pith. sign in

arxiv: 2606.31122 · v1 · pith:QFYZ7MH5new · submitted 2026-06-30 · 💱 q-fin.RM

Generating Plausible Stress Scenarios via Large Deviations

Anand Deo This is my paper

Pith reviewed 2026-07-01 03:17 UTC · model grok-4.3

classification 💱 q-fin.RM
keywords stress testinglarge deviationsscenario generationfinancial risk managementnetwork modelsextreme lossesrisk factor concentration
0
0 comments X

The pith

Financial stress scenarios can be generated systematically from sparse historical data by identifying the most likely risk factor configurations under large losses.

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

The paper develops a method to generate plausible stress scenarios for financial losses driven by exogenous risk factors by exploiting the large deviations principle. This principle states that when a large loss occurs the risk factors tend to concentrate near their most probable stress configurations. The method uses this concentration to define representative stress distributions and to extend observed samples to more extreme cases while keeping the relative plausibility of different stress mechanisms intact. Numerical tests on two financial network models show that the approach recovers the stressed loss law and key diagnostics even when standard generators produce no stressed samples at all.

Core claim

Conditional on a large loss the risk factors concentrate near the most likely stress configurations; this structure defines representative stress distributions and permits extrapolation of observed samples into more extreme scenarios while preserving relative plausibility of stress mechanisms, so that informative stress scenarios can be produced even when historical data contain few or no observations in the stressed regime.

What carries the argument

The large deviations principle applied to the loss model, which identifies the most likely configurations of the exogenous risk factors conditional on an extreme loss.

If this is right

  • The generated scenarios recover the stressed loss law in the two network models tested.
  • Key stress diagnostics are obtained in regimes where benchmark generators produce no stressed samples.
  • Extrapolation preserves the relative plausibility ordering among different stress mechanisms.
  • The procedure remains applicable when historical data are sparse or empty in the stressed regime.

Where Pith is reading between the lines

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

  • The same concentration property could be used to generate stress scenarios for other rare-event systems whose dynamics admit a large deviations description.
  • Real-time updating of the stress distributions might become feasible if new loss observations can be incorporated into the most-likely-configuration calculation.

Load-bearing premise

The financial loss model driven by exogenous risk factors satisfies the conditions for a large deviations principle to apply.

What would settle it

In a controlled simulation with a known loss model, if the scenarios generated by the method produce a conditional loss distribution that differs measurably from the distribution obtained by direct sampling under the large-loss condition, the accuracy claim would be falsified.

Figures

Figures reproduced from arXiv: 2606.31122 by Anand Deo.

Figure 1
Figure 1. Figure 1: Conditional distribution of risk factors given that the loss exceeds a high threshold. Algorithm 1 produces samples that more closely match the true stressed distribution than the Gaussian-copula benchmark, while remaining implementable at stress levels where a naive empirical exceedance resampler would produce no samples. Sample Complexity Reduction using Self-Structuring Transformations: A repre￾sentativ… view at source ↗
Figure 2
Figure 2. Figure 2: Normalized Wasserstein-1 error for the reinsurance-network loss at u = q0.99 (left) and u = q0.999 (middle), and sensitivity to the stretch parameter s at u = q0.999 with n = 1000 (right). Here the marginals are Weibull with αi = 1.2 [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: reports the corresponding normalized Wasserstein errors for the three topologies at u = q0.99. Across all three network structures, the extrapolation-based generator more accurately reproduces the reference stress law than either the Gaussian-copula benchmark or the naive empirical baseline. The former misses the relevant tail dependence, whereas the latter is limited by the small number of observed exceed… view at source ↗
Figure 4
Figure 4. Figure 4: Diagnostic recovery for the reinsurance network at u = q0.999. Left: distribution of the estimated conditional mean of GHHI across replications, with the reference value marked. Middle and right: pointwise 95% empirical bands for the conditional CCDFs of Gfrac and Gloss, with the reference CCDF overlaid. The value of s is chosen so that the self-structuring stress generator produces approximately 50 sample… view at source ↗
Figure 5
Figure 5. Figure 5: Diagnostic recovery for the fully connected clearing network at u = q0.999. Left: distribution of the estimated conditional mean of GHHI across replications, with the reference value marked. Middle and right: pointwise 95% empirical bands for the conditional CCDFs of Gfrac and Gloss, with the reference CCDF overlaid. Proof of Lemma 2.2: Let zt → z. Since ¯pj (x) = (a ⊺ jx − rj )+, we have t −1 p¯j (tzt) = … view at source ↗
read the original abstract

Financial stress tests based on handpicked scenarios can mislead risk management by overlooking genuinely dangerous configurations or overemphasising shocks that are too implausible to be decision-relevant. We develop a systematic method for generating plausible stress scenarios for financial losses driven by exogenous risk factors. The method exploits a large-deviations principle: conditional on a large loss, the risk factors concentrate near the most likely stress configurations. We use this structure to define representative stress distributions and to extrapolate observed samples into more extreme scenarios while preserving the relative plausibility of stress mechanisms. As a result, the procedure can generate informative stress scenarios even when historical data contain few or no observations in the stressed regime. Numerical experiments on two financial network models show that the method recovers the stressed loss law and key stress diagnostics, including in settings where benchmark generators fail to generate any stressed samples.

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 / 2 minor

Summary. The paper develops a method for generating plausible stress scenarios for financial losses driven by exogenous risk factors by exploiting a large-deviations principle: conditional on a large loss, the risk factors concentrate near the most likely stress configurations. This structure is used to define representative stress distributions and to extrapolate observed samples into more extreme scenarios while preserving relative plausibility. Numerical experiments on two financial network models are claimed to show that the method recovers the stressed loss law and key stress diagnostics, including in regimes where benchmark generators produce no stressed samples.

Significance. If the LDP applies and the numerical recovery holds, the approach supplies a systematic, extrapolation-capable alternative to handpicked or purely empirical stress generators, addressing the problem of data scarcity in extreme loss regimes. This could improve decision-relevance of stress tests in risk management by focusing on probabilistically plausible mechanisms rather than arbitrary shocks.

major comments (2)
  1. [Method and numerical experiments sections] The central claim—that informative scenarios can be generated even with zero observations in the stressed regime—rests on the loss functional L(X) obeying a large-deviations principle so that P(X | L(X) > u) concentrates on the rate-function minimizers as u → ∞. The manuscript invokes this structure but supplies neither analytic verification that the chosen network models satisfy the necessary conditions (exponential tightness, lower semicontinuity of the rate function) nor numerical diagnostics (e.g., empirical rate-function estimation or concentration plots for increasing thresholds). This assumption is load-bearing for the extrapolation result.
  2. [Numerical experiments] Table or figure reporting recovery metrics (e.g., the stressed loss law or key diagnostics): the experiments demonstrate performance conditional on the LDP already holding, but do not include explicit checks that the conditional law behaves as predicted by the LDP when the threshold is varied. Without such checks, the numerical results do not independently confirm applicability of the principle to the models.
minor comments (2)
  1. [Abstract] The abstract refers to 'two financial network models' without naming them or citing their definitions; the main text should make these identifications explicit at first use.
  2. [Method] Notation for the loss functional L(X) and the rate function should be introduced with a clear equation reference when first used, to aid readability for readers outside large-deviations theory.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major point below and will revise the manuscript accordingly to strengthen the presentation of the large-deviations assumptions and supporting diagnostics.

read point-by-point responses

  1. Referee: [Method and numerical experiments sections] The central claim—that informative scenarios can be generated even with zero observations in the stressed regime—rests on the loss functional L(X) obeying a large-deviations principle so that P(X | L(X) > u) concentrates on the rate-function minimizers as u → ∞. The manuscript invokes this structure but supplies neither analytic verification that the chosen network models satisfy the necessary conditions (exponential tightness, lower semicontinuity of the rate function) nor numerical diagnostics (e.g., empirical rate-function estimation or concentration plots for increasing thresholds). This assumption is load-bearing for the extrapolation result.

    Authors: We agree that the manuscript does not contain an analytic verification of the LDP conditions (exponential tightness, lower semicontinuity of the rate function) for the two network models, nor explicit numerical diagnostics such as concentration plots or empirical rate-function estimates. The paper develops the stress-generation procedure under the modeling assumption that an LDP holds for the loss functional, which is standard in this literature, but does not prove the conditions from first principles. To address the concern, the revised manuscript will include a dedicated discussion of the relevant LDP conditions together with numerical diagnostics (concentration plots for increasing thresholds) that can be computed from the existing simulation setup. revision: yes

  2. Referee: [Numerical experiments] Table or figure reporting recovery metrics (e.g., the stressed loss law or key diagnostics): the experiments demonstrate performance conditional on the LDP already holding, but do not include explicit checks that the conditional law behaves as predicted by the LDP when the threshold is varied. Without such checks, the numerical results do not independently confirm applicability of the principle to the models.

    Authors: We agree that the current numerical section does not vary the loss threshold u and report the resulting concentration behavior to provide an independent check of the LDP prediction. The existing experiments focus on recovery performance at a fixed high threshold. In the revision we will add experiments that systematically vary u, together with figures or tables documenting the empirical concentration of the conditional distributions around the rate-function minimizers, thereby supplying the requested diagnostic. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external LDP

full rationale

The paper's central construction applies the large deviations principle (LDP) to define representative stress distributions from the conditional concentration of risk factors given large losses. This LDP is invoked as a standard external theorem in probability theory rather than derived or fitted within the paper. No self-definitional steps, fitted inputs renamed as predictions, or load-bearing self-citations appear in the provided abstract or description. The numerical experiments test performance conditional on the LDP holding for the network models, but the derivation chain itself does not reduce to its inputs by construction and remains self-contained against external mathematical benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review; no explicit free parameters, invented entities, or additional axioms beyond the core large-deviations assumption are stated.

axioms (1)
  • domain assumption The loss model driven by exogenous risk factors obeys a large deviations principle that produces concentration on most likely stress configurations conditional on large loss.
    This is the mathematical foundation invoked to define representative stress distributions and to extrapolate samples.

pith-pipeline@v0.9.1-grok · 5658 in / 1156 out tokens · 46045 ms · 2026-07-01T03:17:28.630267+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

296 extracted references · 24 canonical work pages · 6 internal anchors

  1. [1]

    Management Science , volume=

    Machine learning and portfolio optimization , author=. Management Science , volume=. 2018 , publisher=

    2018

  2. [2]

    , author=

    State-dependent importance sampling for rare-event simulation: An overview and recent advances. , author=. Surveys in Operations Research and Management Science , volume=

  3. [3]

    Management Science , volume=

    Systemic risk in financial systems , author=. Management Science , volume=. 2001 , publisher=

    2001

  4. [4]

    European Journal of Operational Research , volume=

    Reverse sensitivity testing: What does it take to break the model? , author=. European Journal of Operational Research , volume=. 2019 , publisher=

    2019

  5. [5]

    Insurance: Mathematics and Economics , volume=

    Multivariate stress scenarios and solvency , author=. Insurance: Mathematics and Economics , volume=. 2012 , publisher=

    2012

  6. [6]

    Journal of Banking & Finance , volume=

    A systematic approach to multi-period stress testing of portfolio credit risk , author=. Journal of Banking & Finance , volume=. 2012 , publisher=

    2012

  7. [7]

    Scientific reports , volume=

    Reverse stress testing interbank networks , author=. Scientific reports , volume=. 2017 , publisher=

    2017

  8. [8]

    European Journal of Operational Research , volume=

    A theory of multivariate stress testing , author=. European Journal of Operational Research , volume=. 2024 , publisher=

    2024

  9. [9]

    Journal of Financial Intermediation , volume=

    Interbank tiering and money center banks , author=. Journal of Financial Intermediation , volume=. 2014 , publisher=

    2014

  10. [10]

    Journal of Financial Stability , volume=

    Multiplex interbank networks and systemic importance: An application to European data , author=. Journal of Financial Stability , volume=. 2018 , publisher=

    2018

  11. [11]

    American Economic Review , volume=

    Systemic risk and stability in financial networks , author=. American Economic Review , volume=. 2015 , publisher=

    2015

  12. [12]

    The Journal of Derivatives , volume=

    Robust risk estimation and hedging: A reverse stress testing approach , author=. The Journal of Derivatives , volume=. 2015 , publisher=

    2015

  13. [13]

    Mathematics of Operations Research , volume=

    Quantifying distributional model risk via optimal transport , author=. Mathematics of Operations Research , volume=. 2019 , publisher=

    2019

  14. [14]

    Mathematical Programming , volume=

    On distributionally robust chance constrained programs with Wasserstein distance , author=. Mathematical Programming , volume=. 2021 , publisher=

    2021

  15. [15]

    Journal of Optimization Theory and Applications , volume=

    Distributionally robust portfolio optimization under marginal and copula ambiguity , author=. Journal of Optimization Theory and Applications , volume=. 2024 , publisher=

    2024

  16. [16]

    Operations Research , volume=

    Distributionally robust optimization under distorted expectations , author=. Operations Research , volume=. 2025 , publisher=

    2025

  17. [17]

    Operations Research , volume=

    Optimal impact portfolios with general dependence and marginals , author=. Operations Research , volume=. 2024 , publisher=

    2024

  18. [18]

    European Journal of Operational Research , volume=

    60 years of portfolio optimization: Practical challenges and current trends , author=. European Journal of Operational Research , volume=. 2014 , publisher=

    2014

  19. [19]

    Management Science , volume=

    Tail risk and robust portfolio decisions , author=. Management Science , volume=. 2021 , publisher=

    2021

  20. [20]

    Stochastic Systems , volume=

    Efficient scenario generation for heavy-tailed chance constrained optimization , author=. Stochastic Systems , volume=. 2024 , publisher=

    2024

  21. [21]

    International statistical review , volume=

    The t copula and related copulas , author=. International statistical review , volume=. 2005 , publisher=

    2005

  22. [22]

    arXiv preprint arXiv:2506.16230 , year=

    On Design of Representative Distributionally Robust Formulations for Evaluation of Tail Risk Measures , author=. arXiv preprint arXiv:2506.16230 , year=

    work page arXiv

  23. [23]

    Portugaliae Mathematica , volume=

    A new class of semi-parametric estimators of the second order parameter , author=. Portugaliae Mathematica , volume=. 2003 , publisher=

    2003

  24. [24]

    Operations Research Letters , volume=

    On approximations of data-driven chance constrained programs over Wasserstein balls , author=. Operations Research Letters , volume=. 2023 , publisher=

    2023

  25. [25]

    Available at SSRN 4926070 , year=

    Characterizing the Minimum Cost to Achieve High Reliability via Asymptotic Analysis of Chance Constrained Programs , author=. Available at SSRN 4926070 , year=

  26. [26]

    The operations research revolution , pages=

    Data-driven stochastic programming using phi-divergences , author=. The operations research revolution , pages=. 2015 , publisher=

    2015

  27. [27]

    Journal of Applied Probability , volume=

    Robust Wasserstein profile inference and applications to machine learning , author=. Journal of Applied Probability , volume=. 2019 , publisher=

    2019

  28. [28]

    Mathematics of Operations Research , volume=

    Optimal transport-based distributionally robust optimization: Structural properties and iterative schemes , author=. Mathematics of Operations Research , volume=. 2022 , publisher=

    2022

  29. [29]

    Open Journal of Mathematical Optimization , volume=

    Frameworks and results in distributionally robust optimization , author=. Open Journal of Mathematical Optimization , volume=

  30. [30]

    Available at SSRN 557214 , year=

    The modelling of operational risk: experience with the analysis of the data collected by the Basel Committee , author=. Available at SSRN 557214 , year=

  31. [31]

    Econometrics and statistics , volume=

    A high quantile estimator based on the log-generalized Weibull tail limit , author=. Econometrics and statistics , volume=. 2018 , publisher=

    2018

  32. [32]

    2019 American Control Conference (ACC) , pages=

    Data-driven chance constrained optimization under Wasserstein ambiguity sets , author=. 2019 American Control Conference (ACC) , pages=. 2019 , organization=

    2019

  33. [33]

    Journal of Statistical Planning and Inference , volume=

    On the estimation of high quantiles , author=. Journal of Statistical Planning and Inference , volume=. 1993 , publisher=

    1993

  34. [34]

    Methodology and Computing in Applied Probability , volume=

    Properties of distortion risk measures , author=. Methodology and Computing in Applied Probability , volume=. 2009 , publisher=

    2009

  35. [35]

    Insurance: Mathematics and Economics , volume=

    Tail distortion risk and its asymptotic analysis , author=. Insurance: Mathematics and Economics , volume=. 2012 , publisher=

    2012

  36. [36]

    Mathematics of Operations Research , volume=

    Distributionally robust stochastic optimization with Wasserstein distance , author=. Mathematics of Operations Research , volume=. 2023 , publisher=

    2023

  37. [37]

    Operations research , volume=

    Constructing risk measures from uncertainty sets , author=. Operations research , volume=. 2009 , publisher=

    2009

  38. [38]

    Principles for Sound Liquidity Risk Management and Supervision (September 2008) , year=

    Basel committee on banking supervision , author=. Principles for Sound Liquidity Risk Management and Supervision (September 2008) , year=

    2008

  39. [39]

    Annual Review of Financial Economics , volume=

    Climate stress testing , author=. Annual Review of Financial Economics , volume=. 2023 , publisher=

    2023

  40. [40]

    1999 , institution=

    A coherent framework for stress-testing , author=. 1999 , institution=

    1999

  41. [41]

    Risk management: value at risk and beyond , volume=

    Stress testing in a value at risk framework , author=. Risk management: value at risk and beyond , volume=. 2002 , publisher=

    2002

  42. [42]

    JP Morgan, New York , volume=

    Creditmetrics-technical document , author=. JP Morgan, New York , volume=

  43. [43]

    International Journal of Forecasting , volume=

    Stress testing banks , author=. International Journal of Forecasting , volume=. 2014 , publisher=

    2014

  44. [44]

    Journal of Monetary Economics , volume=

    Testing macroprudential stress tests: The risk of regulatory risk weights , author=. Journal of Monetary Economics , volume=. 2014 , publisher=

    2014

  45. [45]

    International Journal of Central Banking , year=

    How to Find Plausible, Severe, and Useful Stress Scenarios , author=. International Journal of Central Banking , year=

  46. [46]

    Quantitative Finance , volume=

    Stress scenario selection by empirical likelihood , author=. Quantitative Finance , volume=. 2015 , publisher=

    2015

  47. [47]

    Operations Research , volume=

    Achieving efficiency in black-box simulation of distribution tails with self-structuring importance samplers , author=. Operations Research , volume=. 2025 , publisher=

    2025

  48. [48]

    Top , volume=

    Robust portfolio choice with CVaR and VaR under distribution and mean return ambiguity , author=. Top , volume=. 2014 , publisher=

    2014

  49. [49]

    arXiv preprint arXiv:2206.12835 , year=

    Combining Retrospective Approximation with Importance Sampling for Optimising Conditional Value at Risk , author=. arXiv preprint arXiv:2206.12835 , year=

    work page arXiv

  50. [50]

    2013 , publisher=

    Convergence of probability measures , author=. 2013 , publisher=

    2013

  51. [51]

    2013 , publisher=

    Introduction to rare event simulation , author=. 2013 , publisher=

    2013

  52. [52]

    Portfolio Selection , urldate =

    Harry Markowitz , journal =. Portfolio Selection , urldate =

  53. [53]

    Quantitative Finance , volume=

    Robust risk measurement and model risk , author=. Quantitative Finance , volume=. 2014 , publisher=

    2014

  54. [54]

    Operations Research , volume=

    Recovering best statistical guarantees via the empirical divergence-based distributionally robust optimization , author=. Operations Research , volume=. 2019 , publisher=

    2019

  55. [55]

    Management Science , volume=

    Robust solutions of optimization problems affected by uncertain probabilities , author=. Management Science , volume=. 2013 , publisher=

    2013

  56. [56]

    Mathematics of Operations Research , volume=

    Statistics of robust optimization: A generalized empirical likelihood approach , author=. Mathematics of Operations Research , volume=. 2021 , publisher=

    2021

  57. [57]

    Quantitative Finance , volume=

    Portfolio optimization under expected shortfall: contour maps of estimation error , author=. Quantitative Finance , volume=. 2018 , publisher=

    2018

  58. [58]

    Advances in Neural Information Processing Systems , pages=

    Risk-sensitive and robust decision-making: a cvar optimization approach , author=. Advances in Neural Information Processing Systems , pages=

  59. [59]

    2007 , publisher=

    Extreme value theory: an introduction , author=. 2007 , publisher=

    2007

  60. [60]

    Stochastic Processes and their Applications , volume=

    On regular variation of probability densities , author=. Stochastic Processes and their Applications , volume=. 1987 , publisher=

    1987

  61. [61]

    Mathematical Finance , volume=

    Portfolio value-at-risk with heavy-tailed risk factors , author=. Mathematical Finance , volume=. 2002 , publisher=

    2002

  62. [62]

    Operations Research Letters , volume=

    Asymptotic representations for importance-sampling estimators of value-at-risk and conditional value-at-risk , author=. Operations Research Letters , volume=. 2010 , publisher=

    2010

  63. [63]

    Management Science , volume=

    Simulating sensitivities of conditional value at risk , author=. Management Science , volume=. 2009 , publisher=

    2009

  64. [64]

    Journal of Risk , volume=

    Portfolio optimization with conditional value-at-risk objective and constraints , author=. Journal of Risk , volume=. 2002 , publisher=

    2002

  65. [65]

    Operations Research , volume=

    A probabilistic model for minmax regret in combinatorial optimization , author=. Operations Research , volume=. 2014 , publisher=

    2014

  66. [66]

    Quantitative Finance , volume=

    Robust portfolio techniques for mitigating the fragility of CVaR minimization and generalization to coherent risk measures , author=. Quantitative Finance , volume=. 2013 , publisher=

    2013

  67. [67]

    Operations research , volume=

    Worst-case conditional value-at-risk with application to robust portfolio management , author=. Operations research , volume=. 2009 , publisher=

    2009

  68. [68]

    Operations Research Letters , volume=

    Conditional value-at-risk in portfolio optimization: Coherent but fragile , author=. Operations Research Letters , volume=. 2011 , publisher=

    2011

  69. [69]

    Finance and Stochastics , volume=

    On optimal portfolio diversification with respect to extreme risks , author=. Finance and Stochastics , volume=. 2010 , publisher=

    2010

  70. [70]

    2015 , publisher=

    Quantitative risk management: concepts, techniques and tools-revised edition , author=. 2015 , publisher=

    2015

  71. [71]

    2007 , publisher=

    Heavy-tail phenomena: probabilistic and statistical modeling , author=. 2007 , publisher=

    2007

  72. [72]

    2013 , publisher=

    Extreme values, regular variation and point processes , author=. 2013 , publisher=

    2013

  73. [73]

    Journal of risk , volume=

    Optimization of conditional value-at-risk , author=. Journal of risk , volume=

  74. [74]

    Extremes , volume=

    Multivariate peaks over thresholds models , author=. Extremes , volume=. 2018 , publisher=

    2018

  75. [75]

    Mathematical Finance: An International Journal of Mathematics, Statistics and Financial Economics , volume=

    Nonparametric estimation and sensitivity analysis of expected shortfall , author=. Mathematical Finance: An International Journal of Mathematics, Statistics and Financial Economics , volume=. 2004 , publisher=

    2004

  76. [76]

    Copula Theory and Its Applications: Proceedings of the Workshop Held in Warsaw, 25-26 September 2009 , pages=

    Extreme-value copulas , author=. Copula Theory and Its Applications: Proceedings of the Workshop Held in Warsaw, 25-26 September 2009 , pages=. 2010 , organization=

    2009

  77. [77]

    Mathematics of Operations Research , volume=

    Optimization of convex risk functions , author=. Mathematics of Operations Research , volume=. 2006 , publisher=

    2006

  78. [78]

    Insurance: Mathematics and Economics , volume=

    Insurance pricing and increased limits ratemaking by proportional hazards transforms , author=. Insurance: Mathematics and Economics , volume=. 1995 , publisher=

    1995

  79. [79]

    Stochastic models , volume=

    Risk measures and comonotonicity: a review , author=. Stochastic models , volume=. 2006 , publisher=

    2006

  80. [80]

    arXiv preprint arXiv:2212.05716 , year=

    On generalization and regularization via wasserstein distributionally robust optimization , author=. arXiv preprint arXiv:2212.05716 , year=

    work page arXiv

Showing first 80 references.