Tootfinder

Answer (Jaydn, opus 26) #PianoImprov #NowPlaying

Answer (Jaydn, opus 26) #PianoImprov #NowPlaying
Stan Stewart - @muz4now

Mind the Gap: A Closer Look at Tokenization for Multiple-Choice Question Answering with LLMs
Mario Sanz-Guerrero, Minh Duc Bui, Katharina von der Wense
https://arxiv.org/abs/2509.15020

Mind the Gap: A Closer Look at Tokenization for Multiple-Choice Question Answering with LLMs
When evaluating large language models (LLMs) with multiple-choice question answering (MCQA), it is common to end the prompt with the string "Answer:" to facilitate automated answer extraction via next-token probabilities. However, there is no consensus on how to tokenize the space following the colon, often overlooked as a trivial choice. In this paper, we uncover accuracy differences of up to 11% due to this (seemingly irrelevant) tokenization variation as well as reshuffled model rankings, ra…

"Amerikanska Pen anser att bokutrensningarna är ”statsdriven censur” och kallar dem ett ”strategiskt försök att befästa makten genom rädsla, genom att kringgå rättsliga prejudikat och att tysta olika röster i Floridas offentliga skolor”.
https://www.dn.se/kultur/anne-franks-dagbo

Robust Online Calibration for UWB-Aided Visual-Inertial Navigation with Bias Correction
Yizhi Zhou, Jie Xu, Jiawei Xia, Zechen Hu, Weizi Li, Xuan Wang
https://arxiv.org/abs/2508.10999

Robust Online Calibration for UWB-Aided Visual-Inertial Navigation with Bias Correction
This paper presents a novel robust online calibration framework for Ultra-Wideband (UWB) anchors in UWB-aided Visual-Inertial Navigation Systems (VINS). Accurate anchor positioning, a process known as calibration, is crucial for integrating UWB ranging measurements into state estimation. While several prior works have demonstrated satisfactory results by using robot-aided systems to autonomously calibrate UWB systems, there are still some limitations: 1) these approaches assume accurate robot l…

Medico 2025: Visual Question Answering for Gastrointestinal Imaging
Sushant Gautam, Vajira Thambawita, Michael Riegler, P{\aa}l Halvorsen, Steven Hicks
https://arxiv.org/abs/2508.10869

Medico 2025: Visual Question Answering for Gastrointestinal Imaging
The Medico 2025 challenge addresses Visual Question Answering (VQA) for Gastrointestinal (GI) imaging, organized as part of the MediaEval task series. The challenge focuses on developing Explainable Artificial Intelligence (XAI) models that answer clinically relevant questions based on GI endoscopy images while providing interpretable justifications aligned with medical reasoning. It introduces two subtasks: (1) answering diverse types of visual questions using the Kvasir-VQA-x1 dataset, and (2…

Das Bundesamt für Sicherheit in der Informationstechnik (BSI) warnt vor der weiteren Nutzung von Windows 10 nach dem Support-Ende im Oktober 2025. Millionen Nutzer müssen jetzt handeln.
Zum Artikel: https://heise.de/-10486622?wt_mc=sm.re

»IT-Security – Cyberangriff auf "Nius": Wohl Daten von 5700 Abonnenten geleakt.
Die Angreifer nutzten Sicherheitslücken, um Zugriff auf die Datenbank zu erhalten«
Einer der seltensten Hacks die ich mal nicht nur stillschweigend bejahend bin, sonder auch hier teile.
👉

Cyberangriff auf "Nius": Wohl Daten von 5700 Abonnenten geleakt
Die Angreifer nutzten Sicherheitslücken, um Zugriff auf die Datenbank zu erhalten

#Wordle 1,488 4/6*
⬜🟨⬜⬜🟨 <1% of 214,624 (27)
🟨⬜🟨⬜🟨 6% of 122 (7)
🟨🟨⬜🟨🟩 0 of 19 (2)
🟩🟩🟩🟩🟩
WordleBot
Skill 89/99
Luck 52/99
well that was lol funny. I typed in the answer as my 3rd guess and then thought of another word I liked better so I used that as my 3rd guess. I still like my 3rd guess better as I feel it provided more info though, so I don't feel bad!
Bot had a better 3rd guess that would have left only 1 word no matter which of the 7 words the answer was, instead of the 2 that my guess left.

Anyone who follows me has probably gathered that I've become a fan of Discourse after running a few Discourse instances. They are currently requesting feedback.
I was pleased to see that they used a survey vendor that appears to have meaningful a11y — I didn't look for aria labels so I won't speak to that, but what I did notice is that there is explicitly-labeled support for keyboard-only interaction on every page, and I really appreciated that thoughtfulness.

Towards Reliable and Interpretable Document Question Answering via VLMs
Alessio Chen, Simone Giovannini, Andrea Gemelli, Fabio Coppini, Simone Marinai
https://arxiv.org/abs/2509.10129

Towards Reliable and Interpretable Document Question Answering via VLMs
Vision-Language Models (VLMs) have shown strong capabilities in document understanding, particularly in identifying and extracting textual information from complex documents. Despite this, accurately localizing answers within documents remains a major challenge, limiting both interpretability and real-world applicability. To address this, we introduce \textit{DocExplainerV0}, a plug-and-play bounding-box prediction module that decouples answer generation from spatial localization. This design m…