- ture Distribution Under Sparse Sensing And Uncertainty,

    X. Lin, H. E. Perez, J. B. Siegel, and A. G. Stefanopoulou, “Robust Estimation Of Battery System Tempera- ture Distribution Under Sparse Sensing And Uncertainty,”IEEE Transactions on Control Systems Technology, vol. 28, no. 3, pp. 753–765, 2019

    work page 2019

  5. [5]

    Real-time Sen- sorless Temperature Estimation Of Lithium-ion Batteries Based On Online Operando Impedance Acquisition,

    Y . Zheng, Y . Che, J. Guo, N. A. Weinreich, A. Kulkarni, A. Nadeem, X. Sui, and R. Teodorescu, “Real-time Sen- sorless Temperature Estimation Of Lithium-ion Batteries Based On Online Operando Impedance Acquisition,” IEEE Transactions on Power Electronics, vol. 39, no. 10, pp. 13853–13868, 2024

    work page 2024

  6. [6]

    Detection And Isolation Of Battery Charging Cyberattacks Via Koopman Operator,

    S. Ghosh and T. Roy, “Detection And Isolation Of Battery Charging Cyberattacks Via Koopman Operator,” Available at SSRN 5028845, 2024

    work page 2024

  7. [7]

    Benchmarking Core Temperature Forecasting For Lithium-ion Battery Using Typical Recurrent Neural Networks,

    X. Zhang, H. Xiang, X. Xiong, Y . Wang, and Z. Chen, “Benchmarking Core Temperature Forecasting For Lithium-ion Battery Using Typical Recurrent Neural Networks,”Applied Thermal Engineering, vol. 248, p. 123257, 2024

    work page 2024

  8. [8]

    Data-driven Analysis On Thermal Effects And Temperature Changes Of Lithium-ion Battery,

    S. Zhu, C. He, N. Zhao, and J. Sha, “Data-driven Analysis On Thermal Effects And Temperature Changes Of Lithium-ion Battery,”Journal of Power Sources, vol. 482, p. 228983, 2021

    work page 2021

  9. [9]

    A Data-driven Method For Predicting Thermal Runaway Propagation Of Battery Modules Considering Uncertain Conditions,

    N. Ouyang, W. Zhang, X. Yin, X. Li, Y . Xie, H. He, and Z. Long, “A Data-driven Method For Predicting Thermal Runaway Propagation Of Battery Modules Considering Uncertain Conditions,”Energy, vol. 273, p. 127168, 2023

    work page 2023

  10. [10]

    A Data-driven Approach To Estimate Battery Cell Temperature Using A Nonlinear Autoregressive Exogenous Neural Network Model,

    M. M. Hasan, S. A. Pourmousavi, A. J. Ardakani, and T. K. Saha, “A Data-driven Approach To Estimate Battery Cell Temperature Using A Nonlinear Autoregressive Exogenous Neural Network Model,”Journal of Energy Storage, vol. 32, p. 101879, 2020

    work page 2020

  11. [11]

    Core Temperature Estimation Method For Lithium-ion Battery Based On Long Short-term Memory Model With Transfer Learning,

    N. Wang, G. Zhao, Y . Kang, W. Wang, A. Chen, B. Duan, and C. Zhang, “Core Temperature Estimation Method For Lithium-ion Battery Based On Long Short-term Memory Model With Transfer Learning,”IEEE Journal of Emerging and Selected Topics in Power Electronics, vol. 11, no. 1, pp. 201–213, 2021

    work page 2021

  12. [12]

    Hybrid Electrical Circuit Model And Deep Learning-based Core Temperature Estimation Of Lithium-ion Battery Cell,

    S. Surya, A. Samanta, V . Marcis, and S. Williamson, “Hybrid Electrical Circuit Model And Deep Learning-based Core Temperature Estimation Of Lithium-ion Battery Cell,”IEEE Transactions on Transportation Electrifica- tion, vol. 8, no. 3, pp. 3816–3824, 2022

    work page 2022

  13. [13]

    Co-estimation Of Lithium-ion Battery State Of Charge And State Of Temperature Based On A Hybrid Electrochemical-thermal-neural-network Model,

    F. Feng, S. Teng, K. Liu, J. Xie, Y . Xie, B. Liu, and K. Li, “Co-estimation Of Lithium-ion Battery State Of Charge And State Of Temperature Based On A Hybrid Electrochemical-thermal-neural-network Model,”Journal of Power Sources, vol. 455, p. 227935, 2020

    work page 2020

  14. [14]

    An Online Hybrid Estimation Method For Core Tem- perature Of Lithium-ion Battery With Model Noise Compensation,

    Y . Liu, Z. Huang, Y . Wu, L. Yan, F. Jiang, and J. Peng, “An Online Hybrid Estimation Method For Core Tem- perature Of Lithium-ion Battery With Model Noise Compensation,”Applied Energy, vol. 327, p. 120037, 2022

    work page 2022

  15. [15]

    Neural Network Based Computational Model For Estimation Of Heat Gener- ation In LiFePO4 Pouch Cells Of Different Nominal Capacities,

    S. Arora, W. Shen, and A. Kapoor, “Neural Network Based Computational Model For Estimation Of Heat Gener- ation In LiFePO4 Pouch Cells Of Different Nominal Capacities,”Computers & Chemical Engineering, vol. 101, pp. 81–94, 2017

    work page 2017

  16. [16]

    A CNN-ABC Model For Estimation And Optimization Of Heat Gen- eration Rate And V oltage Distributions Of Lithium-ion Batteries For Electric Vehicles,

    S. Yalc ¸ın, S. Panchal, and M. S. Herdem, “A CNN-ABC Model For Estimation And Optimization Of Heat Gen- eration Rate And V oltage Distributions Of Lithium-ion Batteries For Electric Vehicles,”International Journal of Heat and Mass Transfer, vol. 199, p. 123486, 2022

    work page 2022

  17. [17]

    Core Temperature Estimation Of Lithium-ion Battery Based On Numerical Model Fusion Deep Learning,

    A. Yuan, T. Cai, H. Luo, Z. Song, and B. Wei, “Core Temperature Estimation Of Lithium-ion Battery Based On Numerical Model Fusion Deep Learning,”Journal of Energy Storage, vol. 102, p. 114148, 2024

    work page 2024

  18. [18]

    Capacity Degradation Prediction Of On-road Vehicle Battery Packs By Combining Kolmogorov-Arnold With Squeeze-and-excitation Networks,

    Y . Yang, Z. Liu, and B. Han, “Capacity Degradation Prediction Of On-road Vehicle Battery Packs By Combining Kolmogorov-Arnold With Squeeze-and-excitation Networks,”Measurement Science and Technology, 2025

    work page 2025

  19. [19]

    Battery State Of Charge Estimation For Electric Vehicle Using Kolmogorov-Arnold Networks,

    M. H. Sulaiman, Z. Mustaffa, A. I. Mohamed, A. S. Samsudin, and M. I. M. Rashid, “Battery State Of Charge Estimation For Electric Vehicle Using Kolmogorov-Arnold Networks,”Energy, vol. 311, p. 133417, 2024

    work page 2024

  20. [20]

    Enhanced State Of Health Prediction For Lithium-ion Batteries Using A Hybrid Convolutional–Kolmogorov–Arnold Network,

    Y . Cui and Y . Feng, “Enhanced State Of Health Prediction For Lithium-ion Batteries Using A Hybrid Convolutional–Kolmogorov–Arnold Network,”International Journal of Electrochemical Science, vol. 20, no. 6, p. 101008, 2025

    work page 2025

  21. [21]

    SOH Estimation Of Lithium-ion Batteries Subject To Partly Missing Data: A Kolmogorov–Arnold–Linformer Model,

    L. Shao, Y . Zhang, X. Zheng, R. Yang, and W. Zhou, “SOH Estimation Of Lithium-ion Batteries Subject To Partly Missing Data: A Kolmogorov–Arnold–Linformer Model,”Neurocomputing, vol. 638, p. 130181, 2025

    work page 2025

  22. [22]

    SOH Prediction Of Battery Packs Using Dynamic Graph Convolution Combined With KAN-Driven Methods,

    H. Liu, Y . Wang, J. Wang, A. Zhang, and H. Yang, “SOH Prediction Of Battery Packs Using Dynamic Graph Convolution Combined With KAN-Driven Methods,” in2025 4th International Conference on Green Energy and Power Systems (ICGEPS), pp. 129–134, IEEE, 2025

    work page 2025

  23. [23]

    Remaining Useful Life Prediction Of Lithium-ion Battery Based On Improved Gated Recurrent Unit-generalized Cauchy Process,

    J. He, Z. Ma, Y . Liu, C. Ma, and W. Gao, “Remaining Useful Life Prediction Of Lithium-ion Battery Based On Improved Gated Recurrent Unit-generalized Cauchy Process,”Journal of Energy Storage, vol. 126, p. 117086, 2025

    work page 2025

  24. [24]

    KAN: Kolmogorov-Arnold Networks

    Z. Liu, Y . Wang, S. Vaidya, F. Ruehle, J. Halverson, M. Soljaˇci´c, T. Y . Hou, and M. Tegmark, “Kan: Kolmogorov- arnold Networks,”arXiv preprint arXiv:2404.19756, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  25. [25]

    BMS: Bandwidth-aware Multi-interface Scheduling For Energy-efficient And Delay-constrained Gateway-to-device Communications In IoT,

    H. Chen, H. Qin, W. Chen, N. Li, T. Wang, J. He, G. Yang, and Y . Peng, “BMS: Bandwidth-aware Multi-interface Scheduling For Energy-efficient And Delay-constrained Gateway-to-device Communications In IoT,”Computer Networks, vol. 225, p. 109645, 2023

    work page 2023

  26. [26]

    A Lumped- parameter Electro-thermal Model For Cylindrical Batteries,

    X. Lin, H. E. Perez, S. Mohan, J. B. Siegel, A. G. Stefanopoulou, Y . Ding, and M. P. Castanier, “A Lumped- parameter Electro-thermal Model For Cylindrical Batteries,”Journal of Power Sources, vol. 257, pp. 1–11, 2014

    work page 2014

  27. [27]

    Thermal Fault-tolerance In Lithium-ion Battery Cells: A Barrier Function Based Input-to-state Safety Framework,

    S. D. Vyas, T. Roy, and S. Dey, “Thermal Fault-tolerance In Lithium-ion Battery Cells: A Barrier Function Based Input-to-state Safety Framework,” in2022 IEEE Conference on Control Technology and Applications (CCTA), pp. 1178–1183, IEEE, 2022

    work page 2022

  28. [28]

    A General Energy Balance For Battery Systems,

    D. Bernardi, E. Pawlikowski, and J. Newman, “A General Energy Balance For Battery Systems,”Journal of the electrochemical society, vol. 132, no. 1, p. 5, 1985

    work page 1985

  29. [29]

    Real-time Capacity Estimation Of Lithium-ion Batteries Uti- lizing Thermal Dynamics,

    D. Zhang, S. Dey, H. E. Perez, and S. J. Moura, “Real-time Capacity Estimation Of Lithium-ion Batteries Uti- lizing Thermal Dynamics,”IEEE Transactions on Control Systems Technology, vol. 28, no. 3, pp. 992–1000, 2019

    work page 2019

  30. [30]

    Safer Batteries Via Active Fault Tolerant Control,

    S. Dey, Y . Shi, K. Smith, and M. Khanra, “Safer Batteries Via Active Fault Tolerant Control,” in2019 American Control Conference (ACC), pp. 1561–1566, IEEE, 2019

    work page 2019

  31. [31]

    Development Of The Federal Urban Driving Schedule,

    R. E. Kruse and T. A. Huls, “Development Of The Federal Urban Driving Schedule,”SAE Technical Paper, 1973

    work page 1973

  32. [32]

    A Physics-Inspired Machine Learning Nonlinear Model Of Li-ion Batteries.,

    O. Ahmadzadeh, R. Rodriguez, Y . Wang, and D. Soudbakhsh, “A Physics-Inspired Machine Learning Nonlinear Model Of Li-ion Batteries.,” inACC, vol. 23, pp. 3087–3092, 2023

    work page 2023

  33. [33]

    Automotive li-ion cell usage data set,

    IEEE Dataport, “Automotive li-ion cell usage data set,” 2018

    work page 2018

  34. [34]

    Easy Hyperparameter Search Using Optunity

    M. Claesen, J. Simm, D. Popovic, Y . Moreau, and B. De Moor, “Easy Hyperparameter Search Using Optunity,” arXiv preprint arXiv:1412.1114, 2014

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  35. [35]

    Vanishing Gradient Mitigation With Deep Learning Neural Network Optimization,

    H. H. Tan and K. H. Lim, “Vanishing Gradient Mitigation With Deep Learning Neural Network Optimization,” in2019 7th international conference on smart computing & communications (ICSCC), pp. 1–4, IEEE, 2019

    work page 2019