Pith. sign in

REVIEW 1 major objections 5 minor 25 references

Epistemic and aleatoric uncertainty measures are not primitives to be axiomatized; they are consequences of choosing a strictly proper loss and decomposing an expected subjective risk.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 23:51 UTC pith:5SOS3DGD

load-bearing objection A genuinely unifying framework for UQ via subjective risk decomposition, with clean math behind it—but the epistemic/aleatoric labeling is a semantic choice, not a derivation, so the strongest claims need tempering. the 1 major comments →

arxiv 2607.15196 v1 pith:5SOS3DGD submitted 2026-07-16 stat.ML cs.AIcs.LG

Subjective Risk Decomposition: A New View for Uncertainty Quantification

Raghad Alamri , Michele Caprio , Gavin Brown This is my paper
classification stat.ML cs.AIcs.LG
keywords epistemic uncertaintyaleatoric uncertaintysubjective riskBregman divergencestrictly proper lossbias–variance decompositionmutual informationuncertainty quantification
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Uncertainty quantification frequently presents epistemic and aleatoric uncertainty as concepts that need separate axiom systems. This paper instead claims they emerge from a single higher-level modeling decision: choose a strictly proper loss and evaluate the agent's subjective risk at the true distribution; decompose that expected risk and the variance term is epistemic uncertainty, the expected generalized entropy is aleatoric. The same recipe reproduces the field's leading measures — mutual information, label-wise, variance-based, ensemble, and data-versus-procedural — so their disagreements are seen as differences in loss geometry, not contradictions. If this is right, UQ acquires a common theoretical foundation and a principled way to design new measures. The paper extends the same subjective-risk view to learning theory, linking the terms to excess risk, approximation error, and estimation error.

Core claim

The central claim is that epistemic and aleatoric uncertainty are not quantities to be defined by hand; they are terms induced by a higher-level modeling decision. For any strictly proper loss ℓ, the expected subjective risk Eθ[E_{Y∼qθ} ℓ(p, Y)] decomposes as Bφ(q̄, p) + Eθ[Bφ(qθ, q̄)] + Eθ[Hℓ(qθ)], where Bφ is the divergence generated by the loss and q̄ is its right Bregman centroid (the aggregate predictive belief). The second term, the loss-induced Bregman variance, is identified as epistemic uncertainty; the third, the expected generalized entropy, as aleatoric uncertainty. For the log loss on categorical variables this reduces to the reverse cross-entropy and exactly recovers the classi

What carries the argument

The central object is the Bregman divergence generated by a strictly proper loss through the standard representation of such losses as a convex potential plus an inner-product term — call it the loss-induced divergence. The argument runs on the reverse-argument bias–variance decomposition around the right Bregman centroid q̄ = argmin_z Eθ[Bφ(qθ, z)], the aggregate predictive belief. The divergence is the only loss class with a clean bias–variance split, and its asymmetry separates the 'forward' objective risk from the 'reverse' subjective risk. For the log loss the divergence is the KL divergence, the centroid is the mixture distribution, and the variance term becomes mutual information; for

Load-bearing premise

The load-bearing identification — that the Bregman variance term is epistemic uncertainty and the expected generalized entropy is aleatoric — is a semantic labeling imposed on the decomposition, not a mathematical consequence of it; Theorem 2 alone is an identity about bias, variance, and entropy.

What would settle it

Measure the Bregman variance term Eθ[Bφ(qθ, q̄)] in a fixed model class while the training set grows, under any strictly proper loss. If this 'epistemic' term ever increases while held-out accuracy and calibration improve, the claim that the variance term is reducible uncertainty — uncertainty that diminishes with data — fails in a directly observable way.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • The classic mutual-information/conditional-entropy A/E split, label-wise uncertainty, variance-based regression terms, deep-ensemble terms, and data-versus-procedural decompositions all arise from one expected-subjective-risk identity, so they stop being rival definitions.
  • Disagreements among entropy-based, variance-based, and divergence-based measures are explained by the curvature of the loss's potential near the simplex boundary, not by one measure being wrong.
  • For log loss the mutual information is exactly one component of the subjective estimation error, so the standard epistemic measure systematically under-estimates the reducible part of subjective risk.
  • The subjective best-in-class model equals the objective one only when the projection of the aggregate predictive belief matches the projection of the true distribution; consistency of the predictive mixture restores this agreement asymptotically.
  • Practically, new uncertainty measures are obtained by choosing a strictly proper loss and reading off the variance and entropy terms of the decomposition.

Where Pith is reading between the lines

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

  • If the paper is right, the debate over which A/E measure to use becomes a debate over which loss geometry to use; the key design question is how sensitive the loss's potential must be near probability-zero events, and that can be tuned.
  • The reverse-KL connection gives a principled lens for LLM hallucination scores: the expected subjective risk at the true distribution penalizes confident mass in implausible regions, and this suggests loss-specific 'realism' scores beyond log loss that could be tested on hallucination benchmarks.
  • The paper treats the bias term as external grounding rather than reducible uncertainty; a natural extension is to check whether the Bregman variance term tracks dataset size only when the model class is fixed, while the bias term tracks misspecification.
  • Because the subjective approximation gap can be negative for restricted hypothesis classes, the common 'total uncertainty = AU + EU' equation is incomplete: the additive view is exactly a de-biased expected subjective risk, and empirical UQ users might gain by reporting all three terms rather than two.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 5 minor

Summary. The paper proposes that epistemic and aleatoric uncertainty measures should be viewed not as primitive quantities requiring axioms, but as consequences of a higher-level modelling choice: the evaluation of a strictly proper loss under a subjective risk. Definition 1 introduces subjective risk R_q(z)=E_{Y~q}[ℓ(z,Y)]; Theorem 2 gives a reverse bias–variance–entropy decomposition of its expectation into Bφ(q̄,p), Eθ[Bφ(qθ,q̄)], and Eθ[Hℓ(qθ)]. For log loss, the decomposition recovers Gal's terms (bias + mutual information + conditional entropy). With other strictly proper losses, the same theorem is shown to recover label-wise uncertainty (Sale et al.), variance-based measures (Kendall and Gal), deep ensembles, and data/procedural uncertainty splits. Sections 4–5 develop subjective-risk analogues of excess risk, approximation error, and estimation error; Theorem 4 establishes non-negativity of the subjective estimation error; Theorem 5 and Corollary 6 link mutual information to subjective estimation error; Theorem 8 argues that EPKL approximates mutual information in a small-disagreement regime. The paper concludes that disagreements among existing UQ measures are explained by loss geometry rather than by incompatibility.

Significance. If the semantic identification is accepted, the paper offers a genuinely unifying framework: a single risk decomposition generates many existing aleatoric/epistemic measures as special cases, with no fitted parameters, and it connects these measures to standard learning-theoretic quantities. Theorems 2, 4, 5, and Corollary 6 are mathematically sound, and the examples are worked out in enough detail to be checked. The main caveat is that the epistemic/aleatoric interpretation is imposed rather than derived; the paper's contribution is therefore best read as a unifying identification, not a derivation. With that framing, this could be an influential conceptual contribution to the UQ literature.

major comments (1)
  1. [Abstract; §3.3, Eq. (10)] The paper's central claim — that aleatoric and epistemic uncertainty measures are 'consequences' rather than primitives — is stronger than what Theorem 2 establishes. Theorem 2 is a bias–variance–entropy identity; the identification of the Bregman variance term as epistemic and the expected generalized entropy as aleatoric is asserted after Eq. (10) ('The epistemic uncertainty is the variance... and the aleatoric uncertainty is the corresponding generalized entropy'). The same identity is compatible with other semantic assignments, e.g. viewing the bias term as the reducible component, as argued by Jimenez et al. (2026), which the paper rejects in §5.2 on conceptual grounds. This is not an internal mathematical error, but it is load-bearing: without this mapping, the headline result reduces to 'a strictly proper expected subjective risk decomposes into three nonnegative terms.' I recomme
minor comments (5)
  1. [§5.4, Theorem 8; Appendix C] The statement and proof of Theorem 8 use inconsistent notation. The theorem parameterizes perturbations as qθ(y)=qθ′(y)(1+εh(y)) with ∫h(y)qθ′(y)dν=0, but the proof introduces a reference q0 and functions rθ satisfying ∫q0 rθ=0 and Eθ[rθ]=0, then derives expansions (55)–(56) in terms of rθ. No bridge between h and rθ is given. The theorem as stated is therefore not proved; the proof should be rewritten in a single parameterization (e.g. qθ=q0(1+εrθ), with h identified as rθ−rθ′ or as rθ, depending on the intended statement).
  2. [§5.4, Theorem 7 statement] Typo: the sentence after Eq. (25) says the mutual information 'contains an asymmetric component, and an asymmetric one'; the second occurrence should be 'symmetric'.
  3. [Appendix C, proof of Theorem 7] The proof is essentially a rearrangement of the known identity EPKL = MI + RMI (Malinin and Gales, 2021) rather than a derivation from first principles. Please cite that identity explicitly at the point of use and clarify which parts are being proved.
  4. [Appendix B, Example 5, Eq. (52)] There is a bracket typo in the variance term: 'E_Z[µZ − E_Z[µZ])²' should read 'E_Z[(µZ − E_Z[µZ])²]'.
  5. [§4.1, after Eq. (16)] The sentence 'the expected subjective risk is 3 the subjective risk of the Bayesian Model Average' appears to have a superscript/marker artifact; please fix the formatting.

Circularity Check

0 steps flagged

No significant circularity; the derivations are self-contained and recover existing UQ measures without fitting or self-citation load-bearing steps.

full rationale

The paper's central derivation is Theorem 2, which decomposes the expected subjective risk E_theta[E_{Y~q_theta}[ell(p,Y)]] into bias B_phi(q*_R,p), variance E_theta[B_phi(q_theta,q*_R)], and generalized entropy E_theta[H_ell(q_theta)]. The proof is given in Appendix A and uses only the proper-loss/Bregman representation, the three-point identity, and Jensen-style centroid arguments. No parameter is fitted to data and then renamed as a prediction; no quantity defined in terms of the target result is used to derive it. The examples recover previously published uncertainty measures (Gal mutual information/conditional entropy, Sale label-wise terms, Kendall-Gal variance terms, Huang data/procedural splits, deep ensembles) as special cases, and these are external benchmarks rather than assumptions of the derivation. The only potentially circular-looking move is the interpretive labeling in §3.3, where the Bregman variance term is called epistemic and the expected generalized entropy is called aleatoric. This labeling is asserted rather than derived, and one could alternatively read the bias term as a reducible component, as Jimenez et al. do. But this is a semantic/interpretive choice, not an equation reducing to itself or a fitted input being called a prediction. No load-bearing step depends on a self-citation: Brown and Ali (2024) appears only as a comparison for approximation/estimation relationships, and the relevant identities in Section 4 are proven in the text. External citations such as Pfau (2025) and Heskes (2026) provide background properties of Bregman divergences, not the paper's own conclusions. Under the stated rules, this is a normal non-circular case: score 0.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 0 invented entities

No constants are fitted; the choice of loss ℓ and hypothesis class H are modelling decisions rather than fitted parameters. The central derivation uses standard Bregman/proper-loss machinery plus the paper's interpretive identification of variance with epistemic uncertainty. No new physical entities are introduced.

axioms (5)
  • domain assumption The generalized entropy Hℓ(q)=E_{Y~q}[ℓ(q,Y)] is differentiable on the relative interior of P, so the Savage/Bregman representation of the proper loss holds.
    Used in Theorem 2 and throughout §3 to write Rq(p)=Bφ(q,p)+Hℓ(q). Cited to Savage (1971); standard for proper losses.
  • standard math The right Bregman centroid minimizing Eθ[Bφ(qθ,z)] is the mixture qbar=Eθ[qθ].
    Invoked in Theorem 2 and Examples 1/2 to identify the bias and variance terms. Relies on Frigyik et al. (2008) for functional Bregman divergences.
  • standard math The Bregman three-point identity holds and cross terms vanish after taking expectations since Eθ[qθ−qbar]=0.
    Core algebraic step in the proofs of Theorem 2 and Theorem 9.
  • ad hoc to paper The Bregman variance term is identified as epistemic uncertainty and the expected generalized entropy as aleatoric uncertainty.
    This semantic mapping is the paper's proposal, not a mathematical consequence; the same decomposition could be labeled differently (e.g. bias-as-epistemic in Jimenez et al. 2026). It is load-bearing for the claim that UQ measures are consequences.
  • domain assumption Existence and uniqueness of right Bregman projections ΠR_H(r), and continuity of the projection map at p for the asymptotic statement.
    Required for Proposition 3 and the learning-theoretic decompositions (Eqs 19–22); not proven in the paper.

pith-pipeline@v1.3.0-alltime-deepseek · 20417 in / 14184 out tokens · 117731 ms · 2026-08-01T23:51:50.306565+00:00 · methodology

0 comments
read the original abstract

We present a novel viewpoint for uncertainty quantification. Uncertainty measures are not primitives, in need of axioms and argumentation, but instead consequences, of higher-level modelling decisions. We show how epistemic and aleatoric uncertainty measures can be derived via decomposition of a subjective risk, based on a strictly proper loss. Reverse cross-entropy provides a prominent example, where decomposition recovers the classic information-theoretic uncertainty terms. The same approach recovers numerous measures previously proposed across the UQ literature, providing them a common theoretical foundation. From a practical point of view, this suggests a new approach to UQ: given a modelling scenario and strictly proper loss, the corresponding epistemic and aleatoric terms are induced by the subjective-risk decomposition. We then extend our view to learning theory: we introduce and analyse subjective risk analogues of excess risk, approximation error, and estimation error, and identify the connections to UQ. We consider this a first step towards a full learning-theoretic framework for uncertainty quantification.

Figures

Figures reproduced from arXiv: 2607.15196 by Gavin Brown, Michele Caprio, Raghad Alamri.

Figure 1
Figure 1. Figure 1: LEFT: Log, Brier, and Pseudo-spherical loss, normalized to a common maximum. RIGHT: Second derivatives of the associated generator function, i.e. ∇2ϕ(p). The case of log and pseudo-spherical (with β = 1.5) rise rapidly as p approaches zero, affording them sensitivity to probability mass being placed on implausible events, the far left side of the x-axis. The Brier (squared loss) has a slower growth (consta… view at source ↗
Figure 2
Figure 2. Figure 2: Different posteriors q(θ | D) for the parameter θ of a Bernoulli. As the variance of q(θ | D) increases (i.e. ϵ grows from left to right), the 2nd-order approximation of Theorem 8, EPKL ≈ 2 × MI, becomes less accurate. 6 Conclusion We have presented a new view of uncertainty quantification. Rather than treating aleatoric and epistemic uncertainty as primitive quantities that must be independently defined, … view at source ↗

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

25 extracted references · 8 linked inside Pith

  1. [2]

    Further Examples of Bias/Variance for Subjective Risk Example 4

    However, the generalised entropy is the predictive variance, which for the Bernoulli distribution is Hℓ(µθ) :=R qθ (qθ) =E ˆY∼q θ (µθ − ˆY) 2 =E ˆY∼q θ ˆY 2 − E ˆY∼q θ [ ˆY] 2 (41) Since ˆY∼Ber(µ θ) takes values in{0,1}, we have ˆY 2 = ˆY, so this gives, Hℓ(µθ) =µ θ −µ 2 θ =µ θ ·(1−µ θ).(42) 23 Alamri, Caprio, and Brown Appendix B. Further Examples of Bia...

    2017

  2. [3]

    B´ ela A Frigyik, Santosh Srivastava, and Maya R Gupta

    URLhttps://arxiv.org/abs/2509.26610. B´ ela A Frigyik, Santosh Srivastava, and Maya R Gupta. Functional bregman divergence and bayesian estimation of distributions.IEEE Transactions on Information Theory, 54 (11):5130–5139,

    arXiv

  3. [4]

    Proofs for section 5 We prove the relation between mutual information and EPKL (Schweighofer et al., 2023)

    26 Subjective Risk Decomposition: A New View for Uncertainty Quantification Appendix C. Proofs for section 5 We prove the relation between mutual information and EPKL (Schweighofer et al., 2023). Proof of Theorem 7.Multiplying the expression by 2 and rearranging gives 2·I( ˆY; Θ) =E θ,θ′K(q θ ||q θ′) +E θ h K(q θ ||¯q)−K(¯q||q θ) i I( ˆY; Θ) =E θ,θ′K(q θ ...

    2023

  4. [6]

    Ziyi Huang, Henry Lam, and Haofeng Zhang

    URLhttps://arxiv.org/abs/2404.12215. Ziyi Huang, Henry Lam, and Haofeng Zhang. Quantifying epistemic uncertainty in deep learning.arXiv preprint arXiv:2110.12122,

    Pith/arXiv arXiv

  5. [7]

    Sebastian Jimenez, Mira Juergens, and Willem Waegeman

    URLhttps://arxiv.org/abs/ 2306.05674. Sebastian Jimenez, Mira Juergens, and Willem Waegeman. Position: Epistemic un- certainty estimation methods are fundamentally incomplete. InForty-third Interna- tional Conference on Machine Learning Position Paper Track,

    Pith/arXiv arXiv

  6. [9]

    18 Subjective Risk Decomposition: A New View for Uncertainty Quantification Andrey Malinin.Uncertainty Estimation in Deep Learning with application to Spoken Language Assessment

    URLhttps://proceedings.neurips.cc/paper_files/paper/2017/file/ 9ef2ed4b7fd2c810847ffa5fa85bce38-Paper.pdf. 18 Subjective Risk Decomposition: A New View for Uncertainty Quantification Andrey Malinin.Uncertainty Estimation in Deep Learning with application to Spoken Language Assessment. PhD thesis, University of Cambridge,

    2017

  7. [11]

    Frank P Ramsey

    URL https://arxiv.org/abs/2511.08789. Frank P Ramsey. Truth and probability. InReadings in formal epistemology: Sourcebook, pages 21–45. Springer,

    arXiv

  8. [12]

    org/abs/2401.00276

    URLhttps://arxiv. org/abs/2401.00276. Yusuf Sale, Paul Hofman, Timo L¨ ohr, Lisa Wimmer, Thomas Nagler, and Eyke H¨ ullermeier. Label-wise Aleatoric and Epistemic Uncertainty Quantification. InThe 40th Conference on Uncertainty in Artificial Intelligence,

    Pith/arXiv arXiv

  9. [14]

    Aleatoric and epistemic uncertainty with random forests

    Mohammad Hossein Shaker and Eyke H¨ ullermeier. Aleatoric and epistemic uncertainty with random forests. InAdvances in Intelligent Data Analysis XVIII: 18th International Symposium on Intelligent Data Analysis, IDA 2020, Konstanz, Germany, April 27–29, 2020, Proceedings 18, pages 444–456. Springer,

    2020

  10. [16]

    Lisa Wimmer, Yusuf Sale, Paul Hofman, Bernd Bischl, and Eyke H¨ ullermeier

    X (formerly Twitter) thread, ac- cessed 2026-05-29. Lisa Wimmer, Yusuf Sale, Paul Hofman, Bernd Bischl, and Eyke H¨ ullermeier. Quantifying aleatoric and epistemic uncertainty in machine learning: Are conditional entropy and mutual information appropriate measures? InProceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence, vol...

    2026

  11. [17]

    mlr.press/v216/wimmer23a.html

    URLhttps://proceedings. mlr.press/v216/wimmer23a.html. Wei Xiong, Hanze Dong, Chenlu Ye, Ziqi Wang, Han Zhong, Heng Ji, Nan Jiang, and Tong Zhang. Iterative preference learning from human feedback: Bridging theory and practice for rlhf under kl-constraint.arXiv preprint arXiv:2312.11456,

    Pith/arXiv arXiv

  12. [19]

    reverse-argument

    Hence, Rqθ (p) =E ˆY∼q θ [ℓ(p, ˆY)] =B ϕ(qθ, p) +Hℓ(qθ),(28) and assuming expectations are finite, Eθ Rqθ (p) =E θ Bϕ(qθ, p) +E θ Hℓ(qθ) .(29) By definitionq ∗ R = arg minz∈P Eθ[Bϕ(qθ, z)], which is the mixtureq ∗ R = ¯q=E θ[qθ]. The Bregman three-point identity (Nielsen and Nock, 2009), Bϕ(qθ, p) =Bϕ(qθ,¯q) +Bϕ(¯q, p) + ∇ϕ(¯q)− ∇ϕ(p), qθ −¯q .(30) Taking...

    2009

  13. [21]

    variance-based

    Ignoring affine terms, this inducesϕ(µ) =µ 2, and Theorem 2 becomes, Eθ h E ˆY∼q θ (µp − ˆY) 2 i | {z } expected subjective risk = Eθ[µθ]−µ p 2 | {z } bias +E θ h (µθ −E θ[µθ )2 i | {z } variance +σ 2 q |{z} generalized entropy (43) These are the “variance-based” terms studied by many authors for UQ in regression, e.g. Kendall and Gal (2017); Depeweg et a...

    2017

  14. [22]

    Assume a convex class of distributionsPover an outcome spaceY

    )LetB ϕ be a Bregman divergence generated by a strictly convex, differentiable functionϕ. Assume a convex class of distributionsPover an outcome spaceY. Letq Z ∈ Pbe a model induced by ajointrandom variableZ= (A, B). For the reverse-argument Bregman divergence, the varianceE Z Bϕ(qZ, q∗ R) can be decomposed by the law of total variance as follows: EZ h Bϕ...

    2022

  15. [23]

    (2021), Huang et al

    toZ= (A, B) = (D, θ) splits it into a between-group term over datasets and a within-group term over procedural randomness, EZ (µZ −E Z[µZ])2 =E D (Eθ|D[µZ]−E Z[µZ])2 | {z } data +E D Eθ|D[(µZ −E θ|D[µZ])2] | {z } procedural ,(54) recovering the data/procedural decomposition discussed in Huang et al. (2021), Huang et al. (2023), and Jimenez et al. (2026). ...

    2021

  16. [24]

    as an exemplar. Assume the same setting as Example 4, but now withunknown varianceand a finite set ofmmodels: eachθ i inducesq θi =N(µ i, σ2 i ), so the expectation overθbecomes a finite average 1 m Pm i=1. 1 m mX i=1 " E ˆY∼q θi h (µp − ˆY) 2 i # | {z } expected subjective risk = 1 m mX i=1 µi −µ p !2 | {z } bias + 1 m mX i=1 µ2 i − 1 m mX i=1 µi !2 | {z...

    2025

  17. [1954]

    Kajetan Schweighofer, Lukas Aichberger, Mykyta Ielanskyi, and Sepp Hochreiter

    doi: https: //doi.org/10.1002/nav.3800010316. Kajetan Schweighofer, Lukas Aichberger, Mykyta Ielanskyi, and Sepp Hochreiter. Introduc- ing an improved information-theoretic measure of predictive uncertainty.arXiv preprint arXiv:2311.08309,

    Pith/arXiv arXiv

  18. [2009]

    David Pfau

    doi: 10.1109/TIT.2009.2018176. David Pfau. A generalized bias-variance decomposition for bregman divergences,

    arXiv 2009

  19. [2017]

    ISBN 9781510860964

    Curran Associates Inc. ISBN 9781510860964. Michael Kirchhof, Gjergji Kasneci, and Enkelejda Kasneci. Reexamining the aleatoric and epistemic uncertainty dichotomy. InThe Fourth Blogpost Track at ICLR 2025,

    2025

  20. [2020]

    20 Subjective Risk Decomposition: A New View for Uncertainty Quantification Appendix A

    URLhttps://arxiv.org/abs/1909.08593. 20 Subjective Risk Decomposition: A New View for Uncertainty Quantification Appendix A. Proofs for section 3 Theorem 2, the ‘reverse-argument’ BV decomposition, uses the divergence/entropy decom- position of a strictly proper loss to obtain a Bregman divergence, then applies the right- centroid decomposition from Pfau ...

    Pith/arXiv arXiv 1909

  21. [2022]

    X discussion thread on uncertainty and machine learning.https://x.com/ BlackHC/status/1817556167687569605, July

    19 Alamri, Caprio, and Brown Various. X discussion thread on uncertainty and machine learning.https://x.com/ BlackHC/status/1817556167687569605, July

    arXiv

  22. [2023]

    Yuxian Gu, Li Dong, Furu Wei, and Minlie Huang

    URLhttps://arxiv.org/abs/2210.12256. Yuxian Gu, Li Dong, Furu Wei, and Minlie Huang. MiniLLM: Knowledge distillation of large language models. InInternational Conference on Learning Representations,

    Pith/arXiv arXiv

  23. [2024]

    An Axiomatic Assessment of Entropy- and Variance-based Uncertainty Quantification in Regression.arXiv preprint arXiv:2504.18433,

    Christopher B¨ ulte, Yusuf Sale, Timo L¨ ohr, Paul Hofman, Gitta Kutyniok, and Eyke H¨ ullermeier. An Axiomatic Assessment of Entropy- and Variance-based Uncertainty Quantification in Regression.arXiv preprint arXiv:2504.18433,

    Pith/arXiv arXiv

  24. [2025]

    On LLM knowledge distillation - a comparison between forward KL and reverse KL

    Yihan Cao and Yanbin Kang. On LLM knowledge distillation - a comparison between forward KL and reverse KL. InThe Fourth Blogpost Track at ICLR 2025,

    2025

  25. [2026]

    Paul Hofman, Yusuf Sale, and Eyke H¨ ullermeier

    URLhttps://arxiv.org/abs/2501.18581. Paul Hofman, Yusuf Sale, and Eyke H¨ ullermeier. Quantifying aleatoric and epistemic un- certainty with proper scoring rules,

    arXiv