Determining the acceleration field of a rigid body using three accelerometers and one gyroscope, with applications in mild traumatic brain injury

Benjamin E. Grossman-Ponemona; Haneesh Kesari; Yang Wan

Determining the acceleration field of a rigid body using three accelerometers and one gyroscope, with applications in mild traumatic brain injury

T0 review · 3 major / 0 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The abstract claims a linear algorithm for reconstructing the full acceleration field of a rigid body from three accelerometers and one gyroscope, while the accompanying full text is an unrelated graph-neural-network paper on multi-omics di

desk verdict The abstract and full text are two different papers; the claimed acceleration-field algorithm appears nowhere in the manuscript. read the letter →

arxiv 2508.07464 v1 pith:UT2HWELS submitted 2025-08-10 physics.app-ph

Yang Wan , Benjamin E. Grossman-Ponemona , Haneesh Kesari This is my paper · ORCID

classification physics.app-ph

keywords rigidbodykinematicsaccelerationfieldreconstructiontri-axialaccelerometersgyroscopeheadimpactmeasurementmildtraumaticbraininjurylinearequations

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

The abstract of this submission states a method for reconstructing the full acceleration field of a rigid body from three tri-axial accelerometers and one tri-axial gyroscope, using linear equations derived from rigid body kinematics rather than numerically differentiating noisy angular velocity. The stated motivation is improved head-motion measurement for mild traumatic brain injury, with validation claimed in controlled soccer heading experiments. However, the full text supplied with the submission is a different paper on a tree-generated graph neural network for multi-omics disease classification; it contains none of the algorithm, its derivation, or the experiments the abstract describes. The intended pith, if the abstract is taken as the author's claim, is that three non-collinear accelerometers plus one gyroscope suffice for linear, differentiation-free reconstruction of rigid-body acceleration at unsensed points.

What carries the argument

The load-bearing identity is the rigid-body kinematics formula for the acceleration $\mathbf{a}_i$ at a point with position $\mathbf{r}_i$ relative to a reference point: $\mathbf{a}_i = \mathbf{a}_0 + \dot{\boldsymbol{\omega}} \times \mathbf{r}_i + \boldsymbol{\omega} \times (\boldsymbol{\omega} \times \mathbf{r}_i)$, where $\mathbf{a}_0$ is the reference-point acceleration, $\boldsymbol{\omega}$ is the angular velocity, and $\dot{\boldsymbol{\omega}}$ is the angular acceleration. The gyroscope supplies $\boldsymbol{\omega}$, so with three non-collinear accelerometers the unknowns $\mathbf{a}_0$ and $\dot{\boldsymbol{\omega}}$ enter linearly and can be solved for directly. The claimed trick

What would settle it

A concrete check: simulate a rigid body with prescribed motion, generate noisy measurements from three non-collinear tri-axial accelerometers and one tri-axial gyroscope, and test whether solving the linear system recovers the true angular and translational acceleration at unsensed points without differentiating the gyroscope signal. In parallel, open the submitted full text and look for the derivation of these linear equations and the soccer-heading experimental section; if the body instead presents a multi-omics graph neural network, the claims cannot be verified from this submission.

Watch

Extended reading notes

Core claim

The central claim, as stated in the abstract, is that the full acceleration field of a rigid body can be reconstructed from three tri-axial accelerometers (with the only placement constraint being non-collinearity) and one tri-axial gyroscope, by solving linear equations from rigid body kinematics. The claimed advantage over existing approaches is that it avoids both numerical differentiation of noisy gyroscope angular velocity and the restrictive sensor layouts or nonlinear optimization associated with gyroscope-free methods. The motivation is accurate measurement of head motion for motion- and deformation-based injury criteria in mild traumatic brain injury, and the abstract asserts accura

Load-bearing premise

The abstract's claims stand on the assumption that the submitted full text actually contains the rigid-body derivation and the soccer-heading validation; the full text supplied here is a different paper on multi-omics graph neural networks, so that support is absent.

Editorial extensions

If this is right

Head-impact sensor systems could estimate acceleration at any skull location from a small cluster of three accelerometers and one gyroscope, without noise-amplifying differentiation of angular velocity.
The linear formulation would permit real-time or on-device computation, making wearable mTBI monitors and sideline screening tools more practical.
The only placement constraint—non-collinearity—is mild, so sensors could be distributed flexibly around a helmet or headguard rather than locked into orthogonal triads.
If the claimed soccer-heading validation holds, it would demonstrate the method transfers from laboratory calibration to realistic sports impacts at unsensed sites.

Reading between the lines

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

If the body-text mismatch is a posting error, the two components should be evaluated separately: the abstract's linear-reconstruction algorithm deserves a derivational check, and the graph-neural-network manuscript in the body deserves its own assessment; neither can be judged from this composite document.
A direct numerical stress test of the abstract's claim would be to simulate a rigid body with known motion, add realistic sensor noise, and compare the linear solution against a differentiation-based gyroscope method at unsensed points; a clean win would isolate differentiation avoidance as the active ingredient.
The same rigid-body kinematic identity is generic, so if the algorithm is sound it would transfer to other rigid-body settings—robot link motion, vehicle crash dummies, or instrumented equipment—wherever three non-collinear accelerometers and a gyroscope can be mounted.
Because deformation-based mTBI criteria need strain or strain-rate fields rather than raw acceleration, a natural extension would couple the reconstructed acceleration field to a head finite-element model; the paper does not address this step.

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.

Desk Editor's Note

read the letter

Colleague,

This one is straightforward: the abstract and the full text are two different papers. The abstract promises an algorithm for reconstructing the acceleration field of a rigid head from three accelerometers and one gyroscope, with validation in soccer heading experiments. The body of the manuscript is MOTGNN, a graph neural network paper for multi-omics disease classification. Searching the full text, there is no accelerometer, gyroscope, rigid body kinematics, angular acceleration, soccer heading, or anything related. So the central claim is not merely weak; it is absent.

What the manuscript does contain is a reasonably detailed GNN method paper. It proposes XGBoost-based graph construction per omics modality, modality-specific GNNs, and a feedforward integration network. It reports experiments on three disease datasets and claims 5-10% improvements over baselines, with some robustness and interpretability analysis. On its own terms, that may be a legitimate paper. But it is not the paper described by the title and abstract, and I cannot assess the claimed sensor-fusion result at all. There is no derivation, no equations, no experiments for the head-impact algorithm. The abstract's statement about solving linear equations is unsupported. There is also no comparison to prior work on gyroscope-free methods or gyroscope differentiation, despite the abstract implying those are limitations the algorithm avoids.

The only charitable reading is that the wrong file was uploaded—perhaps this is a submission mix-up. Even under that reading, the current manuscript is incoherent: a referee cannot evaluate whether the claimed algorithm works, whether it is novel, or whether it is correctly derived. The condition for the claim to hold—that the body contains the algorithm and validation—is false. This is not a subtle mathematical gap; it is a total absence of the subject matter.

My call: desk reject, but invite the authors to resubmit the correct manuscript if this was an error. The GNN paper deserves its own venue, not to be appended to an unrelated title. No one should spend referee time on a paper whose abstract and body disagree at every level.

As a piece of submitted research, this fails all axes. If the correct manuscript exists elsewhere, I would be open to looking at it, but the current version is not a research preprint in any usable sense.

Referee Report

3 major / 0 minor

Summary. The submission, labeled arXiv:2508.07464 (physics.app-ph), carries the title 'Determining the acceleration field of a rigid body using three accelerometers and one gyroscope, with applications in mild traumatic brain injury.' The abstract promises an algorithm that reconstructs the full acceleration field of a rigid body from three tri-axial accelerometers and one tri-axial gyroscope, using a linear system derived from rigid-body kinematics, with the only constraint that the accelerometers be non-collinear, and with validation in controlled soccer heading experiments. The supplied full text, however, is a different paper: 'MOTGNN: Interpretable Graph Neural Networks for Multi-Omics Disease Classification.' The body contains no accelerometer, gyroscope, rigid-body kinematic analysis, angular acceleration, soccer heading, or any related formulation. There are no equations defining the proposed linear system, no experimental data, and no validation. The central claim of the title and abstract is therefore completely absent from the manuscript.

Significance. If the claimed algorithm were present and correct, it could be a practically useful contribution to rigid-body motion reconstruction and head-impact biomechanics, especially because it would avoid differentiation of noisy angular velocity signals and would allow flexible sensor placement. The 'non-collinear' condition and the reported validation in soccer heading experiments would be valuable, falsifiable claims. However, none of this content appears in the submitted full text. The actual full text is a graph neural network paper on multi-omics disease classification. That paper may have merit in its own field, but it is not the submitted paper's claimed topic and does not provide any support for the abstract's promises. The manuscript therefore cannot be evaluated as a research contribution to applied physics or head-impact measurement.

major comments (3)

[Abstract vs. full text] The abstract states: 'we present an algorithm for reconstructing the full acceleration field of a rigid body from measurements obtained by three tri-axial accelerometers and one tri-axial gyroscope... We validated the algorithm in controlled soccer heading experiments.' The supplied full text is 'MOTGNN: Interpretable Graph Neural Networks for Multi-Omics Disease Classification.' A search of the full text finds no occurrences of accelerometer, gyroscope, rigid body, angular acceleration, soccer heading, or any equivalent term. The claimed derivation and validation are entirely absent, so the central claim of the paper is unsupported.
[Full text: no derivation] The manuscript contains no equations—or even narrative—describing the rigid-body kinematics that would relate the three accelerometer readings and the gyroscope reading to translational acceleration, angular acceleration, and the acceleration field. Consequently, there is no way to check whether the proposed method is linear, whether the non-collinearity condition is sufficient, or what the observability and noise properties of the system are. The claim that the algorithm 'recovers angular acceleration and translational acceleration by solving a set of linear equations' is unverifiable from the submitted text.
[Full text: no validation] The abstract reports 'controlled soccer heading experiments' with 'accurate prediction of accelerations at unsensed locations across trials.' The full text does not contain these experiments, any sensor data, any error metrics, any comparison baselines, or any trial descriptions. Empirical validation is completely missing. Since the abstract presents the validation as part of the contribution, this is not a minor omission; it removes the evidential basis for the paper's primary claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity can be identified: the submitted full text does not contain the claimed derivation at all, so there is no derivation chain to reduce to its inputs.

full rationale

The abstract claims an algorithm for reconstructing a rigid body's acceleration field from three tri-axial accelerometers and one tri-axial gyroscope, with a linear-solution derivation and validation in soccer heading experiments. However, the supplied full text is an entirely different manuscript, 'MOTGNN: Interpretable Graph Neural Networks for Multi-Omics Disease Classification', by different authors, with no equations, derivations, sensor models, or experiments related to rigid body kinematics. Hard rule 1 requires quoting a specific reduction or fit that constitutes circularity; here there is no derivational content to examine, and absence of a derivation is not itself a circular step. The MOTGNN body may itself be non-circular, but it is irrelevant to the claimed subject. Therefore, on the circularity axis, the honest finding is a non-finding: score 0, with no circular steps identified. The central claim is unsupported and unverifiable from the submitted text, but that is a completeness/correctness problem, not a circularity problem.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

The central claim rests entirely on the abstract; the full text provides no support. The only axioms we can list are the implicit rigid-body assumption and the unmet assumption that the claimed derivation and experiments exist.

assumptions (2)

domain assumption The head is treated as a rigid body during impact
The abstract explicitly models a rigid body; human skull and brain tissue deform, and the manuscript does not justify this approximation.
ad hoc to paper The manuscript contains the derivation and the experimental validation described in the abstract
The full text is a different paper, so this axiom is not satisfied. The claim rests on an unprovided derivation and unprovided experiments.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Determining the acceleration field of a rigid body using three accelerometers and one gyroscope, with applications in mild traumatic brain injury." pith.science (2026). https://pith.science/paper/UT2HWELS

@misc{pith2026250807464,
  author       = {Pith},
  title        = {Pith review of: Determining the acceleration field of a rigid body using three accelerometers and one gyroscope, with applications in mild traumatic brain injury},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UT2HWELS}},
  note         = {Machine review of arXiv:2508.07464}
}

read the original abstract

Mild traumatic brain injury (mTBI) often results from violent head motion or impact. Most prevention strategies explicitly or implicitly rely on motion- or deformation-based injury criteria, both of which require accurate measurements of head motion. We present an algorithm for reconstructing the full acceleration field of a rigid body from measurements obtained by three tri-axial accelerometers and one tri-axial gyroscope. Unlike traditional gyroscope-based methods, which require numerically differentiating noisy angular velocity data, or gyroscope-free methods, which may impose restrictive sensor placement or involve nonlinear optimization, the proposed algorithm recovers angular acceleration and translational acceleration by solving a set of linear equations derived from rigid body kinematics. In the proposed method, the only constraint on sensor placement is that the accelerometers must be non-collinear. We validated the algorithm in controlled soccer heading experiments, demonstrating accurate prediction of accelerations at unsensed locations across trials. The proposed algorithm provides a robust, flexible, and efficient tool for reconstructing rigid body motion, with direct applications in contact sports, robotics, and biomechanical injury prediction.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

Field evaluation of a wearable instrumented headband designed for measuring head kinematics
eess.SP 2025-09 conditional novelty 4.0 of 10

On-field test of a five-sensor headband against an instrumented mouthpiece finds good time-history agreement for angular velocity and translational acceleration, weaker agreement for angular acceleration, and a 40.9% ...

Reference graph

Works this paper leans on

47 extracted references · 44 canonical work pages · cited by 1 Pith paper

[3]

Jenna L Ballard, Zexuan Wang, Wenrui Li, Li Shen, and Qi Long. 2024. Deep learning-based approaches for multi-omics data integration and analysis.�� 17, 1 (2024), 38

work page 2024
[4]

Leo Breiman. 2001. Random forests.�� 45, 1 (2001), 5–32

work page 2001
[5]

Michael M Bronstein, Joan Bruna, Yann LeCun, Arthur Szlam, and Pierre Van- dergheynst. 2017. Geometric deep learning: going beyond euclidean data.�� 34, 4 (2017), 18–42

work page 2017
[6]

Tianqi Chen and Carlos Guestrin. 2016. Xgboost: A scalable tree boosting system. In�� . 785–794

work page 2016
[7]

Zhiqian Chen, Fanglan Chen, Lei Zhang, Taoran Ji, Kaiqun Fu, Liang Zhao, Feng Chen, Lingfei Wu, Charu Aggarwal, and Chang-Tien Lu. 2023. Bridging the gap between spatial and spectral domains: A unified framework for graph neural networks.�� 56, 5 (2023), 1–42

work page 2023
[8]

Bernd Frank, Michael Hoffmeister, Norman Klopp, Thomas Illig, Jenny Chang- Claude, and Hermann Brenner. 2010. Single nucleotide polymorphisms in Wnt signaling and cell death pathway genes and susceptibility to colorectal cancer. ��31, 8 (2010), 1381–1386

work page 2010
[9]

Justin Gilmer, Samuel S Schoenholz, Patrick F Riley, Oriol Vinyals, and George E Dahl. 2017. Neural message passing for quantum chemistry. In�� . PMLR, 1263–1272

work page 2017
[10]

2016.��

Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016.�� . MIT press

work page 2016

Show all 47 references

[11]

Will Hamilton, Zhitao Ying, and Jure Leskovec. 2017. Inductive representation learning on large graphs.�� 30 (2017)

2017
[12]

Leroy Hood and Lee Rowen. 2013. The Human Genome Project: big science transforms biology and medicine.�� 5 (2013), 1–8

2013
[13]

Dan Huang, Bin Yu, Yun Deng, Weiqi Sheng, Zhilei Peng, Wenxin Qin, and Xiang Du. 2010. SFRP4 was overexpressed in colorectal carcinoma.�� 136 (2010), 395–401

2010
[14]

Idaho C3+3 Collaboration. 2022. Falcon: High Performance Supercomputer. https://doi.org/10.7923/falcon.id. Accessed: 2025-06-05

2022 doi
[15]

Wei Jiang, Weicai Ye, Xiaoming Tan, and Yun-Juan Bao. 2025. Network-based multi-omics integrative analysis methods in drug discovery: a systematic review. �� 18, 1 (2025), 27

2025
[16]

Ruth Johnson, Michelle M Li, Ayush Noori, Owen Queen, and Marinka Zitnik
[17]

Ziynet Nesibe Kesimoglu and Serdar Bozdag. 2023. SUPREME: multiomics data in- tegration using graph convolutional networks.�� 5, 2 (2023), lqad063

2023
[18]

Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic opti- mization.�� (2014)

2014
[19]

Thomas N Kipf and Max Welling. 2017. Semi-Supervised Classification with Graph Convolutional Networks. In��

2017
[20]

Yunchuan Kong and Tianwei Yu. 2018. A graph-embedded deep feedforward network for disease outcome classification and feature selection using gene expression data.��34, 21 (2018), 3727–3737

2018
[21]

Yunchuan Kong and Tianwei Yu. 2020. forgeNet: a graph deep neural network model using tree-based ensemble classifiers for feature graph construction.�� 36, 11 (2020), 3507–3515

2020
[22]

Yann Lecun, Yoshua Bengio, and Geoffrey Hinton. 2015. Deep learning.�� 521, 7553 (2015), 436–444. doi:10.1038/nature14539

2015 doi
[23]

Xiao Li, Jie Ma, Ling Leng, Mingfei Han, Mansheng Li, Fuchu He, and Yunping Zhu. 2022. MoGCN: a multi-omics integration method based on graph convo- lutional network for cancer subtype analysis.�� 13 (2022), 806842

2022
[24]

Yuting Liu, Jun Yu, Yang Xie, Mengying Li, Feng Wang, Jing Zhang, and Jian Qi
[25]

Yuxing Lu, Rui Peng, Lingkai Dong, Kun Xia, Renjie Wu, Shuai Xu, and Jinzhuo Wang. 2023. Multiomics dynamic learning enables personalized diagnosis and prognosis for pancancer and cancer subtypes.�� 24, 6 (11 2023). doi:10.1093/bib/bbad378

2023 doi
[26]

Vinod Nair and Geoffrey E Hinton. 2010. Rectified linear units improve restricted boltzmann machines. In��

2010
[27]

black box

Julian D Olden and Donald A Jackson. 2002. Illuminating the “black box”: a randomization approach for understanding variable contributions in artificial neural networks.�� 154, 1-2 (2002), 135–150

2002
[28]

Showmick Guha Paul, Arpa Saha, Md Zahid Hasan, Sheak Rashed Haider Noori, and Ahmed Moustafa. 2024. A systematic review of graph neural network in healthcare-based applications: Recent advances, trends, and future directions. �� 12 (2024), 15145–15170

2024
[29]

Veličković Petar, Cucurull Guillem, Casanova Arantxa, Romero Adriana, Lio Pietro, and B Yoshua. 2018. Graph attention networks. In�� , Vol. 8

2018
[30]

Indhupriya Subramanian, Srikant Verma, Shiva Kumar, Abhay Jere, and Kris- hanpal Anamika. 2020. Multi-omics data integration, interpretation, and its application.�� 14 (2020), 1177932219899051

2020
[31]

Raihanul Bari Tanvir, Md Mezbahul Islam, Masrur Sobhan, Dongsheng Luo, and Ananda Mohan Mondal. 2024. MOGAT: a multi-omics integration framework using graph attention networks for cancer subtype prediction.�� 25, 5 (2024), 2788

2024
[32]

Nektarios A Valous, Ferdinand Popp, Inka Zörnig, Dirk Jäger, and Pornpimol Charoentong. 2024. Graph machine learning for integrated multi-omics analysis. �� 131, 2 (2024), 205–211

2024
[33]

Tongxin Wang, Wei Shao, Zhi Huang, Haixu Tang, Jie Zhang, Zhengming Ding, and Kun Huang. 2021. MOGONET integrates multi-omics data using graph con- volutional networks allowing patient classification and biomarker identification. �� 12, 1 (2021), 3445

2021
[34]

Asim Waqas, Aakash Tripathi, Ravi P Ramachandran, Paul A Stewart, and Ghulam Rasool. 2024. Multimodal data integration for oncology in the era of deep neural networks: a review.�� 7 (2024), 1408843

2024
[35]

John N Weinstein, Eric A Collisson, Gordon B Mills, Kenna R Shaw, Brad A Ozenberger, Kyle Ellrott, Ilya Shmulevich, Chris Sander, and Joshua M Stuart
[36]

Jiecheng Wu, Zhaoliang Chen, Shunxin Xiao, Genggeng Liu, Wenjie Wu, and Shiping Wang. 2024. DeepMoIC: multi-omics data integration via deep graph convolutional networks for cancer subtype classification.�� 25, 1 (2024), 1–13

2024
[37]

Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and Philip S Yu. 2020. A comprehensive survey on graph neural networks.�� 32, 1 (2020), 4–24

2020
[38]

Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. 2019. How powerful are graph neural networks?. In��

2019
[39]

Pengcheng Yu, Weiyang He, Yanqiang Zhang, Can Hu, Yue Wu, Yi Wang, Zhehan Bao, Yuhang Xia, Ruolan Zhang, Mengxuan Cao, et al. 2022. Sfrp4 is a potential biomarker for the prognosis and immunotherapy for gastric cancer.�� 2022, 1 (2022), 8829649

2022
[40]

Shaza B Zaghlool and Omneya Attallah. 2022. A review of deep learning methods for multi-omics integration in precision medicine. In�� . IEEE, 2208–2215

2022
[41]

Xiao-Meng Zhang, Li Liang, Lin Liu, and Ming-Jing Tang. 2021. Graph neural networks and their current applications in bioinformatics.�� 12 (2021), 690049

2021
[42]

Ziwei Zhang, Peng Cui, and Wenwu Zhu. 2020. Deep learning on graphs: A survey. �� 34, 1 (2020), 249–270

2020
[43]

Chen Zhao, Anqi Liu, Xiao Zhang, Xuewei Cao, Zhengming Ding, Qiuying Sha, Hui Shen, Hong Wen Deng, and Weihua Zhou. 2024. CLCLSA: Cross-omics linked embedding with contrastive learning and self attention for integration with incomplete multi-omics data.�� ...

2024
[44]

Yating Zhong, Yuzhong Peng, Yanmei Lin, Dingjia Chen, Hao Zhang, Wen Zheng, Yuanyuan Chen, and Changliang Wu. 2023. MODILM: towards better complex diseases classification using a novel multi-omics data integration learning model. �� 23, ...

2023
[45]

Zhiqiang Zhong, Anastasia Barkova, and Davide Mottin. 2025. Knowledge- augmented Graph Machine Learning for Drug Discovery: A Survey.�� (2025). https://doi.org/10.1145/3744237

2025 doi
[46]

Jie Zhou, Ganqu Cui, Shengding Hu, Zhengyan Zhang, Cheng Yang, Zhiyuan Liu, Lifeng Wang, Changcheng Li, and Maosong Sun. 2020. Graph neural networks: A review of methods and applications.�� 1 (2020), 57–81

2020
[2013]

The cancer genome atlas pan-cancer analysis project.�� 45, 10 (2013), 1113–1120

2013
[2020]

EZH2 regulates sFRP4 expression without affecting the methylation of sFRP4 promoter DNA in colorectal cancer cell lines.�� 20, 5 (2020), 33

2020
[2024]

Graph artificial intelligence in medicine.�� 7, 2024 (2024), 345–368

2024

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [3]

Jenna L Ballard, Zexuan Wang, Wenrui Li, Li Shen, and Qi Long. 2024. Deep learning-based approaches for multi-omics data integration and analysis.�� 17, 1 (2024), 38

work page 2024

[2] [4]

Leo Breiman. 2001. Random forests.�� 45, 1 (2001), 5–32

work page 2001

[3] [5]

Michael M Bronstein, Joan Bruna, Yann LeCun, Arthur Szlam, and Pierre Van- dergheynst. 2017. Geometric deep learning: going beyond euclidean data.�� 34, 4 (2017), 18–42

work page 2017

[4] [6]

Tianqi Chen and Carlos Guestrin. 2016. Xgboost: A scalable tree boosting system. In�� . 785–794

work page 2016

[5] [7]

Zhiqian Chen, Fanglan Chen, Lei Zhang, Taoran Ji, Kaiqun Fu, Liang Zhao, Feng Chen, Lingfei Wu, Charu Aggarwal, and Chang-Tien Lu. 2023. Bridging the gap between spatial and spectral domains: A unified framework for graph neural networks.�� 56, 5 (2023), 1–42

work page 2023

[6] [8]

Bernd Frank, Michael Hoffmeister, Norman Klopp, Thomas Illig, Jenny Chang- Claude, and Hermann Brenner. 2010. Single nucleotide polymorphisms in Wnt signaling and cell death pathway genes and susceptibility to colorectal cancer. ��31, 8 (2010), 1381–1386

work page 2010

[7] [9]

Justin Gilmer, Samuel S Schoenholz, Patrick F Riley, Oriol Vinyals, and George E Dahl. 2017. Neural message passing for quantum chemistry. In�� . PMLR, 1263–1272

work page 2017

[8] [10]

2016.��

Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016.�� . MIT press

work page 2016