Tootfinder

Smart Meters Conference, Salford (2019) - Bringing together industry, academia and policy makers to innovation around the smart meter infrastructure. #UoSSmart2019 - https://www.earth.org.uk/Smart-Meters-…

Smart Meters Conference, Salford (2019)
Bringing together industry, academia and policy makers to innovation around the smart meter infrastructure. #UoSSmart2019

Winter Soldier: Backdooring Language Models at Pre-Training with Indirect Data Poisoning
Wassim Bouaziz, Mathurin Videau, Nicolas Usunier, El-Mahdi El-Mhamdi
https://arxiv.org/abs/2506.14913

Winter Soldier: Backdooring Language Models at Pre-Training with Indirect Data Poisoning
The pre-training of large language models (LLMs) relies on massive text datasets sourced from diverse and difficult-to-curate origins. Although membership inference attacks and hidden canaries have been explored to trace data usage, such methods rely on memorization of training data, which LM providers try to limit. In this work, we demonstrate that indirect data poisoning (where the targeted behavior is absent from training data) is not only feasible but also allow to effectively protect a dat…

Adjustment for Confounding using Pre-Trained Representations
Rickmer Schulte, David R\"ugamer, Thomas Nagler
https://arxiv.org/abs/2506.14329 https://…

Adjustment for Confounding using Pre-Trained Representations
There is growing interest in extending average treatment effect (ATE) estimation to incorporate non-tabular data, such as images and text, which may act as sources of confounding. Neglecting these effects risks biased results and flawed scientific conclusions. However, incorporating non-tabular data necessitates sophisticated feature extractors, often in combination with ideas of transfer learning. In this work, we investigate how latent features from pre-trained neural networks can be leverage…

Enhancing Traffic Accident Classifications: Application of NLP Methods for City Safety
Enes \"Ozeren, Alexander Ulbrich, Sascha Filimon, David R\"ugamer, Andreas Bender
https://arxiv.org/abs/2506.12092

Enhancing Traffic Accident Classifications: Application of NLP Methods for City Safety
A comprehensive understanding of traffic accidents is essential for improving city safety and informing policy decisions. In this study, we analyze traffic incidents in Munich to identify patterns and characteristics that distinguish different types of accidents. The dataset consists of both structured tabular features, such as location, time, and weather conditions, as well as unstructured free-text descriptions detailing the circumstances of each accident. Each incident is categorized into on…

Unusual Valence of Ru and Prediction of Magnetism, Anomalous Hall Conductivity in a Newly Synthesized Double Perovskite Compound Ca_2CoRuO_6
Koushik Pradhan, Soumya Ghorai, Prabuddha Sanyal, Ryan Morrow, Bernd B\"uchner, Thirupathaiah Setti, Tanusri Saha Dasgupta
https://arxiv.org/abs/2506.09423

Unusual Valence of Ru and Prediction of Magnetism, Anomalous Hall Conductivity in a Newly Synthesized Double Perovskite Compound Ca_2CoRuO_6
With a goal to expand on the family of double perovskite compounds, hosting 3d transition metal and 4d or 5d transition metal, two new ordered double perovskite compounds, Ca$_2$FeRuO$_6$ and Ca$_2$CoRuO$_6$ are synthesized following the prediction of a recent high throughput machine-learning study [Phys. Rev. Materials 3, 084418]. Experimentally both compounds are found to stabilize in monoclinic symmetry, which is consistent with the high-throughput prediction for Ca$_2$FeRuO$_6$, but at odd …

Optimization and validation of charge transport simulation for hybrid pixel detectors incorporating the repulsion effect
X. Xie, A. Bergamaschi, M. Br\"uckner, M. Carulla, R. Dinapoli, S. Ebner, K. Ferjaoui, A. Francesca Mazzoleni, J. Franklin Mulvey, V. Gautam, D. Greiffenberg, S. Hasanaj, J. Heymes, V. Hinger, V. Kedych, T. King, S. Li, C. Lopez-Cuenca, D. Mezza, K. Moustakas, A. Mozzanica, M. M\"uller, K. A. Paton, C. Ruder, B. Schmitt, P. Sieberer, S. Silletta, D. Thattil…

Optimization and validation of charge transport simulation for hybrid pixel detectors incorporating the repulsion effect
For emerging applications of hybrid pixel detectors which require high spatial resolution, e.g., subpixel interpolation in X-ray imaging and deep learning-based electron localization, accurate modeling of charge transport processes in the sensor is highly demanded. To address this, two open-source, time-stepping Monte Carlo simulation methods have been developed, both explicitly incorporating charge repulsion, which are found necessary for accurate simulation when charge sharing becomes importa…

🇺🇦 #NowPlaying on KEXP's #DriveTime
Black Country, New Road:
🎵 Two Horses
#BlackCountry #NewRoad
https://blackcountrynewroad.bandcamp.com/track/two-horses
https://open.spotify.com/track/0U8E6CEaYmRwpm2JOFCER5

This https://arxiv.org/abs/2402.18306 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_…

Calibration requirement for Epoch of Reionization 21-cm signal observation -- III. Bias and variance in uGMRT ELAIS-N1 field power spectrum
Power spectrum of \HI 21-cm radiation is one of the promising probes to study large scale structure of the universe and understand galaxy formation and evolution. The presence of foregrounds, that are orders of magnitude larger in the same frequency range of the redshifted 21-cm signal has been one of the largest observational challenges. The foreground contamination also hinders the calibration procedures and introduces residual calibration errors in the interferometric data. It has been shown…