Tootfinder

"In the budget speech this year, commission president Ursula von der Leyen made a bold statement that AI is expected to approach human reasoning by next year."
How the #EU commission is basing its #AI policy (and funding) on the wishful thinking of tech bros. Those, which currently use circu…

Series C, Episode 01 - Aftermath
CALLY: Jenna's gone with Blake.
AVON: So let's get on with it. [Explosion. Avon is unconscious]
CALLY: He's out cold. We must get him into a capsule fast. And then you can put Orac in with him. Come on, quick!
https://blake.torpidity.net/m/301/14

Generalized Zero-Shot Learning for Point Cloud Segmentation with Evidence-Based Dynamic Calibration
Hyeonseok Kim, Byeongkeun Kang, Yeejin Lee
https://arxiv.org/abs/2509.08280 h…

Generalized Zero-Shot Learning for Point Cloud Segmentation with Evidence-Based Dynamic Calibration
Generalized zero-shot semantic segmentation of 3D point clouds aims to classify each point into both seen and unseen classes. A significant challenge with these models is their tendency to make biased predictions, often favoring the classes encountered during training. This problem is more pronounced in 3D applications, where the scale of the training data is typically smaller than in image-based tasks. To address this problem, we propose a novel method called E3DPC-GZSL, which reduces overconf…

Extending Ghouila-Houri's Characterization of Comparability Graphs to Temporal Graphs
Pierre Charbit, Michel Habib, Amalia Sorondo
https://arxiv.org/abs/2510.06849 https://

Extending Ghouila-Houri's Characterization of Comparability Graphs to Temporal Graphs
An orientation of a given static graph is called transitive if for any three vertices $a,b,c$, the presence of arcs $(a,b)$ and $(b,c)$ forces the presence of the arc $(a,c)$. If only the presence of an arc between $a$ and $c$ is required, but its orientation is unconstrained, the orientation is called quasi-transitive. A fundamental result presented by Ghouila-Houri guarantees that any static graph admitting a quasi-transitive orientation also admits a transitive orientation. In a seminal work…

The rise and fall of Little Red Dots could be driven by the environment
Rosa M. M\'erida, Gaia Gaspar, Yoshihisa Asada, Marcin Sawicki, Kiyoaki Christopher Omori, Chris J. Willott, Nicholas S. Martis, Adam Muzzin, Ga\"el Noirot, Gregor Rihtar\v{s}i\v{c}, Ghassan T. E. Sarrouh, Roberta Tripodi
https://arxiv.org/abs/2510.06408

The rise and fall of Little Red Dots could be driven by the environment
The Little Red Dot (LRD) paradigm comprises three main unknowns that are intrinsically connected: (1) What is the nature of these sources? (2) How do they form? (3) How do they evolve? Larger spectroscopic samples and high-resolution data are needed to delve deeper into the mechanisms ruling these sources. Understanding their formation and evolution requires identifying the rise and fall of the key features that characterize these systems, such as their compactness and ``V''-shaped spectral ene…

Bias reduction in g-computation for covariate adjustment in randomized clinical trials
Xin Zhang, Lin Liu, Haitao Chu
https://arxiv.org/abs/2509.07369 https://

Bias reduction in g-computation for covariate adjustment in randomized clinical trials
G-computation is a powerful method for estimating unconditional treatment effects with covariate adjustment in randomized clinical trials. It typically relies on fitting canonical generalized linear models. However, this could be problematic for small sample sizes or in the presence of rare events. Common issues include underestimation of the variance and the potential nonexistence of maximum likelihood estimators. Bias reduction methods are commonly employed to address these issues, including …

Recursive vectorized computation of the Frobenius norm
Vedran Novakovi\'c
https://arxiv.org/abs/2509.06220 https://arxiv.org/pdf/2509.06220

Recursive vectorized computation of the Frobenius norm
Recursive algorithms for computing the Frobenius norm of a real array are proposed, based on hypot, a hypotenuse function. Comparing their relative accuracy bounds with those of the BLAS routine DNRM2 it is shown that the proposed algorithms could in many cases be significantly more accurate. The scalar recursive algorithms are vectorized with the Intel's vector instructions to achieve performance comparable to xNRM2, and are further parallelized with OpenCilk. Some scalar algorithms are uncond…

Coupled tensor models for probability mass function estimation: Part II, Uniqueness of the model
Philippe Flores (CNRS, UGA, Grenoble INP, GIPSA-lab), Konstantin Usevich (CRAN, CNRS, UL), David Brie (CRAN, CNRS, UL)
https://arxiv.org/abs/2509.04931

Coupled tensor models for probability mass function estimation: Part II, Uniqueness of the model
In this paper, uniqueness properties of a coupled tensor model are studied. This new coupled tensor model is used in a new method called Partial Coupled Tensor Factorization of 3D marginals or PCTF3D. This method performs estimation of probability mass functions by coupling 3D marginals, seen as order-3 tensors. The core novelty of PCTF3D's approach (detailed in the part I article) relies on the partial coupling which consists on the choice of 3D marginals to be coupled. Tensor methods are ubiq…