Tootfinder

Benchmarking GPT-5 in Radiation Oncology: Measurable Gains, but Persistent Need for Expert Oversight
Ugur Dinc, Jibak Sarkar, Philipp Schubert, Sabine Semrau, Thomas Weissmann, Andre Karius, Johann Brand, Bernd-Niklas Axer, Ahmed Gomaa, Pluvio Stephan, Ishita Sheth, Sogand Beirami, Annette Schwarz, Udo Gaipl, Benjamin Frey, Christoph Bert, Stefanie Corradini, Rainer Fietkau, Florian Putz

Benchmarking GPT-5 in Radiation Oncology: Measurable Gains, but Persistent Need for Expert Oversight
Introduction: Large language models (LLM) have shown great potential in clinical decision support. GPT-5 is a novel LLM system that has been specifically marketed towards oncology use. Methods: Performance was assessed using two complementary benchmarks: (i) the ACR Radiation Oncology In-Training Examination (TXIT, 2021), comprising 300 multiple-choice items, and (ii) a curated set of 60 authentic radiation oncologic vignettes representing diverse disease sites and treatment indications. For …

Logical Consistency Between Disagreeing Experts and Its Role in AI Safety
Andr\'es Corrada-Emmanuel
https://arxiv.org/abs/2510.00821 https://arxiv.org/…

Logical Consistency Between Disagreeing Experts and Its Role in AI Safety
If two experts disagree on a test, we may conclude both cannot be 100 per cent correct. But if they completely agree, no possible evaluation can be excluded. This asymmetry in the utility of agreements versus disagreements is explored here by formalizing a logic of unsupervised evaluation for classifiers. Its core problem is computing the set of group evaluations that are logically consistent with how we observe them agreeing and disagreeing in their decisions. Statistical summaries of their al…

Regression Language Models for Code
Yash Akhauri, Xingyou Song, Arissa Wongpanich, Bryan Lewandowski, Mohamed S. Abdelfattah
https://arxiv.org/abs/2509.26476 https://

Regression Language Models for Code
We study code-to-metric regression: predicting numeric outcomes of code executions, a challenging task due to the open-ended nature of programming languages. While prior methods have resorted to heavy and domain-specific feature engineering, we show that a single unified Regression Language Model (RLM) can simultaneously predict directly from text, (i) the memory footprint of code across multiple high-level languages such as Python and C++, (ii) the latency of Triton GPU kernels, and (iii) the …

Kid2 had pizza day yesterday. The school decided to order from a new place that offers "healthy pizza." The kids apparently poured onto the playground after lunch, having barely touched the stuff, chanting "We want [corporate pizza chain]!" While I may not agree with their taste in pizza, I'll take collective action any day. And the kids won! The school sent around an email last night saying they were going back to the old place and that they'd pay for a pizza par…

lol so in order to add any kind of 2FA to a Twitch account, I have to give them my phone number and agree to “Twitch may use your phone number to call or send text messages for informational purposes.”
How much 2FA goodwill have thirsty companies like this destroyed?

Towards Improved Speech Recognition through Optimized Synthetic Data Generation
Yanis Perrin, Gilles Boulianne
https://arxiv.org/abs/2508.21631 https://arx…

Towards Improved Speech Recognition through Optimized Synthetic Data Generation
Supervised training of speech recognition models requires access to transcribed audio data, which often is not possible due to confidentiality issues. Our approach to this problem is to generate synthetic audio from a text-only corpus using a state-of-the-art text-to-speech model with voice cloning capabilities. Our goal is to achieve automatic speech recognition (ASR) performance comparable to models trained on real data. We explore ways to optimize synthetic data generation through finetuning…

Quantum internal vibrations in macroscopic systems with classical centers of mass
Gabriel H. S. Aguiar, George E. A. Matsas
https://arxiv.org/abs/2509.22429 https://

Quantum internal vibrations in macroscopic systems with classical centers of mass
Harmonizing classical and quantum worlds is a major challenge for modern physics. A significant portion of the scientific community supports the notion that classical mechanics is an effective theory that arises from quantum mechanics. Recently, the present authors have argued that this should not be the case, as quantum mechanics is not trustworthy for describing the center of mass of systems with masses $m$ much larger than the Planck mass $M_\text{P}$. In this vein, a simple gravitational se…

US-Regierung genehmigt Waffenverkauf an die Ukraine
Das US-Außenministerium hat den Verkauf von Waffen im Wert von geschätzt 825 Millionen US-Dollar (etwa 710 Millionen Euro) an die Ukraine genehmigt. Es gehe um Raketen mit größerer Reichweite und dazugehörige Ausrüstung, die die Ukraine angefragt habe, teilte die zuständige Behörde Defense Security Cooperation in Washington mit. Der Kongress sei darüber informiert worden - w…
📑

Ukraine-Liveblog: ++ Erneut Tote nach russischem Angriff ++
Die Ukraine meldet zwei Tote nach erneuten russischen Angriffen in der zentralukrainischen Region Dnipropetrowsk. Das US-Außenministerium hat den Verkauf von Waffen an die Ukraine genehmigt.

Threat intel company Dataminr plans to acquire cybersecurity threat intel provider ThreatConnect for $290M; Dataminr raised $85M in convertible funding in March (Greg Otto/CyberScoop)
https://cyberscoop.com/dataminr-threatconnect-acquisition/

Dataminr to acquire cybersecurity firm ThreatConnect for $290 million
Dataminr, a New York-based company specializing in real-time threat intelligence, announced plans Tuesday to acquire ThreatConnect, a cybersecurity threat intelligence provider, for $290 million. The acquisition will combine Dataminr’s AI-powered analysis of public data with ThreatConnect’s internal threat management capabilities, creating what the companies describe as “Client-Tailored intelligence” that adapts to individual customer needs. The […]

Evaluating the Evaluators: Metrics for Compositional Text-to-Image Generation
Seyed Amir Kasaei, Ali Aghayari, Arash Marioriyad, Niki Sepasian, MohammadAmin Fazli, Mahdieh Soleymani Baghshah, Mohammad Hossein Rohban
https://arxiv.org/abs/2509.21227

Evaluating the Evaluators: Metrics for Compositional Text-to-Image Generation
Text-image generation has advanced rapidly, but assessing whether outputs truly capture the objects, attributes, and relations described in prompts remains a central challenge. Evaluation in this space relies heavily on automated metrics, yet these are often adopted by convention or popularity rather than validated against human judgment. Because evaluation and reported progress in the field depend directly on these metrics, it is critical to understand how well they reflect human preferences. …