arXiv research

Smooth calibration improves forecast reliability even with leaked information.

Paper detects duality obstruction in smooth calibrations.

Cone structures over minimal products can't be calibrated smoothly.

The study extends calibrated geometry to smooth maps and finds energy bounds.

The paper provides a uniform convergence bound for smooth calibration error and its relationship with functional gradient.

Improves confidence calibration in neural networks by smoothing labels based on class similarity.

The paper studies how to extend local calibration pairs to global ones in various situations. As a result, new discoveries involving mass-minimizing properties are exhibited. In particular, we show that a $\mathbb R$-homologically nontrivial connected submanifold $M$ of a smooth Riemannian manifold $X$ is homologically…

The study examines label smoothing to improve confidence calibration in fine-tuned LLMs.

New algorithms achieve decision calibration without sample complexity dependent on feature dimension.

Calibrations help estimate volumes on odd spheres without gaps.

New calibration measure SSCE ensures truthful prediction, unlike existing measures.

New method calibrates LV surfaces for exotic derivatives with smoother, more stable Greeks.

New method improves calibration of neural networks by targeting robust margins and local smoothness.

We show that any semi-calibration of degree 2 is locally induced by a smooth almost complex structure. We provide some applications of this result in the regularity theory for semi-calibrated 2-currents

Post-processing predictors reduces calibration errors for decision-making.

This work proves $L_2$-regularized ERM controls smCE without post-hoc correction.

New method calibrates multi-class predictions efficiently without sacrificing accuracy.

This is a companion note of [Zhaa] (arXiv:1501.01836) where the extension of local calibration pairs of smooth submanifolds is discussed. Here we emphasize on the case of singular submanifolds. More precisely, we study when a calibration pair around the singular set of a submanifold can extend to a local calibration pa…

We equip many non compact non simply connected surfaces with smooth Riemannian metrics whose isoperimetric profile is smooth, a highly non generic property. The computation of the profile is based on a calibration argument, a rearrangement argument, the Bol-Fiala curvature dependent inequality, together with new result…

DNAMite creates interpretable, calibrated survival analysis models.

T-Cal tests model calibration with a minimax optimal test.

Improves model calibration by addressing adversarially unrobust inputs.

Every graph can be represented as a singular set of a special surface.

Study shows multilingual LLM calibration effects improve model confidence but not accuracy.

Label smoothing improves model performance even with noisy labels.

Unique solutions found for Plateau problems in smooth and continuous calibrations.

In many classification problems it is desirable to output well-calibrated probabilities on the different classes. We propose a robust, non-parametric method of calibrating probabilities called SplineCalib that utilizes smoothing splines to determine a calibration function. We demonstrate how applying certain transforma…

We lay down an elementary yet fundamental lemma concerning a finite algebraicness property of a smooth map from an Azumaya/matrix manifold with a fundamental module to a smooth manifold. This gives us a starting point to build a synthetic (synonymously, $C^{\infty}$-algebraic) symplectic geometry and calibrated geometr…

The task of calibration is to retrospectively adjust the outputs from a machine learning model to provide better probability estimates on the target variable. While calibration has been investigated thoroughly in classification, it has not yet been well-established for regression tasks. This paper considers the problem…

For an element $Ψ$ in the graded vector space $Ω^*(M, TM)$ of tangent bundle valued forms on a smooth manifold $M$, a $Ψ$-submanifold is defined as a submanifold $N$ of $M$ such that $Ψ_{|N} \in Ω^*(N, TN)$. The class of $Ψ$-submanifolds encompasses calibrated submanifolds, complex submanifolds and all Lie subgroups in…

A new perfectly truthful calibration measure improves prediction reliability.

New algorithm reduces prediction errors across various loss functions.

Optimizing proper loss yields calibrated models under specific conditions.

We shall obtain unobstructed deformations of four geometric structures: Calabi-Yau, HyperKähler, $\G$ and Spin(7) structures in terms of closed differential forms (calibrations). We develop a direct and unified construction of smooth moduli spaces of these four geometric structures and show that the local Torelli type …

The generalization and learning speed of a multi-class neural network can often be significantly improved by using soft targets that are a weighted average of the hard targets and the uniform distribution over labels. Smoothing the labels in this way prevents the network from becoming over-confident and label smoothing…

CalArena benchmarks post-hoc calibration methods across various tasks.

New findings show that not all area-minimizing surfaces are calibrated, even on complex manifolds.

Two methods improve Gaussian process predictive distributions' calibration.

We construct a Riemannian metric $g$ on $\mathbb{R}^4$ (arbitrarily close to the euclidean one) and a smooth simple closed curve $Γ\subset \mathbb R^4$ such that the unique area minimizing surface spanned by $Γ$ has infinite topology. Furthermore the metric is almost Kähler and the area minimizing surface is calibrated…

New algorithm tests model calibration in nearly-linear time.

Proposes a method to repair arbitrage in option prices data.

Midicoth compresses online probability estimates by correcting prior smoothing biases.

We propose to use nonparametric Bernstein copulas as bivariate pair-copulas in high-dimensional vine models. The resulting smooth and nonparametric vine copulas completely obviate the error-prone need for choosing the pair-copulas from parametric copula families. By means of a simulation study and an empirical analysis…

This paper is devoted to the application of B-splines to volatility modeling, specifically the calibration of the leverage function in stochastic local volatility models and the parameterization of an arbitrage-free implied volatility surface calibrated to sparse option data. We use an extension of classical B-splines …

We study the out-of-sample properties of robust empirical optimization problems with smooth $φ$-divergence penalties and smooth concave objective functions, and develop a theory for data-driven calibration of the non-negative "robustness parameter" $δ$ that controls the size of the deviations from the nominal model. Bu…

Using Malliavin calculus techniques, we derive an analytical formula for the price of European options, for any model including local volatility and Poisson jump process. We show that the accuracy of the formula depends on the smoothness of the payoff function. Our approach relies on an asymptotic expansion related to …

WamOL uses PINNs to efficiently calibrate IVS from sparse data.

The paper solves Bernstein problems for specific submanifolds in high-dimensional spaces.