The paper presents candela, a differentiable SIREN neural field that learns the photon Green's function for the IceCube Neutrino Observatory. It predicts photon yield and full arrival-time distribution for point-like energy deposits, enabling complete event simulation by superposing responses from multiple deposits. Trained on Monte‑Carlo data, candela produces events 50–100× faster than existing methods while maintaining median yields within 2% of MC expectations and timing distributions at the MC statistical floor across six photon‑count decades.
By Felix J. Yu, Berthy T. Feng, Nicholas Kamp, Carlos A. Arg\"{u}elles
The paper applies sparse autoencoders to a neutrino foundation model trained on IceCube data, uncovering a validated atlas of physical concepts within the model’s internal representation. Causal analysis shows the direction reconstruction head largely ignores this atlas, whereas an uncertainty head trained on the same representation effectively uses quality and brightness features, improving angular resolution from 20.2° to 3.2° at 20% efficiency. These findings demonstrate that mechanistic interpretability can expose latent physics and guide the design of downstream tasks.
By Rapha\"el Bonnet-Guerrini, Johann Ioannou-Nikolaides, Inar Timiryasov, Vincenzo Piuri
arXiv:2512. 06208v3 Announce Type: replace-cross Abstract: Inference of standard convolutional neural networks (CNNs) on FPGAs often incurs high latency and a long initiation interval due to the deep nested loops required to densely convolve every input pixel regardless of its feature value.
By Ho Fung Tsoi, Dylan Rankin, Vladimir Loncar, Philip Harris
The paper presents deep‑learning trigger algorithms for the Hyper‑Kamiokande water Cherenkov detector, targeting low‑energy neutrino events below 7 MeV. It compares a supervised neural‑network classifier with two anomaly‑detection methods—an autoencoder and a Manifold Projection‑Diffusion Recovery model—showing the supervised model achieves a 76.7 % signal efficiency for 3 MeV electrons, far surpassing the 26.4 % efficiency of a traditional hit‑count trigger. GPU‑based runtime tests indicate per‑window inference latencies well below one millisecond.
By Katharina Lachner, Sa\'ul Alonso-Monsalve, Benjamin Richards, Davide Sgalaberna
The paper compares convolutional neural networks (CNN), Vision Transformers (ViT), and hierarchical Swin Transformers for quark‑gluon jet classification using a three‑channel jet‑image representation. CNN and Swin models outperform ViT, indicating that local jet substructure is crucial for discrimination. The study also shows that block‑wise fine‑tuning, Momentum Contrast pretraining, and a compact Swin variant can improve performance while reducing parameters.
By Daeun Kim, Jaeyoon Cho, Jiwon Lee, Wonjun Jeong, Hyeongwoo Noh, Giyeong Kim, Seunghwan Yang, MinJung Kweon
The rapid proliferation of generative models raises the model attribution problem: given only an image, can we determine which model produced it? Existing methods have grown as elaborate as the genera...