Pith. sign in

REVIEW 4 major objections 5 minor 39 references

Exploiting Boosting in Hyperdimensional Computing for Enhanced Reliability in Healthcare

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read BoostHD splits the hyperdimensional space into boosted weak learners and beats single-model HDC on wearable stress data.

desk verdict BoostHD has a genuinely new subspace-ensembling idea and strong WESAD accuracy, but the published algorithm does not actually implement boosting and the theory is invalid, so it needs major revision and code before the claims are credible. read the letter →

arxiv 2411.14612 v2 pith:IQXFNQFO submitted 2024-11-21 cs.LG cs.AI

classification cs.LGcs.AI
keywords hyperdimensionalcomputingboostingensemblelearningwearablestressdetectionOnlineHDoverfittingnoiserobustnessclassimbalance
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 paper argues that a single hyperdimensional classifier underutilizes its own encoding space, and that this underuse is what limits accuracy and reliability. BoostHD splits the $D$-dimensional space into $n$ disjoint $D/n$ dimensional subspaces, trains a weak HDC learner on each, and combines them with AdaBoost-style sample reweighting. On the WESAD wearable stress dataset it reports 98.37% ± 0.32% accuracy versus 96.37% ± 0.40% for OnlineHD, with roughly one-third the standard deviation and smaller accuracy loss under bitflip noise. If the result holds, HDC becomes a more credible low-power option for healthcare monitoring rather than just a speed-oriented alternative to deep networks.

What carries the argument

The mechanism is a partitioned ensemble: the $D$-dimensional encoding is split into $n$ disjoint segments, each the working space of one OnlineHD-style weak learner trained on the same data; boosting reweights misclassified samples between learners, and inference sums the learners' predictions weighted by their error-derived $\alpha_i$. The theoretical support is Marchenko-Pastur analysis of the Gaussian kernel's singular values (Eqs. 2-3): mean $\mu_\lambda$ grows with $D$ while variance $\sigma^2_\lambda$ plateaus, so the kernel ellipse becomes nearly circular, and span utilization $SP = \text{rank}(K)/D$ attenuated by factors $\pi_i$ is used to argue that many small spaces are used more fully than one large space. The baseline-dimensionality condition keeps each segment informative: when $D_{total}/N_L$ falls below a threshold, accuracy collapses (shown for $N_L=100$, $D_{total}=1$K).

What would settle it

Fix $D_{total}=4000$ and $N_L=10$ and run BoostHD and OnlineHD on a dataset where classes are numerous or input features are highly correlated; if accuracy crosses below OnlineHD while each segment is at the same per-learner dimensionality as the single model's best setting, the partition benefit is not general. More directly, compute rank($K$)/$D$ for both on the same data: if BoostHD's span utilization is not higher, the theory's core quantity is contradicted.

Watch

Extended reading notes

Core claim

BoostHD's central claim is that partitioning the hyperdimensional space into boosted weak learners improves both performance and reliability relative to using the same total dimensionality as one strong learner. Each weak learner receives a $D/n$ segment and is trained sequentially; sample weights are updated by each learner's error, and inference weights predictions by model importance. The paper supports this with a random-matrix analysis: as $D$ grows, the singular-value mean of the HDC kernel increases while its variance stabilizes, so the kernel's ellipse becomes more circular, and it argues that the practical span utilization --- rank of the classifier matrix divided by $D$, discounted by cosine-similarity factors --- is higher in BoostHD. Empirically, BoostHD reaches 98.37% ± 0.32% on WESAD, 61.52% ± 0.07% on Nurse Stress, and 68.10% ± 0.09% on Stress-Predict, improving over OnlineHD, and it keeps higher accuracy under class imbalance, bitflip noise, and person-specific splits.

Load-bearing premise

The load-bearing premise is that splitting $D$ into $n$ segments keeps each weak learner above a 'baseline dimensionality' that preserves enough discriminative information; the paper demonstrates degradation when this fails but never defines the threshold or shows how it scales with dataset size.

Editorial extensions

If this is right

  • If the central claim is right, hyperdimensional computing reaches 98.37% on WESAD, beating Random Forest, XGBoost, and OnlineHD while keeping inference near $10^{-4}$ seconds.
  • The roughly threefold lower run-to-run standard deviation and the bounded bitflip loss make BoostHD a candidate for noisy wearable hardware where single-model HDC and DNNs degrade more.
  • Segment independence permits parallel inference; reported per-sample inference times of $1.1\times10^{-4}$ s on WESAD and $1.2\times10^{-4}$ s on Nurse Stress point to on-device deployment.
  • Accuracy holds as the imbalance ratio $r$ increases while OnlineHD drops, suggesting boosting reweighting offsets the overfitting HDC shows on skewed classes.
  • Average person-specific accuracy of 96.19% across hand preference, gender, age, and height subgroups indicates the ensemble narrows demographic accuracy gaps relative to other models.

Reading between the lines

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

  • A natural extension is to derive the baseline-dimensionality threshold from dataset size and class structure; the paper leaves it empirical, so a formula would let $N_L$ and $D_{total}$ be set automatically per dataset.
  • The span-utilization reasoning is a general claim about HDC geometry, not just stress detection; testing BoostHD on image or text benchmarks would show whether subspace ensembles beat a single hypervector outside healthcare.
  • Since the reweighting rule is standard AdaBoost, initializing sample weights with class-balanced or cost-sensitive values could push the imbalance results further; the paper does not test this variant.
  • The bitflip robustness may mix ensemble averaging with boosting effects; logging per-segment errors under noise would separate the two mechanisms.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes BoostHD, a hyperdimensional computing (HDC) ensemble method that splits a D-dimensional hypervector space into n disjoint segments, treats each segment as a weak learner, and allegedly combines the learners through sequential boosting with sample reweighting. The authors report accuracy gains over OnlineHD, Random Forest, XGBoost, SVM, and DNN on three healthcare datasets (WESAD, Nurse Stress, Stress-Predict), with the headline WESAD accuracy of 98.37% ± 0.32% versus 96.37% ± 0.40% for OnlineHD. Additional experiments address stability as a function of dimensionality, robustness to bit-flip noise, behavior under induced class imbalance, and person-specific subgroup performance. A theoretical discussion based on the Marchenko-Pastur distribution and an invented 'span utilization' metric is used to motivate why subspace partitioning improves reliability.

Significance. If the method were correctly specified and reproducible, the central empirical finding would be practically meaningful for low-power wearable stress detection, and the external-baseline comparisons in Table I are a genuine strength because they are not circular. The WESAD accuracy improvement over OnlineHD is internally consistent and the inference-efficiency results in Table II are useful. However, the paper has a load-bearing specification problem: the pseudocode in Algorithm 1 does not describe a working boosting procedure. In addition, the theoretical derivation that motivates the method is internally inconsistent. These issues must be resolved before the claims 'integrates boosting with HDC' and 'enhances performance and reliability' can be assessed.

major comments (4)
  1. [Section III, Algorithm 1] The training procedure as written is not a boosting algorithm. Line 4, 'Train fθi with X and y', never uses the sample-weight vector Ws, so the reweighting update in line 8 cannot influence any subsequent learner. Line 7, 'αi = Ws · eθi', mixes a vector of sample weights with a scalar error rate and is not the AdaBoost confidence α = 0.5 ln((1−e)/e). Inference line 3 sums ŷs · α even though ŷs is not defined as a class-score vector, and the argmax is therefore undefined. Because the paper's novelty is the integration of boosting with HDC, the reported gains cannot currently be attributed to boosting; the authors must either correct the pseudocode to an executable boosting loop or release the implementation.
  2. [Section III, Eqs. (2)–(7)] The Marchenko-Pastur analysis is internally inconsistent. The text defines q = Nc/Nr = Nc/D and states that q has an inverse relationship with D, yet Eqs. (4)–(6) take limits q→∞, which is the opposite scaling regime for large D. Moreover, the support bounds are written as (1+√q)^4 and (q−√q)^4, whereas the standard Marchenko-Pastur support for eigenvalues is (1±√q)^2; the eigenvalue-versus-singular-value convention is never fixed. Since the conclusion that σ²λ stays constant and the kernel becomes 'circular' relies on these limits, the theoretical justification in Section III is unsupported as written and should be corrected or removed.
  3. [Section III and Figure 3] The 'baseline dimensionality' is load-bearing for the stability claim but is never defined. The text states that failure to preserve this baseline causes substantial degradation (e.g., NL = 100 with Dtotal = 1K in Figure 3b), but no formula or scaling law is given to determine when a segment of size D/n is viable, and no connection to sample size or class structure is established. Without such a condition, the recommended operating region for D and NL is uncontrolled, and the claimed stability advantage of BoostHD cannot be generalized beyond the reported grid.
  4. [Section IV-C, Eq. (8)] The overfitting experiment is not interpretable as written. Equation 8 writes D = (x, if y = Ctarget; x × r, if y ≠ Ctarget), but the notation x × r is not defined for a data sample x and a scalar r. The text does not state whether the induced imbalance comes from replicating samples, discarding samples, or scaling feature values. Because Figure 7 is the sole support for the overfitting-resistance claim, this experimental description must be clarified or corrected before the result can be evaluated.
minor comments (5)
  1. [Section III] The 'span utilization' (SP) metric is defined informally: 'rank(K)/D divided by the product of π1, π2, ..., πn' does not specify what πi are or how they are computed. As an explanatory metric it should be formalized if it is retained.
  2. [Figure 8] The horizontal axis labels are confusing: the left panel shows pb values from 20 to 100 while the caption text refers to 10^-6 and 10^-5. The intended axis ranges and any log-scale transformations should be stated explicitly.
  3. [Table III] The text says BoostHD ranked second in two person-specific categories, but in the printed table it appears second only in the Height ≥ 185 row; please reconcile this statement with the table.
  4. [References] References [18] and [39] are incomplete (missing full author lists and, for [18], the publication venue details).
  5. [General] The manuscript does not mention a code or data release. Given the ambiguities in Algorithm 1, providing the implementation as supplementary material would be essential for verifying the reported results.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: BoostHD's reported accuracy gains are benchmarked against external baselines, and the SP and Marchenko-Pastur analyses are explanatory constructs rather than fitted inputs.

full rationale

I examined the claimed derivation chain for reductions to the paper's own inputs. The headline empirical claim (98.37% vs OnlineHD's 96.37% on WESAD) is measured against external baselines (RF, XGBoost, SVM, DNN, OnlineHD), so it does not reduce to the paper's definitions. The 'span utilization' (SP) metric is defined from rank(K)/D and cosine-similarity products, not from the accuracy figures it is later used to rationalize; thus SP is a post-hoc explanatory construct rather than a self-definitional prediction. The Marchenko-Pastur discussion (Eqs. 2-7) invokes an external theorem, and even if the application is mathematically questionable, that is a soundness issue rather than circularity. The only notable self-citation, OnlineHD [18], is used as the base encoder and as a baseline; the boosting claim does not rest on a disputed result from that paper, so the self-citation is not load-bearing. I also note that Algorithm 1 does not actually pass sample weights into weak-learner training (line 4 trains f_theta_i on X and y only), so the 'boosting' mechanism is under-specified; this is a reproducibility and correctness defect, not a circular reduction, because the reported accuracies are not constructed from those weights. No fitted parameter is renamed as a prediction, and no uniqueness theorem is imported to forbid alternatives. The 'baseline dimensionality' condition is asserted without derivation, but an unproven assumption is not a circular equivalence. Verdict: no significant circularity.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The experimental comparison is anchored to external baselines, but the explanation of why BoostHD works depends on an unverified Marchenko-Pastur application, a baseline-dimensionality assumption for weak learners, and the authors' own SP metric. D_total and NL are free hyperparameters whose headline values are not fully disclosed.

free parameters (3)
  • D_total (total hyperdimensional dimensionality) = not stated for Table I; scanned from 10 to 10,000
    Accuracy is reported after scanning D_total from 10 to 10,000; the value used for Table I is not stated, so the headline numbers depend on an unreported selection.
  • NL (number of weak learners) = 10
    The ensemble uses NL=10 for all BoostHD results; stability and performance vary with NL, so this is a free design choice.
  • Moving-average filter window size = 30
    All datasets are preprocessed with a moving-average filter of window size 30; results likely depend on this choice.
assumptions (3)
  • domain assumption Random Gaussian HDC feature matrices follow the Marchenko-Pastur law with q=Nc/Nr, and the conclusions in Equations 2-7 describe the HDC kernel's eigenvalue geometry.
    Invoked in Section III; the paper does not verify the convergence assumptions or handle the q->∞ vs q=Nc/D discrepancy.
  • domain assumption Splitting D into n disjoint segments preserves each weak learner's accuracy provided a baseline dimensionality is maintained.
    Unstated in Algorithm 1; Section III postulates the baseline empirically via Figure 3 without a formal bound.
  • ad hoc to paper Higher span utilization SP of class hypervectors implies higher classification accuracy.
    SP is introduced by the authors and used in Figure 5 to explain BoostHD's superiority; no external evidence validates the SP-accuracy link.
invented entities (1)
  • Span Utilization (SP)
    purpose: Quantifies the effective subspace spanned by class hypervectors, after discounting cosine-similarity overlaps; used to argue BoostHD uses the hyperdimensional space more efficiently.
    SP = rank(K)/D divided by a product of cosine similarities is computed from the model's own class hypervectors; it provides no falsifiable prediction outside the paper and is therefore not independent evidence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exploiting Boosting in Hyperdimensional Computing for Enhanced Reliability in Healthcare." pith.science (2026). https://pith.science/paper/IQXFNQFO

@misc{pith2026241114612,
  author       = {Pith},
  title        = {Pith review of: Exploiting Boosting in Hyperdimensional Computing for Enhanced Reliability in Healthcare},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IQXFNQFO}},
  note         = {Machine review of arXiv:2411.14612}
}
read the original abstract

Hyperdimensional computing (HDC) enables efficient data encoding and processing in high-dimensional space, benefiting machine learning and data analysis. However, underutilization of these spaces can lead to overfitting and reduced model reliability, especially in data-limited systems a critical issue in sectors like healthcare that demand robustness and consistent performance. We introduce BoostHD, an approach that applies boosting algorithms to partition the hyperdimensional space into subspaces, creating an ensemble of weak learners. By integrating boosting with HDC, BoostHD enhances performance and reliability beyond existing HDC methods. Our analysis highlights the importance of efficient utilization of hyperdimensional spaces for improved model performance. Experiments on healthcare datasets show that BoostHD outperforms state-of-the-art methods. On the WESAD dataset, it achieved an accuracy of 98.37%, surpassing Random Forest, XGBoost, and OnlineHD. BoostHD also demonstrated superior inference efficiency and stability, maintaining high accuracy under data imbalance and noise. In person-specific evaluations, it achieved an average accuracy of 96.19%, outperforming other models. By addressing the limitations of both boosting and HDC, BoostHD expands the applicability of HDC in critical domains where reliability and precision are paramount.

Figures

Figures reproduced from arXiv: 2411.14612 by the authors.

Figure 1
Figure 1. Illustration of the BoostHD framework applied to hyperdimensional computing (HDC). Sensor information is encoded into a high-dimensional [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Extreme distribution of terms, Eq. 4, 5, 6 in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Accuracy heatmap based on NL and their respective D. In (a) and (b), NL takes values from 1 to 100 and 10 to 100 with each step 1, 10. For (a), the accuracy is presented for each specified dimension. For (b), the total dimension(Dtotal) is divided among the NL, where each learner possesses a dimension size of Dtotal/NL. factors denoted by Π. These factors are the product sums of cosine similarity values between clas… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: In the process of kernel transformation. Data is mapped into a [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 6
Figure 6. Figure 6: Analysis of the impact of D on the stability of BoostHD and OnlineHD. (a), Accuracy of BoostHD and OnlineHD as a function of D, with error bars representing σ. (b), Variation of the σ in (a) with respect to D. C. Overfitting BoostHD presents a noteworthy advantage over…
Figure 7
Figure 7. Figure 7: Analysis of the impact of r on overfitting BoostHD and OnlineHD. [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 32 canonical work pages

  1. [1]

    Heal: Brain-inspired hyperdimensional efficient active learning,

    Y . Ni, Z. Zou, W. Huang, H. Chen, W. Y . Chung, S. Cho, R. Krishnan, P. Mercati, and M. Imani, “Heal: Brain-inspired hyperdimensional efficient active learning,” arXiv preprint arXiv:2402.11223 , 2024

  2. [2]

    A framework for collaborative learning in secure high- dimensional space,

    M. Imani, Y . Kim, S. Riazi, J. Messerly, P. Liu, F. Koushanfar, and T. Rosing, “A framework for collaborative learning in secure high- dimensional space,” in 2019 IEEE 12th International Conference on Cloud Computing (CLOUD) . IEEE, 2019, pp. 435–446

  3. [3]

    Biohd: an efficient genome sequence search platform using hyperdimensional memorization,

    Z. Zou, H. Chen, P. Poduval, Y . Kim, M. Imani, E. Sadredini, R. Cam- marota, and M. Imani, “Biohd: an efficient genome sequence search platform using hyperdimensional memorization,” in Proceedings of the 49th Annual International Symposium on Computer Architecture , 2022, pp. 656–669

  4. [4]

    Hdpg: Hyperdimensional policy-based reinforcement learning for continuous control,

    Y . Ni, M. Issa, D. Abraham, M. Imani, X. Yin, and M. Imani, “Hdpg: Hyperdimensional policy-based reinforcement learning for continuous control,” in Proceedings of the 59th ACM/IEEE Design Automation Conference, 2022, pp. 1141–1146

  5. [5]

    Algorithm-hardware co-design for efficient brain-inspired hyperdimensional learning on edge. in 2022 design, automation & test in europe conference & exhibition (date),

    Y . Ni, Y . Kim, T. Rosing, and M. Imani, “Algorithm-hardware co-design for efficient brain-inspired hyperdimensional learning on edge. in 2022 design, automation & test in europe conference & exhibition (date),” IEEE, 292 ´s297, 2022

  6. [6]

    Hdreason: Algorithm-hardware codesign for hyperdimensional knowledge graph reasoning,

    H. Chen, Y . Ni, A. Zakeri, Z. Zou, S. Yun, F. Wen, B. Khaleghi, N. Srinivasa, H. Latapie, and M. Imani, “Hdreason: Algorithm-hardware codesign for hyperdimensional knowledge graph reasoning,” arXiv preprint arXiv:2403.05763, 2024

  7. [7]

    Scalable and interpretable brain-inspired hyper-dimensional computing intelligence with hardware- software co-design,

    H. Chen, Y . Ni, W. Huang, and M. Imani, “Scalable and interpretable brain-inspired hyper-dimensional computing intelligence with hardware- software co-design,” in 2024 IEEE Custom Integrated Circuits Confer- ence (CICC). IEEE, 2024, pp. 1–8

  8. [9]

    Hyperdimensional computing for robust and efficient unsupervised learning,

    S. Yun, H. E. Barkam, P. R. Genssler, H. Latapie, H. Amrouch, and M. Imani, “Hyperdimensional computing for robust and efficient unsupervised learning,” in 2023 57th Asilomar Conference on Signals, Systems, and Computers . IEEE, 2023, pp. 281–288

Show all 39 references
  1. [10]

    Spatial-aware image re- trieval: A hyperdimensional computing approach for efficient similarity hashing,

    S. Yun, R. Masukawa, S. Jeong, and M. Imani, “Spatial-aware image re- trieval: A hyperdimensional computing approach for efficient similarity hashing,” arXiv preprint arXiv:2404.11025 , 2024

  2. [11]

    Neurally-inspired hyper- dimensional classification for efficient and robust biosignal processing,

    Y . Ni, N. Lesica, F.-G. Zeng, and M. Imani, “Neurally-inspired hyper- dimensional classification for efficient and robust biosignal processing,” in Proceedings of the 41st IEEE/ACM International Conference on Computer-Aided Design, 2022, pp. 1–9

  3. [12]

    Darl: Distributed reconfigurable accelerator for hyperdimensional reinforcement learning,

    H. Chen, M. Issa, Y . Ni, and M. Imani, “Darl: Distributed reconfigurable accelerator for hyperdimensional reinforcement learning,” in Proceed- ings of the 41st IEEE/ACM International Conference on Computer-Aided Design, 2022, pp. 1–9

  4. [13]

    Brain-inspired computing for in-process melt pool characterization in additive manufacturing,

    R. Chen, M. Sodhi, M. Imani, M. Khanzadeh, A. Yadollahi, and F. Imani, “Brain-inspired computing for in-process melt pool characterization in additive manufacturing,” CIRP Journal of Manufacturing Science and Technology, vol. 41, pp. 380–390, 2023

  5. [14]

    Real-time and robust hyperdimensional classification,

    A. Hern ´andez-Cano, C. Zhuo, X. Yin, and M. Imani, “Real-time and robust hyperdimensional classification,” in Proceedings of the 2021 on Great Lakes Symposium on VLSI , 2021, pp. 397–402

  6. [15]

    Efficient exploration in edge-friendly hyperdimensional reinforcement learning,

    Y . Ni, W. Y . Chung, S. Cho, Z. Zou, and M. Imani, “Efficient exploration in edge-friendly hyperdimensional reinforcement learning,” in Proceed- ings of the Great Lakes Symposium on VLSI 2024 , 2024, pp. 111–118

  7. [16]

    Density-aware parallel hyperdimensional genome sequence matching,

    H. Chen and M. Imani, “Density-aware parallel hyperdimensional genome sequence matching,” in 2022 IEEE 30th Annual Interna- tional Symposium on Field-Programmable Custom Computing Machines (FCCM). IEEE, 2022, pp. 1–4

  8. [17]

    Brain-inspired trustworthy hyperdimensional computing with efficient uncertainty quantification,

    Y . Ni, H. Chen, P. Poduval, Z. Zou, P. Mercati, and M. Imani, “Brain-inspired trustworthy hyperdimensional computing with efficient uncertainty quantification,” in 2023 IEEE/ACM International Conference on Computer Aided Design (ICCAD) . IEEE, 2023, pp. 01–09

  9. [18]

    Onlinehd: Robust, efficient, and single-pass online learning using hyperdimensional system,

    A. Hernandez-Cane et al., “Onlinehd: Robust, efficient, and single-pass online learning using hyperdimensional system,” in DATE, 2021

  10. [19]

    An overview of overfitting and its solutions,

    X. Ying, “An overview of overfitting and its solutions,” in Journal of physics: Conference series, vol. 1168. IOP Publishing, 2019, p. 022022

  11. [20]

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

    C. Rudin, “Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead,” Nature machine intelligence, vol. 1, no. 5, pp. 206–215, 2019

  12. [21]

    Artificial intelligence in disease diagnosis: a systematic literature review, synthesizing frame- work and future research agenda,

    Y . Kumar, A. Koul, R. Singla, and M. F. Ijaz, “Artificial intelligence in disease diagnosis: a systematic literature review, synthesizing frame- work and future research agenda,” Journal of ambient intelligence and humanized computing, pp. 1–28, 2022

  13. [22]

    Explaining adaboost,

    R. E. Schapire, “Explaining adaboost,” in Empirical Inference: Festschrift in Honor of Vladimir N. Vapnik. Springer, 2013, pp. 37–52

  14. [23]

    A survey on hyperdimensional computing aka vector symbolic architectures, part ii: Applications, cognitive models, and challenges,

    D. Kleyko, D. Rachkovskij, E. Osipov, and A. Rahimi, “A survey on hyperdimensional computing aka vector symbolic architectures, part ii: Applications, cognitive models, and challenges,” ACM Computing Surveys, vol. 55, no. 9, pp. 1–52, 2023

  15. [24]

    A theoretical perspective on hyperdimensional computing,

    A. Thomas, S. Dasgupta, and T. Rosing, “A theoretical perspective on hyperdimensional computing,” Journal of Artificial Intelligence Research, vol. 72, pp. 215–249, oct 2021. [Online]. Available: https://doi.org/10.1613%2Fjair.1.12664

  16. [25]

    Greedy function approximation: a gradient boosting machine,

    J. H. Friedman, “Greedy function approximation: a gradient boosting machine,” Annals of statistics , pp. 1189–1232, 2001

  17. [26]

    Lightgbm: A highly efficient gradient boosting decision tree,

    G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T.- Y . Liu, “Lightgbm: A highly efficient gradient boosting decision tree,” Advances in neural information processing systems , vol. 30, 2017

  18. [27]

    Xgboost: extreme gradient boosting,

    T. Chen, T. He, M. Benesty, V . Khotilovich, Y . Tang, H. Cho, K. Chen, R. Mitchell, I. Cano, T. Zhou et al. , “Xgboost: extreme gradient boosting,” R package version 0.4-2 , vol. 1, no. 4, pp. 1–4, 2015

  19. [28]

    Introducing wesad, a multimodal dataset for wearable stress and affect detection,

    P. Schmidt, A. Reiss, R. Duerichen, C. Marberger, and K. Van Laer- hoven, “Introducing wesad, a multimodal dataset for wearable stress and affect detection,” in Proceedings of the 20th ACM international conference on multimodal interaction , 2018, pp. 400–408

  20. [29]

    Deep PPG: Large-Scale heart rate estimation with convolutional neural networks,

    A. Reiss, I. Indlekofer, P. Schmidt, and K. Van Laerhoven, “Deep PPG: Large-Scale heart rate estimation with convolutional neural networks,” Sensors (Basel), vol. 19, no. 14, Jul. 2019

  21. [30]

    Hyperdimensional computing for resilient edge learning,

    H. E. Barkam, S. E. Jeon, S. Yun, C. Yeung, Z. Zou, X. Jiao, N. Srini- vasa, and M. Imani, “Hyperdimensional computing for resilient edge learning,” in 2023 IEEE/ACM International Conference on Computer Aided Design (ICCAD) . IEEE, 2023, pp. 1–8

  22. [31]

    Reliable hyperdimensional reasoning on unreliable emerging technologies,

    H. E. Barkam, S. Yun, H. Chen, P. Gensler, A. Mema, A. Ding, G. Mich- elogiannakis, H. Amrouch, and M. Imani, “Reliable hyperdimensional reasoning on unreliable emerging technologies,” in 2023 IEEE/ACM International Conference on Computer Aided Design (ICCAD) . IEEE, 2023, pp. 1–9

  23. [32]

    Comprehensive analysis of hyperdimensional computing against gra- dient based attacks,

    H. E. Barkam, S. E. Jeon, C. Yeung, Z. Zou, X. Jiao, and M. Imani, “Comprehensive analysis of hyperdimensional computing against gra- dient based attacks,” in 2023 Design, Automation & Test in Europe Conference & Exhibition (DATE) . IEEE, 2023, pp. 1–2

  24. [33]

    Hyper- graf: Hyperdimensional graph-based reasoning acceleration on fpga,

    H. Chen, A. Zakeri, F. Wen, H. E. Barkam, and M. Imani, “Hyper- graf: Hyperdimensional graph-based reasoning acceleration on fpga,” in 2023 33rd International Conference on Field-Programmable Logic and Applications (FPL). IEEE, 2023, pp. 34–41

  25. [34]

    V oicehd: Hyperdi- mensional computing for efficient speech recognition,

    M. Imani, D. Kong, A. Rahimi, and T. Rosing, “V oicehd: Hyperdi- mensional computing for efficient speech recognition,” in 2017 IEEE international conference on rebooting computing (ICRC) . IEEE, 2017, pp. 1–8

  26. [35]

    Hyperdimensional hybrid learning on end- edge-cloud networks,

    M. Issa, S. Shahhosseini, Y . Ni, T. Hu, D. Abraham, A. M. Rahmani, N. Dutt, and M. Imani, “Hyperdimensional hybrid learning on end- edge-cloud networks,” in 2022 IEEE 40th International Conference on Computer Design (ICCD) . IEEE, 2022, pp. 652–655

  27. [36]

    Robust in-memory computing with hyperdimensional stochastic repre- sentation. in 2021 ieee,

    P. Poduval, M. Issa, F. Imani, C. Zhuo, X. Yin, H. Najafi, and M. Imani, “Robust in-memory computing with hyperdimensional stochastic repre- sentation. in 2021 ieee,” in ACM International Symposium on Nanoscale Architectures (NANOARCH), pp. 1–6

  28. [37]

    Stochd: Stochastic hyperdimensional system for efficient and robust learning from raw data,

    P. Poduval, Z. Zou, H. Najafi, H. Homayoun, and M. Imani, “Stochd: Stochastic hyperdimensional system for efficient and robust learning from raw data,” in 2021 58th ACM/IEEE Design Automation Conference (DAC). IEEE, 2021, pp. 1195–1200

  29. [38]

    Rate of convergence in probability to the marchenko-pastur law,

    F. G ¨otze and A. Tikhomirov, “Rate of convergence in probability to the marchenko-pastur law,” Bernoulli, vol. 10, no. 3, pp. 503–548, 2004

  30. [39]

    A multimodal sensor dataset for continuous stress detection of nurses in a hospital,

    S. Hosseini, R. Gottumukkala, S. Katragadda, R. T. Bhupatiraju, Z. Ashkar, C. W. Borst, and K. Cochran, “A multimodal sensor dataset for continuous stress detection of nurses in a hospital,” Scientific Data, vol. 9, no. 1, p. 255, 2022

  31. [40]

    Stress monitoring using wearable sensors: A pilot study and stress-predict dataset,

    T. Iqbal, A. J. Simpkin, D. Roshan, N. Glynn, J. Killilea, J. Walsh, G. Molloy, S. Ganly, H. Ryman, E. Coen et al., “Stress monitoring using wearable sensors: A pilot study and stress-predict dataset,” Sensors, vol. 22, no. 21, p. 8135, 2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.