XGBoost: A Scalable Tree Boosting System
read the original abstract
Tree boosting is a highly effective and widely used machine learning method. In this paper, we describe a scalable end-to-end tree boosting system called XGBoost, which is used widely by data scientists to achieve state-of-the-art results on many machine learning challenges. We propose a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning. More importantly, we provide insights on cache access patterns, data compression and sharding to build a scalable tree boosting system. By combining these insights, XGBoost scales beyond billions of examples using far fewer resources than existing systems.
This paper has not been read by Pith yet.
Forward citations
Cited by 39 Pith papers
-
Observation of the charmless purely baryonic decay $\mathinner{\mathit{\Lambda}^0_b\!\to \mathit{\Lambda} p \overline{p}}$
First observation of Λ_b^0 → Λ p p-bar with 5.1σ significance and relative branching fraction (5.1 ± 1.3(stat) ± 0.3(syst)) × 10^{-2} to the reference mode Λ_b^0 → Λ K^+ K^-.
-
A catalogue of TeV pulsar environments
A homogenized catalogue of 128 TeV sources associated with 66 pulsars, covering all evolutionary stages and showing only weak correlation between luminosity and pulsar age.
-
A catalogue of TeV pulsar environments
A homogenized catalogue of 128 TeV sources linked to 66 pulsars shows weak correlation between TeV luminosity and pulsar age, indicating environmental effects dominate evolution.
-
A satellite foundation model for improved wealth monitoring
Tempov is a self-supervised satellite foundation model that predicts wealth levels and decadal changes at high resolution across Africa from Landsat imagery, outperforming baselines even with limited labels and genera...
-
SynQL: A Controllable and Scalable Rule-Based Framework for SQL Workload Synthesis for Performance Benchmarking
SynQL synthesizes diverse, execution-ready SQL workloads by deterministically traversing foreign-key graphs to populate ASTs, yielding high topological entropy and cost-model training data with R² ≥ 0.79 on held-out sets.
-
A Perfect Storm: First-Nature Geography and Economic Development
A 1825 storm created a new sea connection in Denmark, producing a 27 percent population increase (elasticity 1.6 to market access) driven by fertility and occupational change toward fishing and manufacturing, with sym...
-
MARD: Mirror-Augmented Reasoning Distillation for Mechanism-Level Drug-Drug Interaction Prediction
MARD-7B outperforms baselines and GPT-4o on novel drug pairs for mechanism-level DDI prediction via a new distillation pipeline with verifiable process rewards and releases all resources.
-
VarWISE: Infrared Variability via NEOWISE Single Exposure Photometry
VarWISE catalog identifies 457,080 high-confidence infrared variables (49.81% new) and an extended set of 1.9 million from NEOWISE photometry via spatial clustering, VARnet detection, and XGBoost classification.
-
Identifying Changing-Look AGN Transitions in Light Curve Data with the Zwicky Transient Facility
A criterion of |Δg| > 0.4 mag and |Δ(g-r)| > 0.2 mag detects photometric CL-AGN transitions in 9.6% of known hosts with 1.6% false positive rate from simulations.
-
MuViS: Multimodal Virtual Sensing Benchmark
MuViS is a new unified benchmark showing that neither gradient-boosted trees nor deep neural networks hold a universal advantage in multimodal virtual sensing.
-
Filtering Interlopers with Photometry and Diagnostic Features: A Machine Learning Framework Validated with CSST Slitless Spectroscopy
XGBoost classifier filters interlopers in CSST slitless spectroscopy simulations, retaining 42% of galaxies with 96.6% accurate redshifts and 0.13% outliers.
-
Search for $K_{\mathrm{S(L)}}^{0} \rightarrow \pi^{+}\pi^{-}\mu^{+}\mu^{-}$ decays at LHCb
No evidence for KS0 or KL0 to pi+ pi- mu+ mu- decays; first upper limits set at 1.4e-9 and 6.6e-7 (90% CL).
-
Observation of $W^{+}W^{-}\gamma$ production in $pp$ collisions at $\sqrt{s}$ = 13 TeV with the ATLAS detector and constraints on anomalous quartic gauge-boson couplings
ATLAS reports the observation of W⁺W⁻γ triboson production with 5.9σ significance and derives constraints on anomalous quartic gauge-boson couplings via effective field theory.
-
Measurement of charged-current muon neutrino-argon interactions without pions in the final state using the MicroBooNE detector
Reports flux-integrated single- and double-differential cross sections for νμ CC0π interactions on argon and compares them to neutrino event generators, finding good agreement in single-differential but only partial a...
-
Inclusive Search for Anomalous Single-Photon Production in MicroBooNE
MicroBooNE performs a blind search for anomalous single-photon events and finds agreement overall but a ~2σ excess in the proton-free subsample at energies below 600 MeV.
-
Predicting gestational age at birth in the context of preterm birth from multi-modal fetal MRI
ML pipeline on fetal MRI features predicts gestational age at birth (R²=0.13, MAE=2.74 weeks) and preterm status (accuracy 0.77) from 426 cases.
-
ChronoSurv: A Clinical Pathway-Guided Graph Framework for Multimodal Survival Analysis
ChronoSurv proposes a directed graph framework for multimodal survival analysis that models clinical pathways with hierarchical and heterogeneous message passing.
-
Stellar Obliquities of Young Systems, Atmospheres Undergoing Contraction and Escape (SOYSAUCE) II: a 135 Myr planet on an aligned orbit with transit timing variations
Validation of a 135 Myr, 3.6 R_E transiting planet with aligned obliquity and TTV evidence for a near-resonant companion.
-
Policy-Guided ML for Energy Savings: Cell On/Off Switching under Operator QoS Constraints in Real 5G Networks
ML-based cell on/off strategy for 5G networks trained on real European operator data, using class ratio tuning to enforce QoS policies on throughput and outage while achieving energy savings.
-
Classification of IED-free EEG Responses for Assisted Epilepsy Diagnosis
A stacked ensemble combining temporal, spectral, wavelet and connectivity features from IED-free stimulation EEG reaches up to 97.8% AUC on the TUH corpus and 79.4% AUC on a clinical cohort.
-
Search for pair production of additional neutral scalars within the Inert Doublet Model in a final state with two electrons or two muons in proton-proton collisions at $\sqrt{s}$ = 13 TeV and 13.6 TeV
No significant excess found; new exclusion limits reach m_H = 108 GeV for m_H - m_A = 78 GeV in the Inert Doublet Model.
-
From Gaia to GaiaNIR: II. A new view of the Milky Way bar
Gaia DR3 data shows the Milky Way bar pattern speed is biased high by 14.4 km s^{-1} kpc^{-1}, with a bias-corrected estimate of 29.3 ± 2.3 km s^{-1} kpc^{-1}.
-
Enhanced sensitivity to the $H \to Z\gamma \to \ell^+\ell^-\gamma$ decay at the LHC using machine learning and novel kinematic observables
New correlated observables from the (P_Higgs, θ_Zγ) plane with XGBoost improve H→Zγ signal discrimination from Z/γ* background, raising S/B to 2.1% (electrons) and 3.4% (muons) near the Higgs mass.
-
Stellar flare detection in XMM-Newton with gradient boosted trees
A gradient boosted classifier on X-ray light curve features detects stellar flares at 97.1% test accuracy and generates the largest public catalog of such events.
-
Electromagnetic Shower Reconstruction and Identification in FASER's Emulsion Detector for LHC Forward Neutrino Measurements
Clustering-based reconstruction and multi-level identification (pre-selection, cuts, BDT) achieve 99.9% background rejection, 59-71% efficiency, and 23-25% energy resolution for EM showers in FASERnu emulsion detector...
-
Optical-morphology-based assessment of astrometric quality in Gaia-CRF3 quasars
A multi-survey composite morphology score for Gaia-CRF3 sources reveals a -43.7 μas parallax zero-point shift from extended AGNs and a 15.8% reduction in frame spin for the highest-purity point-source subsample.
-
Analysis of the $C\!P$ structure of the Yukawa coupling between the Higgs boson and tau leptons in proton-proton collisions at $\sqrt{s}$ = 13.6 TeV
Combined CMS result gives α^{Hττ} = 7 ± 16° for the CP mixing angle in H→ττ, consistent with SM expectation of 0 ± 14°.
-
Predicting Redshift in Seyfert Galaxies Using Machine Learning
Random Forest regression on combined optical plus mid-infrared colors yields NMAD of 0.0188, R-squared of 0.9561, and 0.294 percent outliers for photometric redshifts in 23,797 Seyfert II galaxies selected from SDSS and WISE.
-
AMO-ENE: Attention-based Multi-Omics Fusion Model for Outcome Prediction in Extra Nodal Extension and HPV-associated Oropharyngeal Cancer
An attention-based fusion model combining semi-supervised CT segmentation, radiomics, and clinical features predicts metastatic recurrence, overall survival, and disease-free survival in HPV+ oropharyngeal cancer with...
-
Exotic Higgs Decays at a Muon Collider
Muon colliders at 3 TeV and 10 TeV can probe branching ratios for h to SS decays in 4b and 2b2μ channels down to 10^{-3}–10^{-5}, improving on HL-LHC projections using machine learning.
-
Probing Heavy Neutral Higgs Bosons via Single Vector-Like Bottom Quark Production at the HL-LHC
XGBoost multivariate analysis extends the 5-sigma discovery reach for singly produced vector-like bottom quarks decaying via heavy neutral Higgs bosons to 1.6 TeV at the HL-LHC with 3 ab^{-1}.
-
Study of Higgs boson pair production in the $HH \rightarrow b \overline{b} \gamma \gamma$ final state with 308 fb$^{-1}$ of data collected at $\sqrt{s} =$ 13 TeV and 13.6 TeV by the ATLAS experiment
Updated ATLAS search for HH → bbγγ with 308 fb⁻¹ yields observed μ_HH = 0.9^{+1.4}_{-1.1}, 95% CL limit μ_HH < 3.7, and κ_λ in [-1.6, 6.6].
-
Weibull-Stationary Stochastic Differential Equations for Conditional Long-Horizon Wind Power Forecasting
A conditional probabilistic framework using monthly Weibull parameters forecasted by Kalman filter on VAR(1), three Weibull-stationary SDE models, and XGBoost power curve mapping achieves CRPS of 1.57 m/s and low Wass...
-
From Network Experience to Subscriber Retention: An Explainable AI Framework for Mobile Operators
Framework for telco churn prediction using XAI shows QoE indicators provide stronger signals than network counters on real operator data.
-
Search for a new heavy scalar resonance decaying to a pair of Z bosons in the four-lepton final state in proton-proton collisions at $\sqrt{s}$ = 13 TeV
No significant excess observed in CMS four-lepton data; 95% CL upper limits set on heavy scalar resonance production cross section times BR(ZZ) ranging 0.05-0.1 pb at low mass to 0 pb at high mass.
-
7 Tesla Quantitative MRI and Machine Learning for Exploratory Motor Subtype Stratification and Diagnosis in Parkinson's Disease
7T qMRI with DL segmentation and feature-selected ML reached 82% accuracy for PD vs HC, 100% for PIGD vs TD, and 73% multiclass on 45 subjects.
-
Prospects for Measuring $H\to \rm{invisble}$ at the FCCee
FCC-ee at 240 GeV with 10.8 ab^{-1} could set a 95% CL upper limit of 0.15% on the branching ratio of Higgs to invisible particles.
-
Probing the electron Yukawa coupling via resonant Higgs boson production at FCC-ee via $e^+e^- \to H \to WW^*$ in lepton-plus-jets final states
Simulation projects 2.0 sigma significance for resonant Higgs production at FCC-ee, yielding an upper limit of kappa_e less than or equal to 1.35 at 95% CL on the electron Yukawa coupling modifier with 10 ab inverse l...
-
Machine learning study on single production of a singlet vectorlike lepton at the Large Hadron Collider
XGBoost machine learning improves discrimination in LHC searches for singlet vector-like leptons, yielding projected 2σ mass exclusion limits of 620 GeV (three-lepton) and 490 GeV (four-lepton) at 14 TeV with 3000 fb^{-1}.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.