Tootfinder

Il y a, paraît-il, des gens qui écrivent le livre qu'ils voudraient lire (transposable Š d'autres arts, je suppose). C'est peut-être pour ça qu'il m'est Š moi si difficile d'écrire: par d'heureux (?) hasards, je tombe régulièrement sur le livre d'un autre qui est celui que j'aurais voulu écrire, voire suis en train d'écrire.
Dernier en date : Don DeLillo, "Body Art" ( je le commence tout juste, peut-être finira-t-il par me décevoir).

FLAMES: Improving LLM Math Reasoning via a Fine-Grained Analysis of the Data Synthesis Pipeline
Parker Seegmiller, Kartik Mehta, Soumya Saha, Chenyang Tao, Shereen Oraby, Arpit Gupta, Tagyoung Chung, Mohit Bansal, Nanyun Peng
https://arxiv.org/abs/2508.16514

FLAMES: Improving LLM Math Reasoning via a Fine-Grained Analysis of the Data Synthesis Pipeline
Recent works improving LLM math reasoning with synthetic data have used unique setups, making comparison of data synthesis strategies impractical. This leaves many unanswered questions about the roles of different factors in the synthetic data pipeline, such as the impact of filtering low-quality problems. To address this gap, we introduce FLAMES, a Framework for LLM Assessment of Math rEasoning Data Synthesis, and perform a systematic study of 10 existing data synthesis strategies and multiple…

Breaking the Layer Barrier: Remodeling Private Transformer Inference with Hybrid CKKS and MPC
Tianshi Xu, Wen-jie Lu, Jiangrui Yu, Chen Yi, Chenqi Lin, Runsheng Wang, Meng Li
https://arxiv.org/abs/2508.19525

Breaking the Layer Barrier: Remodeling Private Transformer Inference with Hybrid CKKS and MPC
This paper presents an efficient framework for private Transformer inference that combines Homomorphic Encryption (HE) and Secure Multi-party Computation (MPC) to protect data privacy. Existing methods often leverage HE for linear layers (e.g., matrix multiplications) and MPC for non-linear layers (e.g., Softmax activation functions), but the conversion between HE and MPC introduces significant communication costs. The proposed framework, dubbed BLB, overcomes this by breaking down layers into …

"L'économie ressemble aujourd'hui Š la théologie catholique de l'Europe médiévale : une doctrine rigide gardée par une prêtrise moderne qui prétend détenir la seule vérité. Les dissidents sont évités. On dit aux non-économistes de "penser comme un économiste" ou de ne pas penser du tout. Ce n'est pas de l'éducation. C'est de l'endoctrinement." (Ha-Joon Chang)

Simultaneous estimation of the effective reproduction number and the time series of daily infections: Application to Covid-19
Hans R. K\"unsch, Fabio Sigrist
https://arxiv.org/abs/2506.21027

Simultaneous estimation of the effective reproduction number and the time series of daily infections: Application to Covid-19
The time varying effective reproduction number is an important parameter for communication and policy decisions during an epidemic. It is difficult to estimate because it depends on latent variables such as new infections and other characteristics of an epidemic which have to be inferred from available data. In this paper, we present new statistical methods for a popular model which defines the effective reproduction number based on self-exciting dynamics of new infections. Such a model is conc…

One more entry in my epic date parsing adventure. I’ve shipped a breaking change to this library to ensure that all valid date strings supported in Eleventy v4 will be RFC 9557-compatible, ensuring friendliness with the standardized Temporal APIs coming to a JavaScript runtime near you. Our full test suite runs against both!
Read (a little) more:

DRARL: Disengagement-Reason-Augmented Reinforcement Learning for Efficient Improvement of Autonomous Driving Policy
Weitao Zhou, Bo Zhang, Zhong Cao, Xiang Li, Qian Cheng, Chunyang Liu, Yaqin Zhang, Diange Yang
https://arxiv.org/abs/2506.16720

DRARL: Disengagement-Reason-Augmented Reinforcement Learning for Efficient Improvement of Autonomous Driving Policy
With the increasing presence of automated vehicles on open roads under driver supervision, disengagement cases are becoming more prevalent. While some data-driven planning systems attempt to directly utilize these disengagement cases for policy improvement, the inherent scarcity of disengagement data (often occurring as a single instances) restricts training effectiveness. Furthermore, some disengagement data should be excluded since the disengagement may not always come from the failure of dri…

Enhanced Velocity-Adaptive Scheme: Joint Fair Access and Age of Information Optimization in Vehicular Networks
Xiao Xu, Qiong Wu, Pingyi Fan, Kezhi Wang, Nan Cheng, Wen Chen, Khaled B. Letaief
https://arxiv.org/abs/2507.18328

Enhanced Velocity-Adaptive Scheme: Joint Fair Access and Age of Information Optimization in Vehicular Networks
In this paper, we consider the fair access problem and the Age of Information (AoI) under 5G New Radio (NR) Vehicle-to-Infrastructure (V2I) Mode 2 in vehicular networks. Specifically, vehicles follow Mode 2 to communicate with Roadside Units (RSUs) to obtain accurate data for driving assistance.Nevertheless, vehicles often have different velocity when they are moving in adjacent lanes, leading to difference in RSU dwelltime and communication duration. This results in unfair access to network re…

CLAImate: AI-Enabled Climate Change Communication through Personalized and Localized Narrative Visualizations
Mashrur Rashik, Jean-Daniel Fekete, Narges Mahyar
https://arxiv.org/abs/2507.11677

CLAImate: AI-Enabled Climate Change Communication through Personalized and Localized Narrative Visualizations
Communicating climate change remains challenging, as climate reports, though rich in data and visualizations, often feel too abstract or technical for the public. Although personalization can enhance communication, most tools still lack the narrative and visualization tailoring needed to connect with individual experiences. We present CLAImate, an AI-enabled prototype that personalizes conversation narratives and localizes visualizations based on users' climate knowledge and geographic location…

Emission Impossible: privacy-preserving carbon emissions claims
Jessica Man, Sadiq Jaffer, Patrick Ferris, Martin Kleppmann, Anil Madhavapeddy
https://arxiv.org/abs/2506.16347

Emission Impossible: privacy-preserving carbon emissions claims
Information and Communication Technologies (ICT) have a significant climate impact, and data centres account for a large proportion of the carbon emissions from ICT. To achieve sustainability goals, it is important that all parties involved in ICT supply chains can track and share accurate carbon emissions data with their customers, investors, and the authorities. However, businesses have strong incentives to make their numbers look good, whilst less so to publish their accounting methods along…