Tootfinder

In Germany it's a day to remind volunteering "Tag des #ehrenamt ".
Well I can't contribute a lot besides reminding about the efforts of the trail maintenance teams of the alpine clubs.
Recently I made a video for the local alpine club about that (I should ask if I'm allowed to publish it on my peertube as well 🤔)

Investigating the Sulfur Mystery in Protoplanetary Disks Through Chemical Modeling
Becky J. Williams, L. Ilsedore Cleeves, Rachel E. Gross, Jackson Baker
https://arxiv.org/abs/2510.02176

Investigating the Sulfur Mystery in Protoplanetary Disks Through Chemical Modeling
Sulfur is a critical element to life on Earth, and with detections of sulfur-bearing molecules in exoplanets and comets, questions arise as to how sulfur is incorporated into planets in the first place. In order to understand sulfur's journey from molecular clouds to planets, we need to understand the molecular forms that sulfur takes in protoplanetary disks, where the rotational emission from sulfur-bearing molecules in the gas phase indicates a very low abundance. To address this question, we…

Accelerating cosmological simulations on GPUs: a portable approach using OpenMP
M. D. Lepinzan, G. Lacopo, D. Goz, G. Taffoni, P. Monaco, P. J. Elahi, U. Varetto, M. Cytowski
https://arxiv.org/abs/2510.02873

Accelerating cosmological simulations on GPUs: a portable approach using OpenMP
In this work we present the porting to Graphics Processing Units (GPUs, using OpenMP target directives) and optimization of a key module within the cosmological {\pinocchio} code, a Lagrangian Perturbation Theory (LPT)-based framework widely used for generating dark matter (DM) halo catalogs. Our optimization focuses on a specific segment of the code responsible for calculating the collapse time of each particle involved in the simulation. Due to the embarrassingly parallel nature of this compu…

Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards
Yiran Shen, Yu Xia, Jonathan Chang, Prithviraj Ammanabrolu
https://arxiv.org/abs/2510.01167 h…

Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable Rewards
Aligning large language models to human preferences is inherently multidimensional, yet most pipelines collapse heterogeneous signals into a single optimizeable objective. We seek to answer what it would take to simultaneously align a model across various domains spanning those with: verifiable rewards (mathematical accuracy), non-verifiable subjective preferences (human values), and complex interactive scenarios (multi-turn AI tutoring dialogues). Such multi-objective reinforcement learning se…