arXiv research

Fine-tunes diffusion models to generate diverse samples with high genuine rewards.

New metrics improve scRNA-seq perturbation modeling by reducing mode collapse.

The paper examines how updates to probabilistic models influence behavior based on evidence.

We propose Turing Learning, a novel system identification method for inferring the behavior of natural or artificial systems. Turing Learning simultaneously optimizes two populations of computer programs, one representing models of the behavior of the system under investigation, and the other representing classifiers. …

We show that if a closed atoroidal 3-manifold M contains a genuine lamination, then it is group negatively curved in the sense of Gromov. Specifically, we exploit the structure of the non-product complementary regions of the genuine lamination and then apply the first author's Ubiquity Theorem to show that M satisfies …

We extend to the conformal realm the concept of genuine deformations of submanifolds, introduced by Dajczer and the first author for the isometric case. Analogously to that case, we call a conformal deformation of a submanifold $M^n$ genuine if no open subset of $M^n$ can be included as a submanifold of a higher dimens…

New method decomposes Markov chain rewards into persistent and transient components.

Signed compression progress on a sealed audit is goodhart-resistant.

We classify hypersurfaces of rank two of Euclidean space $\R^{n+1}$ that admit genuine isometric deformations in $\R^{n+2}$. That an isometric immersion $\hat f\colon\,M^n\to\R^{n+2}$ is a genuine isometric deformation of a hypersurface $f\colon\, M^n\to\R^{n+1}$ means that $\hat f$ is nowhere a composition $\hat f=\ha…

We extend the concept of genuine rigidity of submanifolds by allowing mild singularities, mainly to obtain new global rigidity results and unify the known ones. As one of the consequences, we simultaneously extend and unify Sacksteder and Dajczer-Gromoll theorems by showing that any compact $n$-dimensional submanifold …

This paper extends classifications of hypersurface immersions to higher dimensions.

In this paper we classify Euclidean hypersurfaces $f\colon M^n \rightarrow \mathbb{R}^{n+1}$ with a principal curvature of multiplicity $n-2$ that admit a genuine conformal deformation $\tilde{f}\colon M^n \rightarrow \mathbb{R}^{n+2}$. That $\tilde{f}\colon M^n \rightarrow \mathbb{R}^{n+2}$ is a genuine conformal defo…

A basic question in submanifold theory is whether a given isometric immersion $f\colon M^n\to\R^{n+p}$ of a Riemannian manifold of dimension $n\geq 3$ into Euclidean space with low codimension $p$ admits, locally or globally, a genuine infinitesimal bending. That is, if there exists a genuine smooth variation of $f$ by…

Develops a method for solving optimal stopping problems with multiple exercise rights.

We construct a pair of transverse genuine laminations on an atoroidal 3-manifold admitting transversely orientable uniform 1-cochain. The laminations are induced by the uniform 1-cochain and they are indeed the "straightening" of the coarse laminations defined in [Ca], by using minimal surface techniques. Moreover, whe…

In the Minority, Majority and Dollar Games (MG, MAJG, $G), synthetic agents compete for rewards, at each time-step acting in accord with the previously best-performing of their limited sets of strategies. Different components and/or aspects of real-world financial markets are modelled by these games. In the MG, agents …

New framework controls generalization for heavy-tailed data in RLHF and SGLD.

Concerning the problem of classifying complete submanifolds of Euclidean space with codimension two admitting genuine isometric deformations, until now the only known examples with the maximal possible rank four are the real Kaehler minimal submanifolds classified by Dajczer-Gromoll \cite{dg3} in parametric form. These…

Polynomial-time algorithm for list-decodable linear regression with batches.

Survey compares methods for generating artificial outliers.

Verified numerics prove existence of a curvature solution with known symmetries.

Audit financial machine learning workflows to detect spurious predictability.

As advances in signature recognition have reached a new plateau of performance at around 2% error rate, it is interesting to investigate alternative approaches. The approach detailed in this paper looks at using Variational Auto-Encoders (VAEs) to learn a latent space representation of genuine signatures. This is then …

An observable for nonabelian, higher-dimensional forms is introduced, its properties are discussed and its expectation value in BF theory is described. This is shown to produce potential and genuine invariants of higher-dimensional knots.

We examine the difference between several notions of curvature homogeneity and show that the notions introduced by Kowalski and Vanžurová are genuine generalizations of the ordinary notion of $k$-curvature homogeneity. The homothety group plays an essential role in the analysis.

We study left-invariant symmetric Killing 2-tensors on 2-step nilpotent Lie groups endowed with a left-invariant Riemannian metric, and construct genuine examples, which are not linear combinations of parallel tensors and symmetric products of Killing vector fields.

New concept SB-generation helps classify transformation groups.

Deep hedging uses RL to minimize risk in financial markets.

We present a simpler proof for the existence of adiabatic limits. Moreover, we added a new section where the adiabatic process is reversed and in some nondegenerate cases we deform the adiabatic limits to genuine irreducible solutions of the SW equations.

We show that among the Euclidean submanifolds with codimension two the ones of rank two that are parabolic but nonruled are isometrically rigid. This generalizes the result in [10] that these submanifolds are genuinely rigid. In addition, we give a parametric classifications of all parabolic submanifolds.

BCPO optimizes offline RL policies by converting uncertainty into conservative bounds.

Novel construction of Bauer--Furuta invariant using sheaves of spectra.

AI needs causal inference to avoid being just a correlation machine.

Gurau argued in [arXiv:1006.0714] that the gluing spaces arising as Feynman diagrams of three-dimensional group field theory are not all pseudo-manifolds. I dispute this conclusion: albeit not properly triangulated, these spaces are genuine pseudo-manifolds, viz. their singular locus is of codimension at least two.

Study on fake stationary Volterra Heston model for non-stationary processes.

Paper extends SI method for detecting CPs in complex systems' frequency domain.

An overwhelming number of true and false news stories are posted and shared in social networks, and users diffuse the stories based on multiple factors. Diffusion of news stories from one user to another depends not only on the stories' content and the genuineness but also on the alignment of the topical interests betw…

Study categorizes time series anomaly detection metrics based on evaluation challenges.

Using ideas from an article of P. Bieliavsky, M. Rooman and Ph. Spindel on BTZ black holes, I construct a family of interesting examples of quasi-Poisson actions as defined by A. Alekseev and Y. Kosmann-Schwarzbach. As an application, I obtain a genuine Poisson structure on $SL(2,R)$ which induces a Poisson structure o…

UNREAL selectively ensembles distinct models to improve active learning performance.

Reward hacking exploits misspecified rewards, affecting agent capabilities and true performance.

Paper introduces PRMs to learn non-Markovian stochastic rewards for reinforcement learning.

In this paper we prove that, in the category of chain complexes, partial algebras can be functorially replaced by quasi-isomorphic algebras. In particular, partial algebras contain all of the important homological and homotopical information that genuine algebras do. Applying this result to McClure's partial algebra in…

This work analyzes the value of future reward information in RL.

Self-supervised reward prediction improves RL in sparse reward settings.

The study categorizes reward errors in reinforcement learning, finding some can be beneficial.

Reward collapse occurs when ranking-based reward models yield uniform rewards for different prompts.

Reward models need more than just accuracy for effective RLHF.