Pith. sign in

REVIEW 3 major objections 6 minor 47 references

Shrinking the Haystack: One-Class Machine Learning Detection of Magnetosheath Current Sheets in MMS Burst Data

T0 review · 3 major / 6 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A one-class detector shrinks 22,775 MMS burst windows to 270 candidates, cutting the reconnection search space by 98.8%.

desk verdict A solid, honestly-scoped proof of concept for MMS burst data reduction; treat the 78% precision as in-sample until there's an out-of-sample test. read the letter →

arxiv 2608.00252 v1 pith:ES45CQO3 submitted 2026-07-31 physics.space-ph

Deep Ghuge , Daniel J. Gershman , Vadim Uritsky , Julia E. Stawarz This is my paper · ORCID
classification physics.space-ph
keywords magneticreconnectionmagnetosheathturbulenceMMSburstdataone-classclassificationdeepSVDDstructurefunctioncurrentsheetsreduction
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

This paper argues that a morphology-first, one-class machine-learning pipeline can reduce the search for small, short reconnecting current sheets in MMS burst-mode data to a manageable review queue. The key idea is to train only on synthetic surrogates of a single well-understood reference event, using the second-order structure function S2(τ) to ensure the surrogates share the same multi-scale fingerprint. Applied to 15 magnetosheath intervals, the pipeline compresses 22,775 sliding windows to 270 detections, of which 211 are visually confirmed as sheet-like or reconnection-like. If this holds, it offers a practical way to find rare reconnection events in noisy turbulence without a curated negative class.

What carries the argument

The load-bearing object is the S2(τ)-anchored surrogate library: a Monte Carlo generator built around a single reference event (MMS1, 2017-01-28 09:09:01 UTC). The generator produces candidate BL traces from an asymmetric Harris sheet template plus two-band colored noise, admits them only if their S2(τ) matches the seed's multi-scale fingerprint, and then uses these surrogates to train a one-class Deep SVDD encoder that maps each window to a 32-dimensional latent embedding. The SVDD distance to a fixed center, converted to a percentile-based match score, is what separates sheet-like windows from everything else.

What would settle it

Re-run the pipeline with a different seed event (e.g., an electron-only reconnection event from Phan et al. 2018) and compare the candidate lists; if the detector recovers dramatically different detections that do not overlap the original 270, the single-seed assumption fails.

Watch

Extended reading notes

Core claim

The central claim is that one-class Deep SVDD, trained exclusively on 3,000 Monte Carlo surrogates of a single reference event, can serve as an effective Phase-1 data-reduction stage. Each surrogate is accepted only if its S2(τ) profile stays within a 0.20 dex band of the reference and passes five shape checks. The detector then scores real windows by their distance in a 32-dimensional latent space, reporting windows with match score ≥0.7. Across 15 intervals, the pipeline retains 270 of 22,775 windows (98.8% reduction), and manual review identifies 93 candidate reconnection events and 118 sheet-like events, with the retained queue at 78% useful.

Load-bearing premise

The surrogate library built from a single reference event is representative of the morphological diversity of magnetosheath current sheets across all 15 test intervals.

Editorial extensions

If this is right

  • A search space of tens of thousands of windows can be reduced to a few hundred reviewable candidates, making manual screening feasible for long burst-mode datasets.
  • The candidate-reconnection pool (93 events) is larger than the 24 catalog events overlapped, demonstrating that a morphology-first detector can surface events missed by threshold-based multi-spacecraft catalogs.
  • Because the score measures sheet-likeness, not reconnection, the method is a natural first stage before flow-channel or multi-spacecraft validation.
  • Retargeting the detector to another regime only requires swapping the seed event and re-running the Monte Carlo and training steps.
  • The detected population includes electron-scale sheets (e.g., 10.6 de thickness) that match the electron-only reconnection regime.

Reading between the lines

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

  • The single-seed design implies that the detector's validity is bounded by how well the surrogate library represents the full diversity of magnetosheath current sheets; a systematic study varying the seed would clarify the actual sensitivity.
  • The 78% visual-screening yield suggests that the SVDD score, by ignoring plasma flow channels, likely misses reconnection events whose magnetic signature is weak, even if they have strong jets or energy conversion.
  • Adding flow-channel features (e.g., electron velocity or J·E′) to the surrogate generator could push label-1-versus-label-2 discrimination into the automated stage, as the paper itself suggests via symbolic regression or self-supervised refinement.
  • The split between the 93 label-1 and 118 label-2 detections hints that a large fraction of magnetosheath current sheets are not actively reconnecting, which could be tested with multi-spacecraft follow-ups on the retained candidates.
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

3 major / 6 minor

Summary. The paper presents a two-stage, morphology-first pipeline for reducing MMS burst-mode data to a manageable set of reconnection-candidate current sheets. Stage 1 is an LMN frame-quality gate based on minimum-variance eigenvalue ratios. Stage 2 is a one-class Deep SVDD encoder trained on 3,000 synthetic surrogates generated by Monte Carlo from a single published reference event, with surrogate acceptance controlled by an S2(τ) acceptance band and several shape-based checks. Applied to 15 magnetosheath intervals from Stawarz et al. (2022), the pipeline reduces 22,775 sliding windows to 270 detections (98.8% compression). Manual visual screening labels 93 as candidate reconnection events and 118 as sheet-like, giving a combined 78.1% 'sheet-or-better' fraction. A cross-check against the Stawarz et al. reconnection catalog gives 24/78 (30.8%) recall over overlapping intervals. The paper positions the work as a Phase-1 data-reduction tool, with reconnection confirmation deferred to multi-spacecraft follow-up.

Significance. If the reported precision transfers out of sample, the pipeline would be a genuinely useful Phase-1 filter: it converts tens of thousands of burst windows into a few hundred visually reviewable candidates, and the one-class design is an elegant response to the absence of a curated negative class. The paper is transparent about its engineering choices and limitations, and the compression arithmetic is internally consistent. The explicitly physics-anchored surrogate construction, with S2(τ) as an acceptance criterion, is a worthwhile contribution. However, the headline 78% figure rests on threshold selection and visual labeling performed on the same 15 intervals, so its out-of-sample value is unestablished; the single-seed surrogate library also leaves the representativeness of the training distribution untested. These issues are load-bearing for the central claim that the queue is mostly worth reviewing.

major comments (3)
  1. [§3.3 and §4.2] The central quantitative claim — that 78.1% of the 270 detections are sheet-like or reconnection-like (Table 2) — is an in-sample figure. The SVDD decision threshold M≥0.7 was chosen by evaluating thresholds {0.3, 0.5, 0.7, 0.85} on exactly the 15 intervals used for evaluation (§3.3), and the visual labels in §4.2 were assigned by the authors on those same intervals. The text mentions 'blind manual screening' only in §4.4, without describing any blinding protocol; the reviewers knew which windows the pipeline selected. As a result, the abstract's headline 'retaining 78% of the queue' has no out-of-sample estimate and is likely optimistic. This is fixable: hold out intervals or use leave-one-interval-out for threshold selection, report precision with confidence intervals, and have an independent, blinded reviewer label at least a subset of detections. I would also ask for a table showing
  2. [§3.2, §4.4, §5] The single-seed design sets the entire training distribution, and the paper explicitly acknowledges in §5 that 'the single seed sets the regime of validity.' But the quantitative consequence is not assessed. Against the Stawarz et al. catalog, the pipeline recovers only 24/78 events (30.8%), i.e., 54 catalog events are missed. The paper attributes these misses to the LMN gate and to the single-seed morphology (§4.4), but provides no decomposition of how many catalog events fail the LMN gate versus the SVDD score. Without this breakdown, it is impossible to tell whether the morphology scorer is the limiting factor or merely inherits the gate's exclusions. Please add a failure-mode analysis for the 54 missed catalog events (gate-fail vs score-fail vs grouping-fail) and, ideally, a multi-seed ablation to show how sensitive recall is to the seed choice. This is directly relevant to the claim
  3. [§1 and §4.5] The paper motivates the morphology-based approach by contrasting it with threshold/PVI methods, stating that pure-threshold methods cannot distinguish a clean small-amplitude crossing from a diffuse large fluctuation. However, no baseline method is actually run on the same 15 intervals. The ρJ comparison in §4.5 is indirect: it shows that catalog-matched detections have more isolated |J| peaks than non-overlap candidates, but it does not demonstrate that a PVI- or |J|-threshold search would produce a worse queue (e.g., more false positives or fewer true events). Since the paper's value proposition is data-reduction quality, a direct comparison with a standard threshold baseline on the same intervals would substantially strengthen the contribution.
minor comments (6)
  1. [§4.2/§4.4] The term 'blind manual screening' appears in §4.4 but no blinding protocol is described in §4.2. Please clarify whether the labelers were blind to the pipeline's score or to the catalog, and how disagreements were resolved.
  2. [Table 2] The row 'Catalog events touched 24 (out of overlapped subset)' is confusing because the text says 22 detection windows overlap catalog events while touching 24 unique rows. Please define the denominator and explain why these numbers differ.
  3. [Title page] The author list contains a typo: 'V adim Uritsky' should read 'Vadim Uritsky'.
  4. [§3.2, Eq. (3)] The text mentions an 'odd-polynomial term inside the tanh argument' but never defines it. Either provide the explicit form or remove the phrase.
  5. [§3.2] The five 'shape-based checks' plus the S2(τ) band are described as six rejection criteria in Figure 4 and as 'five shape-based checks' in the text. Make the counting consistent (band + 5 checks = 6).
  6. [§5] The sentence 'the candidate-reconnection pool includes all 18 detections that both overlap a published catalog event and pass the visual criterion' is clear, but the earlier phrase 'all 18 detections' could be misread as all 18 catalog-overlap detections; consider rewording to 'all 18 of the 22 catalog-overlap detections that passed the visual criterion.'

Circularity Check

2 steps flagged · score 6.0 of 10

Headline 78% retention rate is partly self-referential (label-2 features equal the training features) and partly an in-sample threshold-tuned statistic.

  1. self definitional [Section 3.2–3.3 (Eqs. 3–8) and Section 4.2 label definitions]
    "the second training channel is the absolute derivative |dBL/dt|, which serves as a current-density proxy at fixed cadence ... A window is reported as a positive detection if M(x) ≥ 0.7, i.e. its distance ranks below the 30th percentile of the training cloud. ... A detection is assigned label 2, sheet-like, if both (i) BL shows a clear sharp change across the marked window, often reversing sign or rotating strongly, and (ii) |J| shows a clear local enhancement near that same window."

    The positive class is built from Harris tanh BL reversals plus |dBL/dt| as current proxy, accepted only when S2(τ) is inside a ±0.20 dex band around the single seed. M is then defined as the percentile rank of a window's latent distance within that seed-anchored cloud. The visual 'sheet-like' criterion independently requires exactly the same two features (sharp BL change and |J| enhancement). Thus the 78.1% 'sheet-or-better' figure largely confirms the detector's own training definition rather than providing an independent physical validation; the paper concedes 'the score measures how sheet-like a window is, not how reconnecting it is.'

  2. fitted input called prediction [Section 3.3 threshold choice; Section 4.2/Table 2 result]
    "The value 0.7 was chosen pragmatically: we evaluated thresholds of 0.3, 0.5, 0.7, and 0.85 and observed qualitatively that lower thresholds admitted many low-quality detections that were costly to review, while 0.85 excluded a non-trivial fraction of clear sheet-like crossings. A systematic sensitivity analysis of this threshold is left for future work."

    The decision threshold M≥0.7 is a free parameter selected by inspecting the very same 15 intervals that are then used to report the headline outcome. The abstract's 'retaining 78% of the queue' and Table 2's 78.1% are computed on that same evaluation set after this selection, so they are in-sample, threshold-tuned statistics rather than out-of-sample predictions. A different threshold would change both 270 and the 78% figure, and no held-out split or sensitivity estimate is provided; the quantitative evidence for pipeline usefulness is therefore partly fitted, not predictive.

full rationale

The main circularity is in the evaluation rather than in a claimed physical derivation. The model is deliberately anchored to one reference event: surrogates enter the library only if their S2(τ) tracks the seed's within ±0.20 dex, the second channel is |dBL/dt| as a current proxy, and the detection score is a percentile rank within that cloud. Consequently, 'sheet-like' is partly defined by the same features the human label-2 criterion requires (sharp BL change plus |J| enhancement), so the 78% retention claim is a partially self-referential measure. Separately, the M≥0.7 threshold was chosen on the same 15 intervals used to produce the headline numbers, making the reported precision an in-sample, threshold-tuned statistic. The paper is transparent about the core limitation — 'the single seed sets the regime of validity' — and it explicitly avoids claiming that S2 uniquely identifies sheet-like morphology. The cross-check against the Stawarz et al. (2022) catalog (24/78 recall; 18/22 agreement on overlaps) is a genuine external benchmark, but it tests recall only and does not rescue the in-sample precision claim. The raw 98.8% search-space reduction is an independent operational fact, so the work has real content; nevertheless the central quantitative evidence of usefulness (78% sheet-or-better) partially reduces to the method's own construction and threshold fitting, giving a partial-circularity score of 6.

Assumptions & free parameters 10 free parameters · 6 assumptions · 0 invented entities

The method introduces no new physical entities. The free parameters are numerous and mostly engineering choices tuned on the fly; several (SVDD threshold, S2 tolerances, shape-check thresholds) directly affect the reported performance numbers, and the lack of sensitivity analysis weakens the claim that the configuration is robust. The axioms are domain assumptions about the validity of MVA, S2 as a fingerprint, and the representative power of the surrogate library.

free parameters (10)
  • S2 acceptance band (0.20 dex, 60% lags) = ±0.20 dex; ≥60% of lags
    Equation 5; engineering tolerance tuned so visually acceptable surrogates pass; no sensitivity study.
  • Kinetic-range slope tolerance = ±0.6
    Shape check (3); chosen to match reference slope.
  • Shape-check thresholds (ramp/curvature, multi-peak, clone, leakage) = e.g., variance >99.8%, curvature <2e-3, peaks >3, MSE <0.02, corr >0.985, leakage <35%
    Section 3.2; 'engineering choices rather than the product of a systematic sensitivity study.'
  • SVDD decision threshold M≥0.7 = 0.7
    Section 3.3; chosen pragmatically by inspecting detection quality at 0.3, 0.5, 0.7, 0.85 on the 15 intervals.
  • LMN eigenvalue ratios r12, r23 ≥5 = ≥5
    Section 3.4; gate to ensure well-defined frame; no sensitivity analysis.
  • Merge gap and window/stride = merge gap 0.3 s; window 2 s; stride 0.25 s
    Sections 3.3-3.4; chosen to balance locality and shoulder content.
  • Reconstruction regularizer λ = 0.25
    Equation 6; chosen to block trivial collapse.
  • Surrogate library size N = 3000
    Section 3.2; chosen for percentile stability; 'acceptance fractions are typically a few percent.'
  • Noise model parameters (band slopes, envelope) = not specified numerically
    Section 3.2; two-band colored noise fit to seed S2; exact slopes and envelope not given.
  • Harris parameter priors (A, w, t0) = centered on least-squares fit to reference
    Section 3.2; define surrogate morphology.
assumptions (6)
  • domain assumption MVA with r≥5 gives a physically meaningful LMN frame.
    Section 3.4; if eigenvalue ratios are not well separated, BL is a noisy mixture and the score loses meaning.
  • domain assumption S2(τ) is a sufficient multi-scale fingerprint for sheet-like morphology.
    Section 3.2; the acceptance test and library generation rely on this; the paper acknowledges S2 alone is not unique but argues the shape checks mitigate.
  • ad hoc to paper The Harris+noise generator spans the real variability of magnetosheath current sheets.
    Section 3.2; the surrogate library defines the in-class distribution; if the generator misses real morphologies, detection will be biased.
  • domain assumption One-class training on synthetic surrogates generalizes to real MMS windows.
    Section 3.3; the Deep SVDD is trained only on surrogates and applied to real data; generalization is assumed.
  • domain assumption Manual visual labels (Section 4.2 criteria) are a valid proxy for reconnection candidacy.
    The 78% precision measure depends on the authors' labels; no independent validation.
  • domain assumption Stawarz et al. (2022) catalog is a reliable sanity check.
    The catalog is used to measure overlap; however, it is from the same author group and is more conservative by design.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Shrinking the Haystack: One-Class Machine Learning Detection of Magnetosheath Current Sheets in MMS Burst Data." pith.science (2026). https://pith.science/paper/ES45CQO3

@misc{pith2026260800252,
  author       = {Pith},
  title        = {Pith review of: Shrinking the Haystack: One-Class Machine Learning Detection of Magnetosheath Current Sheets in MMS Burst Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ES45CQO3}},
  note         = {Machine review of arXiv:2608.00252}
}
abstract

We present a morphology-first framework for narrowing the search for magnetic-reconnection candidates in Magnetospheric Multiscale (MMS) burst-mode data. The target is the small, short, and frequently electron-only reconnecting current sheets that occur in turbulent magnetosheath plasma. The pipeline operates in two stages. A local frame-quality gate based on minimum-variance analysis first retains only windows whose current-sheet coordinates are well defined. A one-class Deep Support Vector Data Description neural network then scores those windows against a library of 3,000 physically calibrated synthetic current sheets generated by Monte Carlo from a single published reference event. Acceptance into the surrogate library is governed by the second-order structure function $S_2(\tau)$: a candidate is admitted only if its multi-scale fingerprint tracks that of the seed event inside a tolerance band, together with a small number of shape-based checks. This $S_2(\tau)$-anchored construction defines the in-class distribution directly from a well-understood reference event and sidesteps the absence of a curated negative class in turbulent magnetosheath data. Applied to 15 magnetosheath turbulence intervals from the literature (1.58 h of burst-mode coverage), the framework compresses 22,775 sliding windows to 270 candidate detections (a 98.8% reduction). Manual visual screening identifies 93 of these as candidate reconnection events and a further 118 as sheet-like, retaining 78% of the queue for follow-up; the candidate-reconnection pool extends well beyond the 22 detections that overlap the published reconnection-event catalog used here as a sanity check. The framework is intended as the data-reduction stage of a broader reconnection-search workflow, offered here as an initial proof of concept before extending the one-class design to additional feature channels.

Figures

Figures reproduced from arXiv: 2608.00252 by the authors.

Figure 1
Figure 1. Reference event (MMS1, 2017 Jan 28, 09:09:01 UTC) used to anchor the surrogate generator. Panels: (a) B in LMN coordinates and |B|; (b) ne, ni; (c)–(d) Te, Ti; (e)–(f) Ve,Vi in LMN; (g) J from Equation 1; (h) |J|; (i) |dBL/dt|; (j) J · E ′ . Red dashed lines bound the 2 s analysis window. The S2(τ ) profile of the boxed window defines the acceptance band used by the surrogate generator (Equation 5). –5– [PITH_FULL_… view at source ↗
Figure 2
Figure 2. End-to-end pipeline. Top row (inference): MMS burst-mode FGM, FPI, and EDP data are rotated into the local LMN frame by minimum-variance analysis (MVA), screened by an LMN-quality gate, scored by the trained SVDD model, grouped into packets, and reduced to one representative window per packet for downstream review. Bottom row (training): a single MMS reference event seeds a Monte Carlo surrogate generator; the resul… view at source ↗
Figure 1
Figure 1. This event is well-suited as a seed because every signature the surrogate library [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figures from the paper (8 more)
Figure 3
Figure 3. Figure 3: Structure-function acceptance test. (a) BL(t) profile for an accepted surrogate (top, blue), a low-S2 candidate that strips small-scale variability (middle, salmon), and an over￾noisy candidate that lets broadband fluctuations leak into the reversal core (bottom, orang…
Figure 4
Figure 4. Figure 4: Block diagram of the rejection-sampling Monte Carlo generator. Inside the sampling loop (left), Harris parameters and a two-band coloured-noise process produce a candidate BL trace, which is bandlimited, z-scored, and given a random sign. Each candidate is then evaluat…
Figure 5
Figure 5. Figure 5: Surrogate ensemble. (a) Reference trace (black) and 15 surrogate traces for both training channels: BL (top) and |dBL/dt| (bottom). The surrogates show clear shape diversity around the reference while preserving the multi-scale fingerprint by construction. (b) FWHM-bas…
Figure 6
Figure 6. Figure 6: Score-packet logic. (a) A 10 s sliding-window scan covering 2015-10-21 07:01:52.752– 07:02:02.752 UTC and centred on a real detected packet. Blue circles show SVDD match scores for LMN-pass windows; the dashed red line is the threshold (M ≥ 0.7); orange markers are acc…
Figure 7
Figure 7. Figure 7: Label-1 (candidate reconnection) detection in interval 1, 2015-10-21 07:01:57.900– 58.500 UTC. Standard 11-panel detection plot with user-placed manual boundaries (red dashed) and the automated 2 s detection window (faded). Manual single-spacecraft workup: M = 0.84, ∆t…
Figure 8
Figure 8. Figure 8: Label-1 detection in interval 14, 2016-12-11 15:29:15.680–16.090 UTC. Plot conventions as in [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: Label-1 detection in interval 13, 2016-12-09 09:27:01.560–01.975 UTC. Plot conventions as in [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: Example of a label-3 (not-convincing) detection. The 2 s window is selected by the SVDD scorer because the magnetic channel contains a strong local gradient, but the co-located plasma response that distinguishes a current-sheet crossing is absent: |J| has no clean loc…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 15 canonical work pages

  1. [1]

    , Lacombe, C

    Alexandrova2012 APACrefauthors Alexandrova, O. , Lacombe, C. , Mangeney, A. , Grappin, R. \ Maksimovic, M. APACrefauthors \ 2012 . Solar wind turbulent spectrum at plasma kinetic scales Solar wind turbulent spectrum at plasma kinetic scales . The Astrophysical Journal 760 2 121 . APACrefDOI doi:10.1088/0004-637X/760/2/121 APACrefDOI

    doi:10.1088/0004-637x/760/2/121 2012

  2. [2]

    , Moore, T E

    Burch2016 APACrefauthors Burch, J L. , Moore, T E. , Torbert, R B. \ Giles, B L. APACrefauthors \ 2016 . Magnetospheric Multiscale Overview and Science Objectives Magnetospheric Multiscale overview and science objectives . Space Science Reviews 199 1--4 5--21 . APACrefDOI doi:10.1007/s11214-015-0164-9 APACrefDOI

    doi:10.1007/s11214-015-0164-9 2016

  3. [3]

    , Torbert, R B

    BurchScience2016 APACrefauthors Burch, J L. , Torbert, R B. , Phan, T D. \ . APACrefauthors \ 2016 . Electron-scale measurements of magnetic reconnection in space Electron-scale measurements of magnetic reconnection in space . Science 352 6290 aaf2939 . APACrefDOI doi:10.1126/science.aaf2939 APACrefDOI

    doi:10.1126/science.aaf2939 2016

  4. [4]

    APACrefauthors \ 2019

    Camporeale2019 APACrefauthors Camporeale, E. APACrefauthors \ 2019 . The challenge of machine learning in space weather: Nowcasting and forecasting The challenge of machine learning in space weather: Nowcasting and forecasting . Space Weather 17 8 1166--1207 . APACrefDOI doi:10.1029/2018SW002061 APACrefDOI

    doi:10.1029/2018sw002061 2019

  5. [5]

    \ Shay, M A

    CassakShay2007 APACrefauthors Cassak, P A. \ Shay, M A. APACrefauthors \ 2007 . Scaling of asymmetric magnetic reconnection: General theory and collisional simulations Scaling of asymmetric magnetic reconnection: General theory and collisional simulations . Physics of Plasmas 14 102114 . APACrefDOI doi:10.1063/1.2795630 APACrefDOI

    work page doi:10.1063/1.2795630 2007

  6. [6]

    , Sanchez-Gonzalez, A

    Cranmer2020 APACrefauthors Cranmer, M. , Sanchez-Gonzalez, A. , Battaglia, P. , Xu, R. , Cranmer, K. , Spergel, D. \ Ho, S. APACrefauthors \ 2020 . Discovering symbolic models from deep learning with inductive biases Discovering symbolic models from deep learning with inductive biases . Advances in Neural Information Processing Systems 33 17429--17442

    2020

  7. [7]

    Denton2018 APACrefauthors Denton, R E. \ . APACrefauthors \ 2018 . Determining L-M-N Current Sheet Coordinates at the Magnetopause From Magnetospheric Multiscale Data Determining L-M-N current sheet coordinates at the magnetopause from magnetospheric multiscale data . Journal of Geophysical Research: Space Physics 123 3 2274--2295 . APACrefDOI doi:10.1002...

    work page doi:10.1002/2017ja024619 2018

  8. [8]

    , Balogh, A

    Dunlop2002 APACrefauthors Dunlop, M W. , Balogh, A. , Glassmeier, K H. \ Robert, P. APACrefauthors \ 2002 . Four-point Cluster application of magnetic field analysis tools: The Curlometer Four-point Cluster application of magnetic field analysis tools: The curlometer . Journal of Geophysical Research: Space Physics 107 A11 1384 . APACrefDOI doi:10.1029/20...

    doi:10.1029/2001ja005088 2002

Show all 47 references
  1. [9]

    Finley2024 APACrefauthors Finley, M G. \ . APACrefauthors \ 2024 . Generalized time-series analysis for in situ spacecraft observations: anomaly detection and data prioritization using principal components analysis and unsupervised clustering Generalized time-series analysis f...

    2024 doi

  2. [10]

    APACrefauthors \ 1995

    Frisch1995 APACrefauthors Frisch, U. APACrefauthors \ 1995 . Turbulence: The Legacy of A. N. Kolmogorov Turbulence: The legacy of A. N. Kolmogorov . Cambridge University Press . APACrefDOI doi:10.1017/CBO9781139170666 APACrefDOI

    1995 doi

  3. [11]

    Genestreti2025 APACrefauthors Genestreti, K J. \ . APACrefauthors \ 2025 . Structure of the Electron Diffusion Region During Magnetic Reconnection Structure of the electron diffusion region during magnetic reconnection . Space Science Reviews 221 5 59 . APACrefDOI doi:10.1007/...

    2025 doi

  4. [12]

    , Bhattacharjee, D

    Ghuge2025 APACrefauthors Ghuge, D. , Bhattacharjee, D. \ Subramanian, P. APACrefauthors \ 2025 . Turbulent power: a discriminator between sheaths and CMEs Turbulent power: a discriminator between sheaths and CMEs . Solar Physics 300 4 47 . APACrefDOI doi:10.1007/s11207-025-024...

    2025 doi

  5. [13]

    Gingell2020 APACrefauthors Gingell, I. \ . APACrefauthors \ 2020 . Statistics of Reconnecting Current Sheets in the Transition Region of Earth's Bow Shock Statistics of reconnecting current sheets in the transition region of Earth's bow shock . Journal of Geophysical Research:...

    2020 doi

  6. [14]

    , Chuychai, P

    Greco2008 APACrefauthors Greco, A. , Chuychai, P. , Matthaeus, W H. , Servidio, S. \ Dmitruk, P. APACrefauthors \ 2008 . Intermittent MHD structures and classical discontinuities Intermittent MHD structures and classical discontinuities . Geophysical Research Letters 35 L19111...

    2008 doi

  7. [15]

    Grimes2022 APACrefauthors Grimes, E W. \ . APACrefauthors \ 2022 . The Space Physics Environment Data Analysis System in Python The space physics environment data analysis system in python . Frontiers in Astronomy and Space Sciences 9 1020815 . APACrefDOI doi:10.3389/fspas.202...

    2022

  8. [16]

    APACrefauthors \ 1962

    Harris1962 APACrefauthors Harris, E G. APACrefauthors \ 1962 . On a plasma sheath separating regions of oppositely directed magnetic field On a plasma sheath separating regions of oppositely directed magnetic field . Il Nuovo Cimento 23 1 115--121 . APACrefDOI doi:10.1007/BF02...

    1962 doi

  9. [17]

    Hasegawa2024 APACrefauthors Hasegawa, H. \ . APACrefauthors \ 2024 . Advanced Methods for Analyzing in-Situ Observations of Magnetic Reconnection Advanced methods for analyzing in-situ observations of magnetic reconnection . Space Science Reviews 220 6 68 . APACrefDOI doi:10.1...

    2024 doi

  10. [18]

    Hesse2020 APACrefauthors Hesse, M. \ . APACrefauthors \ 2020 . Magnetic reconnection in the space sciences: Past, present, and future Magnetic reconnection in the space sciences: Past, present, and future . Journal of Geophysical Research: Space Physics 125 2 e2018JA025935 . A...

    2020 doi

  11. [19]

    Hou2021 APACrefauthors Hou, C. , He, J. , Zhu, X. \ Wang, Y. APACrefauthors \ 2021 . Contribution of Magnetic Reconnection Events to Energy Dissipation in Space Plasma Turbulence Contribution of magnetic reconnection events to energy dissipation in space plasma turbulence . Th...

    2021 doi

  12. [20]

    Hubbert2022 APACrefauthors Hubbert, M. \ . APACrefauthors \ 2022 . Electron-Only Reconnection as a Transition Phase From Quiet Magnetotail Current Sheets to Traditional Magnetotail Reconnection Electron-only reconnection as a transition phase from quiet magnetotail current she...

    2022 doi

  13. [21]

    APACrefauthors \ 1941

    Kolmogorov1941 APACrefauthors Kolmogorov, A N. APACrefauthors \ 1941 . The local structure of turbulence in incompressible viscous fluid for very large Reynolds numbers The local structure of turbulence in incompressible viscous fluid for very large Reynolds numbers . Doklady ...

    1941

  14. [22]

    Lenouvel2021 APACrefauthors Lenouvel, F. \ . APACrefauthors \ 2021 . Identification of Electron Diffusion Regions with a Machine Learning Approach on MMS Data at the Earth's Magnetopause Identification of electron diffusion regions with a machine learning approach on MMS data ...

    2021 doi

  15. [23]

    \ Hutter, F

    Loshchilov2019 APACrefauthors Loshchilov, I. \ Hutter, F. APACrefauthors \ 2019 . Decoupled Weight Decay Regularization Decoupled weight decay regularization . International Conference on Learning Representations. International conference on learning representations. APACrefUR...

    2019

  16. [24]

    , Matthaeus, W H

    Osman2014 APACrefauthors Osman, K T. , Matthaeus, W H. , Gosling, J T. , Greco, A. , Servidio, S. , Hnat, B. Phan, T D. APACrefauthors \ 2014 . Magnetic reconnection and intermittent turbulence in the solar wind Magnetic reconnection and intermittent turbulence in the solar wi...

    2014 doi

  17. [25]

    , Shen, C

    Pang2022 APACrefauthors Pang, G. , Shen, C. , Cao, L. \ van den Hengel, A. APACrefauthors \ 2022 . Deep Learning for Anomaly Detection Deep learning for anomaly detection . ACM Computing Surveys 54 2 1--38 . APACrefDOI doi:10.1145/3439950 APACrefDOI

    2022 doi

  18. [26]

    Paschmann1979 APACrefauthors Paschmann, G. \ . APACrefauthors \ 1979 . Plasma acceleration at the Earth's magnetopause: evidence for reconnection Plasma acceleration at the Earth's magnetopause: evidence for reconnection . Nature 282 5736 243--246 . APACrefDOI doi:10.1038/2822...

    1979 doi

  19. [27]

    Phan2018 APACrefauthors Phan, T D. \ . APACrefauthors \ 2018 . Electron magnetic reconnection without ion coupling in Earth's turbulent magnetosheath Electron magnetic reconnection without ion coupling in earth's turbulent magnetosheath . Nature 557 7704 202--206 . APACrefDOI ...

    2018 doi

  20. [28]

    Pollock2016 APACrefauthors Pollock, C. \ . APACrefauthors \ 2016 . Fast Plasma Investigation for Magnetospheric Multiscale Fast plasma investigation for magnetospheric multiscale . Space Science Reviews 199 1--4 331--406 . APACrefDOI doi:10.1007/s11214-016-0245-4 APACrefDOI

    2016 doi

  21. [29]

    , Vandermeulen, R

    Ruff2018 APACrefauthors Ruff, L. , Vandermeulen, R. , G \"o rnitz, N. , Deecke, L. , Siddiqui, S A. , Binder, A. Kloft, M. APACrefauthors \ 2018 . Deep One-Class Classification Deep one-class classification . Proceedings of the 35th International Conference on Machine Learning...

    2018

  22. [30]

    o rnitz, N. , Binder, A. , M \

    Ruff2020 APACrefauthors Ruff, L. , Vandermeulen, R A. , G \"o rnitz, N. , Binder, A. , M \"u ller, E. , M \"u ller, K R. \ Kloft, M. APACrefauthors \ 2020 . Deep Semi-Supervised Anomaly Detection Deep semi-supervised anomaly detection . International Conference on Learning Rep...

    2020

  23. [31]

    Russell2016 APACrefauthors Russell, C T. \ . APACrefauthors \ 2016 . The Magnetospheric Multiscale Magnetometers The magnetospheric multiscale magnetometers . Space Science Reviews 199 1--4 189--256 . APACrefDOI doi:10.1007/s11214-014-0057-3 APACrefDOI

    2016 doi

  24. [32]

    , Platt, J C

    Scholkopf2001 APACrefauthors Sch \"o lkopf, B. , Platt, J C. , Shawe-Taylor, J. , Smola, A J. \ Williamson, R C. APACrefauthors \ 2001 . Estimating the support of a high-dimensional distribution Estimating the support of a high-dimensional distribution . Neural Computation 13 ...

    2001 doi

  25. [33]

    , Greco, A

    Servidio2011 APACrefauthors Servidio, S. , Greco, A. , Matthaeus, W H. , Osman, K T. \ Dmitruk, P. APACrefauthors \ 2011 . Statistical association of discontinuities and reconnection in magnetohydrodynamic turbulence Statistical association of discontinuities and reconnection ...

    2011 doi

  26. [34]

    \ Cahill, L J

    Sonnerup1967 APACrefauthors Sonnerup, B U \"O . \ Cahill, L J. APACrefauthors \ 1967 . Magnetopause structure and attitude from Explorer 12 observations Magnetopause structure and attitude from Explorer 12 observations . Journal of Geophysical Research 72 1 171--183 . APACrefD...

    1967 doi

  27. [35]

    \ Scheible, M

    Sonnerup1998 APACrefauthors Sonnerup, B U \"O . \ Scheible, M. APACrefauthors \ 1998 . Minimum and Maximum Variance Analysis Minimum and maximum variance analysis . G. Paschmann\ P W. Daly\ ( ), Analysis Methods for Multi-Spacecraft Data Analysis methods for multi-spacecraft d...

    1998

  28. [36]

    Stawarz2019 APACrefauthors Stawarz, J E. \ . APACrefauthors \ 2019 . Properties of the Turbulence Associated with Electron-Only Magnetic Reconnection in Earth's Magnetosheath Properties of the turbulence associated with electron-only magnetic reconnection in earth's magnetoshe...

    2019 doi

  29. [37]

    Stawarz2022 APACrefauthors Stawarz, J E. \ . APACrefauthors \ 2022 . Turbulence-driven magnetic reconnection and the magnetic correlation length: Observations from Magnetospheric Multiscale in Earth's magnetosheath Turbulence-driven magnetic reconnection and the magnetic corre...

    2022 doi

  30. [38]

    \ Duin, R P W

    TaxDuin2004 APACrefauthors Tax, D M J. \ Duin, R P W. APACrefauthors \ 2004 . Support vector data description Support vector data description . Machine Learning 54 1 45--66 . APACrefDOI doi:10.1023/B:MACH.0000008084.60811.49 APACrefDOI

    2004

  31. [39]

    Torbert2016 APACrefauthors Torbert, R B. \ . APACrefauthors \ 2016 . The FIELDS Instrument Suite on MMS : Scientific Objectives, Measurements, and Data Products The FIELDS instrument suite on MMS : Scientific objectives, measurements, and data products . Space Science Reviews ...

    2016 doi

  32. [40]

    Torbert2018 APACrefauthors Torbert, R B. \ . APACrefauthors \ 2018 . Electron-scale dynamics of the diffusion region during symmetric magnetic reconnection in space Electron-scale dynamics of the diffusion region during symmetric magnetic reconnection in space . Science 362 64...

    2018 doi

  33. [41]

    \ Tegmark, M

    UdrescuTegmark2020 APACrefauthors Udrescu, S M. \ Tegmark, M. APACrefauthors \ 2020 . AI Feynman : A physics-inspired method for symbolic regression AI Feynman : A physics-inspired method for symbolic regression . Science Advances 6 16 eaay2631 . APACrefDOI doi:10.1126/sciadv....

    2020 doi

  34. [42]

    , Slavin, J A

    Uritsky2011 APACrefauthors Uritsky, V M. , Slavin, J A. , Khazanov, G V. , Donovan, E F. , Boardsen, S A. , Anderson, B J. \ Korth, H. APACrefauthors \ 2011 . Kinetic-scale magnetic turbulence and finite Larmor radius effects at Mercury Kinetic-scale magnetic turbulence and fi...

    2011 doi

  35. [43]

    Vaivads2004 APACrefauthors Vaivads, A. \ . APACrefauthors \ 2004 . Structure of the magnetic reconnection diffusion region from four-spacecraft observations Structure of the magnetic reconnection diffusion region from four-spacecraft observations . Physical Review Letters 93 1...

    2004 doi

  36. [44]

    Waters2025 APACrefauthors Waters, C L. \ . APACrefauthors \ 2025 . Bridging in Situ Satellite Measurements and Simulations of Magnetic Reconnection Using Recurrent Neural Networks Bridging in situ satellite measurements and simulations of magnetic reconnection using recurrent ...

    2025 doi

  37. [45]

    , Kulsrud, R

    Yamada2010 APACrefauthors Yamada, M. , Kulsrud, R. \ Ji, H. APACrefauthors \ 2010 . Magnetic reconnection Magnetic reconnection . Reviews of Modern Physics 82 1 603--664 . APACrefDOI doi:10.1103/RevModPhys.82.603 APACrefDOI

    2010 doi

  38. [46]

    , V \"o r \"o s, Z

    Yordanova2020 APACrefauthors Yordanova, E. , V \"o r \"o s, Z. , Raptis, S. \ Karlsson, T. APACrefauthors \ 2020 . Current Sheet Statistics in the Magnetosheath Current sheet statistics in the magnetosheath . Frontiers in Astronomy and Space Sciences 7 2 . APACrefDOI doi:10.33...

    2020

  39. [47]

    \ Yamada, M

    Zweibel2009 APACrefauthors Zweibel, E G. \ Yamada, M. APACrefauthors \ 2009 . Magnetic reconnection in astrophysical and laboratory plasmas Magnetic reconnection in astrophysical and laboratory plasmas . Annual Review of Astronomy and Astrophysics 47 1 291--332 . APACrefDOI do...

    2009 doi

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.