Tootfinder

Oh nice! Account notes carry over when someone migrates. @…

This https://arxiv.org/abs/2503.21224 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csLG_…

Efficient Learning for Entropy-Regularized Markov Decision Processes via Multilevel Monte Carlo
Designing efficient learning algorithms with complexity guarantees for Markov decision processes (MDPs) with large or continuous state and action spaces remains a fundamental challenge. We address this challenge for entropy-regularized MDPs with Polish state and action spaces, assuming access to a generative model of the environment. We propose a novel family of multilevel Monte Carlo (MLMC) algorithms that integrate fixed-point iteration with MLMC techniques and a generic stochastic approximat…

Egun on! Bideo motibazional bat ondo hasteko eguna:

Beyond Interval MDPs: Tight and Efficient Abstractions of Stochastic Systems
Ibon Gracia, Morteza Lahijanian
https://arxiv.org/abs/2507.02213 https://

Beyond Interval MDPs: Tight and Efficient Abstractions of Stochastic Systems
This work addresses the general problem of control synthesis for continuous-space, discrete-time stochastic systems with probabilistic guarantees via finite abstractions. While established methods exist, they often trade off accuracy for tractability. We propose a unified abstraction framework that improves both the tightness of probabilistic guarantees and computational efficiency. First, we introduce multi-interval MDPs (MI-MDPs), a generalization of interval-valued MDPs (IMDPs), which allows…

Ich nutze #Mastodon seit 2017 (!) und habe heute (2025) mitbekommen, dass man in der Suchmaske "from:me" eingeben kann, um die Suche auf eigenen Beiträge einzuschränken! 🤩
Ich finde das jetzt schon super hilfreich, um längerfristig an Themen dranzubleiben und nach einer Weile nochmal in ältere Threads reinzuschauen und ggf. nachzuhaken.
Ist das ein alter Hut oder ein nütz…

Topology induced modifications in the critical behavior of the Yaldram Khan catalytic reaction model
Paulo F. Gomes, Henrique A. Fernandes, Roberto da Silva
https://arxiv.org/abs/2506.04485

Topology induced modifications in the critical behavior of the Yaldram Khan catalytic reaction model
In this work, we investigated how the use of complex networks as catalytic surfaces can affect the phase diagram of the Yaldram-Khan model, as well as how the order of the phase transitions present in the seminal work behaves when the randomness is added to the model. The study was conducted by taking into consideration two well-known random networks, the Erdos-Renyi network (ERN), with its long-range randomness, and the random geometric graph (RGG), with its spatially constrained randomness. W…

Hall algebras and Hecke modifications of vector bundles
Roberto Alvarenga, Leonardo Mo\c{c}o
https://arxiv.org/abs/2506.00186 https://

Hall algebras and Hecke modifications of vector bundles
In this article, we investigate Hecke modifications of vector bundles on a smooth projective curve $X$ defined over an arbitrary field. We obtain structural results that allow us to reduce the classification problem of Hecke modifications to the case of vector bundles of lower rank. Moreover, when the base field is a finite field and $X$ is the projective line, we apply the Hall algebra of coherent sheaves to provide a full classification of the Hecke modifications, including their multipliciti…

Pfff... I knew about that since like forever.
Grant even showed up into at least one of Ferguson's shows, operating Geoff.
https://noc.social/@todayilearned/114795935441049956

Today I learned (@todayilearned@noc.social)
Attached: 1 image TIL that Geoff Peterson, the animatronic human skeleton who served as the sidekick on The Late Late Show with Craig Ferguson, was designed and built by MythBusters' Grant Imahara. https://en.wikipedia.org/wiki/Geoff_Peterson #til #todayilearned https://www.reddit.com/r/todayilearned/comments/1lrmt52/til_that_geoff_peterson_the_animatronic_human/

Thinking Beyond Visibility: A Near-Optimal Policy Framework for Locally Interdependent Multi-Agent MDPs
Alex DeWeese, Guannan Qu
https://arxiv.org/abs/2506.04215

Thinking Beyond Visibility: A Near-Optimal Policy Framework for Locally Interdependent Multi-Agent MDPs
Decentralized Partially Observable Markov Decision Processes (Dec-POMDPs) are known to be NEXP-Complete and intractable to solve. However, for problems such as cooperative navigation, obstacle avoidance, and formation control, basic assumptions can be made about local visibility and local dependencies. The work DeWeese and Qu 2024 formalized these assumptions in the construction of the Locally Interdependent Multi-Agent MDP. In this setting, it establishes three closed-form policies that are tr…

This https://arxiv.org/abs/2505.19946 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csLG_…

Inverse Q-Learning Done Right: Offline Imitation Learning in $Q^π$-Realizable MDPs
We study the problem of offline imitation learning in Markov decision processes (MDPs), where the goal is to learn a well-performing policy given a dataset of state-action pairs generated by an expert policy. Complementing a recent line of work on this topic that assumes the expert belongs to a tractable class of known policies, we approach this problem from a new angle and leverage a different type of structural assumption about the environment. Specifically, for the class of linear $Q^π$-rea…