Max Andersson | Searcharxiv

arXiv2025

Temperature dependence of the long-term annealing behavior of neutron irradiated diodes from 8-inch p-type silicon wafers

To face the higher levels of radiation due to the 10-fold increase in integrated luminosity during the High-Luminosity LHC, the CMS detector will replace the current Calorimeter Endcap (CE) using the High-Granularity Calorimeter (HGCAL) concept. The high-radiation regions of the the CE, where fluences between $1\cdot10^{14}~n_{eq}/cm^{2}$ and $1\cdot10^{16}~n_{eq}/cm^{2}$ and doses of up to 2\,MGy are expected considering an integrated luminosity of $3\,ab^{-1}$, will be equipped with silicon pad sensors. This includes the entire electromagnetic as well as parts of the hadronic section of the CE. The silicon sensors are processed on 8-inch p-type wafers with an active thickness of 300\,\textmu m, 200\,\textmu m and 120\,\textmu m and cut into hexagonal shapes for optimal use of the full wafer area and tiling. With each main sensor, several small test structures (e.g. pad diodes) are hosted on the wafers, used for quality assurance and radiation hardness tests. In order to investigate the radiation-induced bulk damage, these diodes have been irradiated with reactor neutrons at JSI (Jozef Stefan Institute, Ljubljana, Slovenia) to fluences between $5\cdot10^{14}~n_{eq}/cm^{2}$ and $1.5\cdot10^{16}~n_{eq}/cm^{2}$. This study focuses on the isothermal annealing behavior of the bulk material at different temperatures between 5.5{\deg}C and 60{\deg}C using electrical characterization and charge collection measurements. The results are used to extract the annealing time constants for this material and fluence range based on the Hamburg model approach to allow an estimation of the expected annealing effects in silicon sensors during the year-end technical stops and the long HL-LHC shutdowns. The annealing parameters found will make it possible to model the annealing behavior of p-type silicon detector projects at HL-LHC fluence ranges better than the existing Hamburg model.

physics.ins-det↗

arXiv2025

Exploiting the Asymmetric Uncertainty Structure of Pre-trained VLMs on the Unit Hypersphere

Vision-language models (VLMs) as foundation models have significantly enhanced performance across a wide range of visual and textual tasks, without requiring large-scale training from scratch for downstream tasks. However, these deterministic VLMs fail to capture the inherent ambiguity and uncertainty in natural language and visual data. Recent probabilistic post-hoc adaptation methods address this by mapping deterministic embeddings onto probability distributions; however, existing approaches do not account for the asymmetric uncertainty structure of the modalities, and the constraint that meaningful deterministic embeddings reside on a unit hypersphere, potentially leading to suboptimal performance. In this paper, we address the asymmetric uncertainty structure inherent in textual and visual data, and propose AsymVLM to build probabilistic embeddings from pre-trained VLMs on the unit hypersphere, enabling uncertainty quantification. We validate the effectiveness of the probabilistic embeddings on established benchmarks, and present comprehensive ablation studies demonstrating the inherent nature of asymmetry in the uncertainty structure of textual and visual data.

cs.LG↗