pith. sign in

arxiv: 2606.02117 · v1 · pith:JJRM6QKInew · submitted 2026-06-01 · 📊 stat.ML · cs.LG· stat.ME

ProbRes: Volatility Learning for Probabilistic Time-Series Forecasting

Tingting Wang , Yunyi Zhang , Benyou Wang This is my paper

Pith reviewed 2026-06-28 12:43 UTC · model grok-4.3

classification 📊 stat.ML cs.LGstat.ME
keywords probabilistic forecastingtime seriesvolatilityheteroskedasticitypredictive distributionscalibrationresidual resamplinguncertainty quantification
0
0 comments X

The pith

ProbRes learns conditional volatility separately from the mean to produce calibrated predictive distributions for heteroskedastic time series.

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

The paper introduces ProbRes as a post-hoc method that adds probabilistic calibration to time-series forecasters by explicitly learning volatility dynamics. It trains two architecture-agnostic modules to capture the conditional mean and the conditional volatility separately, then forms predictive distributions at inference by resampling normalized residuals. The approach targets data with time-varying variance and non-Gaussian errors. A sympathetic reader would care because many applications, especially financial ones, require reliable uncertainty estimates rather than point forecasts alone. Theoretical results establish validity of the distributions, and experiments on synthetic and real-world data confirm accurate capture of predictive distributions along with well-calibrated intervals.

Core claim

ProbRes is a post-hoc probabilistic calibration method that employs two architecture-agnostic modules to separately model the conditional mean and conditional volatility during training. At inference it generates predictive distributions by resampling normalized residuals. The method applies to both univariate and multivariate time series and remains robust under a wide range of error distributions, including non-Gaussian innovations with conditional heteroskedasticity. Theoretical results demonstrate ProbRes's validity and experiments on both synthetic and real-world datasets show that ProbRes accurately captures predictive distributions and produces well-calibrated prediction intervals.

What carries the argument

Two architecture-agnostic modules that separately model conditional mean and conditional volatility, followed by resampling of normalized residuals to construct predictive distributions.

If this is right

  • The method can attach to any existing forecasting architecture because the modules are architecture-agnostic.
  • It extends directly to both univariate and multivariate time series.
  • It yields valid predictive distributions even when errors are non-Gaussian and exhibit conditional heteroskedasticity.
  • It produces well-calibrated prediction intervals on both synthetic and real-world datasets.

Where Pith is reading between the lines

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

  • The mean-volatility separation might transfer to other sequential prediction settings where variance dynamics matter, such as energy demand or traffic flow.
  • Combining the resampling step with modern neural forecasters could improve uncertainty estimates in settings with long-range volatility clustering.
  • In risk-sensitive domains the calibrated intervals could support more reliable downstream decisions like portfolio allocation without requiring changes to the base model.

Load-bearing premise

The assumption that separately modeling conditional mean and conditional volatility with architecture-agnostic modules, followed by resampling of normalized residuals, will produce valid predictive distributions under a wide range of error distributions including non-Gaussian innovations with conditional heteroskedasticity.

What would settle it

A dataset with strong conditional heteroskedasticity and non-Gaussian innovations where the prediction intervals produced by ProbRes fail to achieve nominal coverage rates, or a theoretical counter-example showing the resampling step does not yield valid distributions for some error process.

Figures

Figures reproduced from arXiv: 2606.02117 by Benyou Wang, Tingting Wang, Yunyi Zhang.

Figure 1
Figure 1. Figure 1: Kernel density estimates of the data (blue) and bootstrapped samples [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Histograms of the normalized fitted residuals [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Autocorrelation coefficients plot of the logarithm of square fitted [PITH_FULL_IMAGE:figures/full_fig_p021_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: 90% Prediction intervals generated by ProbRes with Dlinear backbone [PITH_FULL_IMAGE:figures/full_fig_p025_4.png] view at source ↗
read the original abstract

Probabilistic time series forecasting has attracted increasing attention in financial applications due to the need to quantify risk and uncertainty in future observations. We propose ProbRes, a post-hoc probabilistic calibration method that explicitly learns and incorporates volatility dynamics into probabilistic forecasting, enabling effective handling of heteroskedastic data. During training, ProbRes employs two architecture-agnostic modules to separately model the conditional mean and conditional volatility. At the inference stage, it generates predictive distributions by resampling normalized residuals. ProbRes is applicable to both univariate and multivariate time series and remains robust under a wide range of error distributions, including non-Gaussian innovations with conditional heteroskedasticity. Theoretical results demonstrate ProbRes's validity and experiments on both synthetic and real-world datasets show that ProbRes accurately captures predictive distributions and produces well-calibrated prediction intervals.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The manuscript proposes ProbRes, a post-hoc probabilistic calibration method for time-series forecasting. It employs two architecture-agnostic modules to separately model the conditional mean and conditional volatility during training, then generates predictive distributions at inference by resampling normalized residuals. The authors assert that this approach is theoretically valid and, via experiments on synthetic and real-world datasets, accurately captures predictive distributions while producing well-calibrated intervals even for non-Gaussian innovations with conditional heteroskedasticity; the method is claimed to apply to both univariate and multivariate series.

Significance. If the central claim holds, ProbRes would supply a flexible, model-agnostic route to incorporating volatility dynamics into existing forecasters, which is practically relevant for risk quantification in financial time series. The separation of mean and volatility modules plus residual resampling is a clean design choice that could avoid joint optimization difficulties, provided the resampling step is justified.

major comments (2)
  1. [Abstract] Abstract: The statement that 'theoretical results demonstrate ProbRes's validity' is not accompanied by any regularity conditions on the volatility estimator (or on the base forecaster) that would ensure the normalized residuals are approximately exchangeable. Without such conditions the resampling procedure cannot be guaranteed to recover the correct conditional law for non-Gaussian, conditionally heteroskedastic errors; this assumption is load-bearing for the central claim of robustness across error distributions.
  2. [Abstract (and implied theoretical section)] The manuscript provides no argument or rate result showing that the architecture-agnostic volatility module produces a consistent estimator at a rate sufficient for the normalized residuals to converge in distribution. Any misspecification in the volatility module therefore propagates directly into invalid predictive intervals, undermining the claim that the method remains valid under a wide range of error distributions.
minor comments (1)
  1. [Abstract] The abstract states applicability to multivariate series but supplies no indication of how the resampling step is extended when residuals are vector-valued.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and for identifying the need for explicit regularity conditions and consistency arguments to support the central theoretical claims. We agree that these elements are currently insufficient in the manuscript and will revise accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The statement that 'theoretical results demonstrate ProbRes's validity' is not accompanied by any regularity conditions on the volatility estimator (or on the base forecaster) that would ensure the normalized residuals are approximately exchangeable. Without such conditions the resampling procedure cannot be guaranteed to recover the correct conditional law for non-Gaussian, conditionally heteroskedastic errors; this assumption is load-bearing for the central claim of robustness across error distributions.

    Authors: We accept the point. The manuscript invokes residual resampling under the assumption that normalized residuals are approximately exchangeable once volatility is modeled, but does not state the required regularity conditions. In revision we will add an explicit subsection listing sufficient conditions on the volatility estimator (uniform consistency, moment bounds, and continuity of the conditional distribution) and on the base forecaster that justify the exchangeability argument for non-Gaussian heteroskedastic innovations. revision: yes

  2. Referee: [Abstract (and implied theoretical section)] The manuscript provides no argument or rate result showing that the architecture-agnostic volatility module produces a consistent estimator at a rate sufficient for the normalized residuals to converge in distribution. Any misspecification in the volatility module therefore propagates directly into invalid predictive intervals, undermining the claim that the method remains valid under a wide range of error distributions.

    Authors: We agree that a convergence-rate argument is absent. The existing theory treats the volatility estimate as given and focuses on the resampling step; it does not quantify how fast the volatility module must converge. In the revision we will insert a paragraph discussing the necessary rate (e.g., o_p(1) uniform consistency) and note that the separate, architecture-agnostic volatility module is intended to facilitate such rates, while acknowledging that full misspecification analysis would require additional technical work. revision: yes

Circularity Check

0 steps flagged

No circularity; derivation relies on standard residual resampling without self-referential reduction.

full rationale

The provided abstract and description outline a post-hoc method that separately models conditional mean and volatility via architecture-agnostic modules, then resamples normalized residuals to form predictive distributions. No equations, fitted parameters renamed as predictions, or self-citations are quoted that would make any claimed validity or calibration equivalent to the inputs by construction. The theoretical validity claim is presented as external to the fitting process itself, and the approach aligns with established statistical resampling techniques without evident tautology. This is the most common honest finding for papers whose central procedure does not reduce to a fit or self-citation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no equations, methods sections, or experimental details are provided to identify free parameters, axioms, or invented entities.

pith-pipeline@v0.9.1-grok · 5663 in / 1104 out tokens · 21661 ms · 2026-06-28T12:43:51.148735+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

85 extracted references · 3 canonical work pages · 2 internal anchors

  1. [1]

    Probabilistic forecasting.Annual Review of Statistics and Its Application, 1(Volume 1, 2014):125–151, 2014

    Tilmann Gneiting and Matthias Katzfuss. Probabilistic forecasting.Annual Review of Statistics and Its Application, 1(Volume 1, 2014):125–151, 2014

    2014

  2. [2]

    A neural stochas- tic volatility model.Proceedings of the AAAI Conference on Artificial Intelligence, 32(1), Apr

    Rui Luo, Weinan Zhang, Xiaojun Xu, and Jun Wang. A neural stochas- tic volatility model.Proceedings of the AAAI Conference on Artificial Intelligence, 32(1), Apr. 2018

    2018

  3. [3]

    Nam Nguyen and Brian Quanz. Temporal latent auto-encoder: A method for probabilistic multivariate time series forecasting.Proceedings of the AAAI Conference on Artificial Intelligence, 35(10):9117–9125, May 2021

    2021

  4. [4]

    Kejin Wu and Dimitris N. Politis. Bootstrap prediction inference of nonlinear autoregressive models.Journal of Time Series Analysis, 45(5):800–822, 2024

    2024

  5. [5]

    Probabilistic forecasting with generative networks via scoring rule minimization.Journal of Machine Learning Research, 25(45):1–64, 2024

    Lorenzo Pacchiardi, Rilwan A Adewoyin, Peter Dueben, and Ritabrata Dutta. Probabilistic forecasting with generative networks via scoring rule minimization.Journal of Machine Learning Research, 25(45):1–64, 2024

    2024

  6. [6]

    KooNPro: A variance- aware koopman probabilistic model enhanced by neural process for time series forecasting

    Ronghua Zheng, Hanru Bai, and Weiyang Ding. KooNPro: A variance- aware koopman probabilistic model enhanced by neural process for time series forecasting. InThe Thirteenth International Conference on Learning Representations, 2025

    2025

  7. [7]

    Analysis of financial time series.John Eiley and Sons, 2005

    Ruey S Tsay. Analysis of financial time series.John Eiley and Sons, 2005

    2005

  8. [8]

    Li Pan and Dimitris N. Politis. Bootstrap prediction intervals for linear, nonlinear and nonparametric autoregressions.Journal of Statistical Planning and Inference, 177:1–27, 2016

    2016

  9. [9]

    Learning quantile functions without quantile crossing for distribution-free time series forecasting

    Youngsuk Park, Danielle Maddix, François-Xavier Aubet, Kelvin Kan, Jan Gasthaus, and Yuyang Wang. Learning quantile functions without quantile crossing for distribution-free time series forecasting. In Gustau Camps-Valls, Francisco J. R. Ruiz, and Isabel Valera, editors,Proceedings of The 25th International Conference on Artificial Intelligence and Statis...

    2022

  10. [10]

    Nonparametric quantile regression with reLU-activated recurrent neural networks

    Hang Yu, Lyumin Wu, Wenxin Zhou, and Zhao Ren. Nonparametric quantile regression with reLU-activated recurrent neural networks. InThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2026

    2026

  11. [11]

    Predict, refine, synthe- size: Self-guiding diffusion models for probabilistic time series forecasting

    Marcel Kollovieh, Abdul Fatir Ansari, Michael Bohlke-Schneider, Jasper Zschiegner, Hao Wang, and Yuyang (Bernie) Wang. Predict, refine, synthe- size: Self-guiding diffusion models for probabilistic time series forecasting. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, 26 editors,Advances in Neural Information Processing Systems, ...

    2023

  12. [12]

    Maddix, Hao Wang, Michael W

    Abdul Fatir Ansari, Lorenzo Stella, Ali Caner Turkmen, Xiyuan Zhang, Pe- dro Mercado, Huibin Shen, Oleksandr Shchur, Syama Sundar Rangapuram, Sebastian Pineda Arango, Shubham Kapoor, Jasper Zschiegner, Danielle C. Maddix, Hao Wang, Michael W. Mahoney, Kari Torkkola, Andrew Gordon Wilson, Michael Bohlke-Schneider, and Bernie Wang. Chronos: Learning the lan...

  13. [13]

    Expert Certification

  14. [14]

    Robert F. Engle. Autoregressive conditional heteroscedasticity with esti- mates of the variance of united kingdom inflation.Econometrica, 50(4):987– 1007, 1982

    1982

  15. [15]

    Generalized autoregressive conditional heteroskedasticity

    Tim Bollerslev. Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3):307–327, 1986

    1986

  16. [16]

    Predictive quantile regressions under persistence and conditional heteroskedasticity.Journal of Econometrics, 213(1):261–280,

    Rui Fan and Ji Hyung Lee. Predictive quantile regressions under persistence and conditional heteroskedasticity.Journal of Econometrics, 213(1):261–280,

  17. [17]

    Annals: In Honor of Roger Koenker

  18. [18]

    Options and volatility.Economic Review, 81(3-6):21, 1996

    Peter A Abken and Saikat Nandi. Options and volatility.Economic Review, 81(3-6):21, 1996

    1996

  19. [19]

    McMurran, and Megan N

    Harindra De Silva, Gregory M. McMurran, and Megan N. Miller. 14 - diversification and the volatility risk premium. In Emmanuel Jurczenko, editor,Factor Investing, pages 365–387. Elsevier, 2017

    2017

  20. [20]

    Are transformers effective for time series forecasting?Proceedings of the AAAI Conference on Artificial Intelligence, 37(9):11121–11128, Jun

    Ailing Zeng, Muxi Chen, Lei Zhang, and Qiang Xu. Are transformers effective for time series forecasting?Proceedings of the AAAI Conference on Artificial Intelligence, 37(9):11121–11128, Jun. 2023

    2023

  21. [21]

    Cyclenet: Enhancing time series forecasting through modeling periodic patterns

    Shengsheng Lin, Weiwei Lin, Xinyi HU, Wentai Wu, Ruichao Mo, and Haocheng Zhong. Cyclenet: Enhancing time series forecasting through modeling periodic patterns. InThe Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024

    2024

  22. [22]

    B. Efron. Bootstrap methods: another look at the jackknife.Ann. Statist., 7(1):1–26, 1979

    1979

  23. [23]

    C. F. J. Wu. Jackknife, Bootstrap and Other Resampling Methods in Regression Analysis.The Annals of Statistics, 14(4):1261 – 1295, 1986

    1986

  24. [24]

    Robert A. Stine. Bootstrap prediction intervals for regression.Journal of the American Statistical Association, 80(392):1026–1031, 1985

    1985

  25. [25]

    A wild boot- strap for degenerate kernel tests

    Kacper Chwialkowski, Dino Sejdinovic, and Arthur Gretton. A wild boot- strap for degenerate kernel tests. In Z. Ghahramani, M. Welling, C. Cortes, N. Lawrence, and K.Q. Weinberger, editors,Advances in Neural Information Processing Systems, volume 27. Curran Associates, Inc., 2014. 27

    2014

  26. [26]

    Interval estimation for reinforcement- learning algorithms in continuous-state domains

    Martha White and Adam White. Interval estimation for reinforcement- learning algorithms in continuous-state domains. In J. Lafferty, C. Williams, J. Shawe-Taylor, R. Zemel, and A. Culotta, editors,Advances in Neural Information Processing Systems, volume 23. Curran Associates, Inc., 2010

    2010

  27. [27]

    Asymptotics of the bootstrap via stability with applications to inference with model selection

    Morgane Austern and Vasilis Syrgkanis. Asymptotics of the bootstrap via stability with applications to inference with model selection. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P.S. Liang, and J. Wortman Vaughan, editors, Advances in Neural Information Processing Systems, volume 34, pages 10705–10717. Curran Associates, Inc., 2021

    2021

  28. [28]

    Bootstrap AutoEncoders with contrastive paradigm for self-supervised gaze estimation

    Yaoming Wang, Jin Li, Wenrui Dai, Bowen Shi, Xiaopeng Zhang, Chenglin Li, and Hongkai Xiong. Bootstrap AutoEncoders with contrastive paradigm for self-supervised gaze estimation. In Ruslan Salakhutdinov, Zico Kolter, Katherine Heller, Adrian Weller, Nuria Oliver, Jonathan Scarlett, and Felix Berkenkamp, editors,Proceedings of the 41st International Confer...

    2024

  29. [29]

    Metamath: Bootstrap your own mathematical questions for large language models

    Longhui Yu, Weisen Jiang, Han Shi, Jincheng YU, Zhengying Liu, Yu Zhang, James Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu. Metamath: Bootstrap your own mathematical questions for large language models. In The Twelfth International Conference on Learning Representations, 2024

    2024

  30. [30]

    Conformalized quantile regression.Advances in neural information processing systems, 32, 2019

    Yaniv Romano, Evan Patterson, and Emmanuel Candes. Conformalized quantile regression.Advances in neural information processing systems, 32, 2019

    2019

  31. [31]

    Sequential predictive conformal inference for time series

    Chen Xu and Yao Xie. Sequential predictive conformal inference for time series. InProceedings of the 40th International Conference on Machine Learning, ICML’23. JMLR.org, 2023

    2023

  32. [32]

    Con- formal pid control for time series prediction.Advances in neural information processing systems, 36:23047–23074, 2023

    Anastasios Angelopoulos, Emmanuel Candes, and Ryan J Tibshirani. Con- formal pid control for time series prediction.Advances in neural information processing systems, 36:23047–23074, 2023

    2023

  33. [33]

    Confor- mal prediction for time series with modern hopfield networks

    Andreas Auer, Martin Gauch, Daniel Klotz, and Sepp Hochreiter. Confor- mal prediction for time series with modern hopfield networks. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors, Advances in Neural Information Processing Systems, volume 36, pages 56027–56074. Curran Associates, Inc., 2023

    2023

  34. [34]

    Conformalized time series with semantic features.Advances in Neural Information Processing Systems, 37:121449–121474, 2024

    Baiting Chen, Zhimei Ren, and Lu Cheng. Conformalized time series with semantic features.Advances in Neural Information Processing Systems, 37:121449–121474, 2024

    2024

  35. [35]

    Le, Timothy D

    Ichiro Takeuchi, Quoc V. Le, Timothy D. Sears, and Alexander J. Smola. Nonparametric quantile estimation.Journal of Machine Learning Research, 7(45):1231–1264, 2006. 28

    2006

  36. [36]

    Distributed high-dimensional quantile regression: estimation efficiency and support recovery

    Caixing Wang and Ziliang Shen. Distributed high-dimensional quantile regression: estimation efficiency and support recovery. InProceedings of the 41st International Conference on Machine Learning, ICML’24. JMLR.org, 2024

    2024

  37. [37]

    Au- toregressive denoising diffusion models for multivariate probabilistic time series forecasting

    Kashif Rasul, Calvin Seward, Ingmar Schuster, and Roland Vollgraf. Au- toregressive denoising diffusion models for multivariate probabilistic time series forecasting. In Marina Meila and Tong Zhang, editors,Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pages 8857–8868. PMLR, 1...

    2021

  38. [38]

    Multivariate probabilistic time series forecasting via conditioned normalizing flows

    Kashif Rasul, Abdul-Saboor Sheikh, Ingmar Schuster, Urs M Bergmann, and Roland Vollgraf. Multivariate probabilistic time series forecasting via conditioned normalizing flows. InInternational Conference on Learning Representations, 2021

    2021

  39. [39]

    Generative time series forecasting with diffusion, denoise, and disentanglement

    Yan Li, Xinjiang Lu, Yaqing Wang, and Dejing Dou. Generative time series forecasting with diffusion, denoise, and disentanglement. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, and A. Oh, editors,Advances in Neural Information Processing Systems, volume 35, pages 23009–23022. Curran Associates, Inc., 2022

    2022

  40. [40]

    Recurrent interpolants for probabilistic time series prediction

    Yu Chen, Marin Biloš, Sarthak Mittal, Wei Deng, Kashif Rasul, and Ander- son Schneider. Recurrent interpolants for probabilistic time series prediction. arXiv preprint arXiv:2409.11684, 2024

    work page arXiv 2024

  41. [41]

    Probabilistic forecasting with stochastic interpolants and föllmer processes

    Yifan Chen, Mark Goldstein, Mengjian Hua, Michael Samuel Albergo, Nicholas Matthew Boffi, and Eric Vanden-Eijnden. Probabilistic forecasting with stochastic interpolants and föllmer processes. In Ruslan Salakhutdinov, Zico Kolter, Katherine Heller, Adrian Weller, Nuria Oliver, Jonathan Scar- lett, and Felix Berkenkamp, editors,Proceedings of the 41st Inte...

    2024

  42. [42]

    Csdi: condi- tional score-based diffusion models for probabilistic time series imputation

    Yusuke Tashiro, Jiaming Song, Yang Song, and Stefano Ermon. Csdi: condi- tional score-based diffusion models for probabilistic time series imputation. InProceedings of the 35th International Conference on Neural Information Processing Systems, NIPS ’21, Red Hook, NY, USA, 2021. Curran Associates Inc

    2021

  43. [43]

    Transformer-modulated diffusion models for probabilistic multivariate time series forecasting

    Yuxin Li, Wenchao Chen, Xinyue Hu, Bo Chen, baolin sun, and Mingyuan Zhou. Transformer-modulated diffusion models for probabilistic multivariate time series forecasting. InThe Twelfth International Conference on Learning Representations, 2024

    2024

  44. [44]

    Flow matching with gaussian process priors for probabilistic 29 time series forecasting

    Marcel Kollovieh, Marten Lienen, David Lüdke, Leo Schwinn, and Stephan Günnemann. Flow matching with gaussian process priors for probabilistic 29 time series forecasting. InThe Thirteenth International Conference on Learning Representations, 2025

    2025

  45. [45]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors,Advances in Neural Information Processing Systems, volume 33, pages 6840–6851. Curran Associates, Inc., 2020

    2020

  46. [46]

    Score-Based Generative Modeling through Stochastic Differential Equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations.arXiv preprint arXiv:2011.13456, 2020

    work page internal anchor Pith review Pith/arXiv arXiv 2011

  47. [47]

    Auto-regressive moving diffusion models for time series forecasting.Proceedings of the AAAI Conference on Artificial Intelligence, 39(16):16727–16735, Apr

    Jiaxin Gao, Qinglong Cao, and Yuntian Chen. Auto-regressive moving diffusion models for time series forecasting.Proceedings of the AAAI Conference on Artificial Intelligence, 39(16):16727–16735, Apr. 2025

    2025

  48. [48]

    Deep state space models for time series forecasting

    Syama Sundar Rangapuram, Matthias W Seeger, Jan Gasthaus, Lorenzo Stella, Yuyang Wang, and Tim Januschowski. Deep state space models for time series forecasting. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, and R. Garnett, editors,Advances in Neural Information Processing Systems, volume 31. Curran Associates, Inc., 2018

    2018

  49. [49]

    Learning interpretable deep state space model for probabilistic time series forecasting

    Longyuan Li, Junchi Yan, Xiaokang Yang, and Yaohui Jin. Learning interpretable deep state space model for probabilistic time series forecasting. InProceedings of the 28th International Joint Conference on Artificial Intelligence, IJCAI’19, page 2901–2908. AAAI Press, 2019

    2019

  50. [50]

    End-to-end learning of coherent probabilistic forecasts for hierarchical time series

    Syama Sundar Rangapuram, Lucien D Werner, Konstantinos Benidis, Pedro Mercado, Jan Gasthaus, and Tim Januschowski. End-to-end learning of coherent probabilistic forecasts for hierarchical time series. In Marina Meila and Tong Zhang, editors,Proceedings of the 38th International Conference on Machine Learning, volume 139 ofProceedings of Machine Learning R...

    2021

  51. [51]

    Multi-scale attention flow for probabilistic time series forecasting.IEEE Trans

    Shibo Feng, Chunyan Miao, Ke Xu, Jiaxiang Wu, Pengcheng Wu, Yang Zhang, and Peilin Zhao. Multi-scale attention flow for probabilistic time series forecasting.IEEE Trans. on Knowl. and Data Eng., 36(5):2056–2068, May 2024

    2056

  52. [52]

    Probabilistic time series forecasting with shape and temporal diversity

    Vincent Le Guen and Nicolas Thome. Probabilistic time series forecasting with shape and temporal diversity. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors,Advances in Neural Information Processing Systems, volume 33, pages 4427–4440. Curran Associates, Inc., 2020

    2020

  53. [53]

    Proba- bilistic forecasting: A level-set approach

    Hilaf Hasson, Bernie Wang, Tim Januschowski, and Jan Gasthaus. Proba- bilistic forecasting: A level-set approach. In M. Ranzato, A. Beygelzimer, Y. Dauphin, P.S. Liang, and J. Wortman Vaughan, editors,Advances in Neural Information Processing Systems, volume 34, pages 6404–6416. Curran Associates, Inc., 2021. 30

    2021

  54. [54]

    Sutranets: Sub-series autoregressive networks for long-sequence, probabilistic forecasting

    Shane Bergsma, Tim Zeyl, and Lei Guo. Sutranets: Sub-series autoregressive networks for long-sequence, probabilistic forecasting. In A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors,Advances in Neural Information Processing Systems, volume 36, pages 30518–30533. Curran Associates, Inc., 2023

    2023

  55. [55]

    Autoregressive conditional heteroscedasticity with estimates of the variance of united kingdom inflation.Econometrica: Journal of the econometric society, pages 987–1007, 1982

    Robert F Engle. Autoregressive conditional heteroscedasticity with estimates of the variance of united kingdom inflation.Econometrica: Journal of the econometric society, pages 987–1007, 1982

    1982

  56. [56]

    HyperIV: Real-time implied volatility smoothing

    Yongxin Yang, Wenqi Chen, Chao Shu, and Timothy Hospedales. HyperIV: Real-time implied volatility smoothing. InForty-second International Con- ference on Machine Learning, 2025

    2025

  57. [57]

    Ni, Jun Pan, and Allen M

    Sophie X. Ni, Jun Pan, and Allen M. Poteshman. Volatility information trading in the option market.The Journal of Finance, 63(3):1059–1091, 2008

    2008

  58. [58]

    Volatility, the macroeconomy, and asset prices.The Journal of Finance, 69(6):2471–2511, 2014

    Ravi Bansal, Dana Kiku, Ivan Shaliastovich, and Amir Yaron. Volatility, the macroeconomy, and asset prices.The Journal of Finance, 69(6):2471–2511, 2014

    2014

  59. [59]

    Deep smoothing of the implied volatility surface

    Damien Ackerer, Natasa Tagasovska, and Thibault Vatter. Deep smoothing of the implied volatility surface. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors,Advances in Neural Information Processing Systems, volume 33, pages 11552–11563. Curran Associates, Inc., 2020

    2020

  60. [60]

    Arbitrage-free implied volatility surface generation with variational autoencoders.SIAM Journal on Financial Mathematics, 14(4):1004–1027, 2023

    Brian (Xin) Ning, Sebastian Jaimungal, Xiaorong Zhang, and Maxime Berg- eron. Arbitrage-free implied volatility surface generation with variational autoencoders.SIAM Journal on Financial Mathematics, 14(4):1004–1027, 2023

    2023

  61. [61]

    Operator deep smoothing for implied volatility

    Ruben Wiedemann, Antoine Jacquier, and Lukas Gonon. Operator deep smoothing for implied volatility. InThe Thirteenth International Conference on Learning Representations, 2025

    2025

  62. [62]

    Forecasting realized volatility with spillover effects: Perspectives from graph neural networks.International Journal of Forecasting, 41(1):377–397, 2025

    Chao Zhang, Xingyue Pu, Mihai Cucuringu, and Xiaowen Dong. Forecasting realized volatility with spillover effects: Perspectives from graph neural networks.International Journal of Forecasting, 41(1):377–397, 2025

    2025

  63. [63]

    Incorporating prior financial domain knowledge into neural networks for implied volatility surface predic- tion

    Yu Zheng, Yongxin Yang, and Bowei Chen. Incorporating prior financial domain knowledge into neural networks for implied volatility surface predic- tion. InProceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, KDD ’21, page 3968–3975, New York, NY, USA,

  64. [64]

    Association for Computing Machinery

  65. [65]

    Deepar: Probabilistic forecasting with autoregressive recurrent networks

    David Salinas, Valentin Flunkert, Jan Gasthaus, and Tim Januschowski. Deepar: Probabilistic forecasting with autoregressive recurrent networks. International Journal of Forecasting, 36(3):1181–1191, 2020. 31

    2020

  66. [66]

    Kejin Wu and Dimitris N. Politis. Scalable subsampling inference for deep neural networks.ACM / IMS J. Data Sci., 2(1), February 2025

    2025

  67. [67]

    Leveraging temporal dependency in probabilistic electric load forecasting.Applied Soft Computing, 169:112611, 2025

    Yaoli Zhang, Ye Tian, and Yunyi Zhang. Leveraging temporal dependency in probabilistic electric load forecasting.Applied Soft Computing, 169:112611, 2025

    2025

  68. [68]

    Asymptotic theory for a vector arma- garch model.Econometric Theory, 19(2):280–310, 2003

    Shiqing Ling and Michael McAleer. Asymptotic theory for a vector arma- garch model.Econometric Theory, 19(2):280–310, 2003

    2003

  69. [69]

    Non-stationary diffusion for probabilistic time series forecasting, 2025

    Weiwei Ye, Zhuopeng Xu, and Ning Gui. Non-stationary diffusion for probabilistic time series forecasting, 2025

    2025

  70. [70]

    Crossformer: Transformer utilizing cross- dimension dependency for multivariate time series forecasting

    Yunhao Zhang and Junchi Yan. Crossformer: Transformer utilizing cross- dimension dependency for multivariate time series forecasting. InThe Eleventh International Conference on Learning Representations, 2023

    2023

  71. [71]

    Bootstrap model aggregation for distributed statistical learning

    JUN HAN and Qiang Liu. Bootstrap model aggregation for distributed statistical learning. In D. Lee, M. Sugiyama, U. Luxburg, I. Guyon, and R. Garnett, editors,Advances in Neural Information Processing Systems, volume 29. Curran Associates, Inc., 2016

    2016

  72. [72]

    Small resamples, sharp guarantees: Convergence rates for resampled studentized quantile estimators

    Imon Banerjee and Sayak Chakrabarty. Small resamples, sharp guarantees: Convergence rates for resampled studentized quantile estimators. InThe Thirty-ninth Annual Conference on Neural Information Processing Systems, 2026

    2026

  73. [73]

    A time series is worth 64 words: Long-term forecasting with transformers

    Yuqi Nie, Nam H Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. A time series is worth 64 words: Long-term forecasting with transformers. In The Eleventh International Conference on Learning Representations, 2023

    2023

  74. [74]

    Timemixer: Decomposable multiscale mixing for time series forecasting

    Shiyu Wang, Haixu Wu, Xiaoming Shi, Tengge Hu, Huakun Luo, Lintao Ma, James Y Zhang, and JUN ZHOU. Timemixer: Decomposable multiscale mixing for time series forecasting. InInternational Conference on Learning Representations (ICLR), 2024

    2024

  75. [75]

    Estimating conditional quantiles with the help of the pinball loss.Bernoulli, 17(1):211 – 225, 2011

    Ingo Steinwart and Andreas Christmann. Estimating conditional quantiles with the help of the pinball loss.Bernoulli, 17(1):211 – 225, 2011

    2011

  76. [76]

    Online quantile regression

    Yinan Shen, Dong Xia, and Wen-Xin Zhou. Online quantile regression. Journal of Machine Learning Research, 26(231):1–55, 2025

    2025

  77. [77]

    High-dimensional multivariate forecasting with low-rank gaussian copula processes

    David Salinas, Michael Bohlke-Schneider, Laurent Callot, Roberto Medico, and Jan Gasthaus. High-dimensional multivariate forecasting with low-rank gaussian copula processes. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett, editors,Advances in Neural Infor- mation Processing Systems, volume 32. Curran Associates, Inc., 2019

    2019

  78. [78]

    Sampling-based multi- dimensional recalibration

    Youngseog Chung, Ian Char, and Jeff Schneider. Sampling-based multi- dimensional recalibration. InForty-first International Conference on Ma- chine Learning, 2024. 32

    2024

  79. [79]

    Skip sampling: subsampling in the frequency domain.Biometrika, 111(4):1241–1256, 08 2024

    Tucker McElroy and Dimitris N Politis. Skip sampling: subsampling in the frequency domain.Biometrika, 111(4):1241–1256, 08 2024

    2024

  80. [80]

    Springer Berlin Heidelberg, Berlin, Heidelberg, 2011

    Olimjon Shukurovich Sharipov.Glivenko-Cantelli Theorems, pages 612–614. Springer Berlin Heidelberg, Berlin, Heidelberg, 2011

    2011

Showing first 80 references.