Tootfinder

Beyond Random Masking: A Dual-Stream Approach for Rotation-Invariant Point Cloud Masked Autoencoders
Xuanhua Yin, Dingxin Zhang, Yu Feng, Shunqi Mao, Jianhui Yu, Weidong Cai
https://arxiv.org/abs/2509.14975

Beyond Random Masking: A Dual-Stream Approach for Rotation-Invariant Point Cloud Masked Autoencoders
Existing rotation-invariant point cloud masked autoencoders (MAE) rely on random masking strategies that overlook geometric structure and semantic coherence. Random masking treats patches independently, failing to capture spatial relationships consistent across orientations and overlooking semantic object parts that maintain identity regardless of rotation. We propose a dual-stream masking approach combining 3D Spatial Grid Masking and Progressive Semantic Masking to address these fundamental l…

HBAR entropy in causal diamond geometry: A near-horizon perspective
Nada Eissa, Carlos R. Ord\'o\~nez, Gustavo Valdivia-Mera
https://arxiv.org/abs/2508.13493 https://…

HBAR entropy in causal diamond geometry: A near-horizon perspective
In this article we extend the Horizon Brightened Acceleration Radiation (HBAR) framework to the causal diamond (CD) spacetime. We study a cloud of two-level atoms, injected at random times in the asymptotic past of the CD, freely falling toward its causal horizon and emitting scalar radiation via a weak dipole coupling to a quantum field. In the near-horizon region, an emergent conformal symmetry-captured by conformal quantum mechanics (CQM)-governs the field dynamics and allows analytic contro…

Charge accumulation by Direct Magnetoelectric Effect in ScAlN/Ni Nanoscale Devices
Federica Luciano, Emma Van Meirvenne, Ephraim Spindler, Philipp Pirro, Bart Sor\'ee, Mathias Weiler, Stefan De Gendt, Florin Ciubotaru, Christoph Adelmann
https://arxiv.org/abs/2508.13674

Charge accumulation by Direct Magnetoelectric Effect in ScAlN/Ni Nanoscale Devices
This work investigates the direct magnetoelectric effect in thin-film lab scale composite heterostructures comprising a 100 nm thick piezoelectric Sc0.4Al0.6N (ScAlN) and a magnetostrictive Ni with 100-200 nm thickness, fabricated on Si/SiO2 substrates. The films are patterned into square pillar arrays with lateral dimensions down to 500 nm x 500 nm. Vibrating sample magnetometry (VSM) measurements reveal in-plane magnetic anisotropy in the Ni films, attributed to strain induced by the underlyi…

Dispersion of collective modes in spinful fractional quantum Hall states on the sphere
Rakesh K. Dora, Ajit C. Balram
https://arxiv.org/abs/2509.13100 https://

Dispersion of collective modes in spinful fractional quantum Hall states on the sphere
Collective modes capture the dynamical aspects of fractional quantum Hall (FQH) fluids. Depending on the active degrees of freedom, different types of collective modes can arise in a FQH state. In this work, we consider spinful FQH states in the lowest Landau level (LLL) along the Jain sequence of fillings $ν{=}n/(2n{\pm}1)$ and compute the Coulomb dispersion of their spin-flip and spin-conserving collective modes in the spherical geometry. We use the LLL-projected density-wave and composite f…

Spectrometry of Captured Highly Charged Ions Produced Following Antiproton Annihilations
F. P. Gustafsson, M. Volponi, J. Zielinski, A. Asare, I. Hwang, S. Alfaro Campos, M. Auzins, D. Bhanushali, A. Bhartia, M. Berghold, R. S. Brusa, K. Calik, A. Camper, R. Caravita, F. Castelli, G. Cerchiari, S. Chandran, A. Chehaimi, S. Choudapurkar, R. Ciury{\l}o, P. Conte, G. Consolati, M. Doser, R. Ferguson, M. Germann, A. Giszczak, L. T. Gl\"oggler, {\L}. Graczykowski, M. Grosbart, F. Guati…

Spectrometry of Captured Highly Charged Ions Produced Following Antiproton Annihilations
We report the first capture and time-of-flight spectrometry of highly charged ions produced following antiproton annihilations in a Penning-Malmberg trap. At the AEgIS experiment, we employed a multi-step nested-trap technique to isolate ions from antiproton annihilations with ultra-low-density helium and argon gas. The capture and identification of highly-charged argon ions in charge-states up to $Ar^{5+}$ demonstrates a new method for in-trap synthesis. This work establishes a clear path towa…

Towards Adaptive ML Benchmarks: Web-Agent-Driven Construction, Domain Expansion, and Metric Optimization
Hangyi Jia, Yuxi Qian, Hanwen Tong, Xinhui Wu, Lin Chen, Feng Wei
https://arxiv.org/abs/2509.09321

Towards Adaptive ML Benchmarks: Web-Agent-Driven Construction, Domain Expansion, and Metric Optimization
Recent advances in large language models (LLMs) have enabled the emergence of general-purpose agents for automating end-to-end machine learning (ML) workflows, including data analysis, feature engineering, model training, and competition solving. However, existing benchmarks remain limited in task coverage, domain diversity, difficulty modeling, and evaluation rigor, failing to capture the full capabilities of such agents in realistic settings. We present TAM Bench, a diverse, realistic, and st…

A Quantum Generative Framework for Modeling Single-Cell Transcriptomes with Gene-Gene and Cell-Cell Interactions
Selim Romero, Vignesh Kumar, Robert S. Chapkin, James J. Cai
https://arxiv.org/abs/2510.12776

A Quantum Generative Framework for Modeling Single-Cell Transcriptomes with Gene-Gene and Cell-Cell Interactions
Single-cell RNA sequencing (scRNA-seq) data simulation is limited by classical methods that rely on linear correlations, failing to capture the intrinsic, nonlinear dependencies and the simultaneous gene-gene and cell-cell interactions. We introduce qSimCells, a novel hybrid quantum-classical simulator that leverages quantum entanglement to model single-cell transcriptomes. The core innovation is a quantum kernel that uses a parameterized quantum circuit with CNOT gates to encode complex, nonli…

Decoupled Multimodal Fusion for User Interest Modeling in Click-Through Rate Prediction
Alin Fan, Hanqing Li, Sihan Lu, Jingsong Yuan, Jiandong Zhang
https://arxiv.org/abs/2510.11066

Decoupled Multimodal Fusion for User Interest Modeling in Click-Through Rate Prediction
Modern industrial recommendation systems improve recommendation performance by integrating multimodal representations from pre-trained models into ID-based Click-Through Rate (CTR) prediction frameworks. However, existing approaches typically adopt modality-centric modeling strategies that process ID-based and multimodal embeddings independently, failing to capture fine-grained interactions between content semantics and behavioral signals. In this paper, we propose Decoupled Multimodal Fusion (…

Bidirectional Representations Augmented Autoregressive Biological Sequence Generation:Application in De Novo Peptide Sequencing
Xiang Zhang, Jiaqi Wei, Zijie Qiu, Sheng Xu, Zhi Jin, ZhiQiang Gao, Nanqing Dong, Siqi Sun
https://arxiv.org/abs/2510.08169

Bidirectional Representations Augmented Autoregressive Biological Sequence Generation:Application in De Novo Peptide Sequencing
Autoregressive (AR) models, common in sequence generation, are limited in many biological tasks such as de novo peptide sequencing and protein modeling by their unidirectional nature, failing to capture crucial global bidirectional token dependencies. Non-Autoregressive (NAR) models offer holistic, bidirectional representations but face challenges with generative coherence and scalability. To transcend this, we propose a hybrid framework enhancing AR generation by dynamically integrating rich c…

NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents
Tianshi Zheng, Kelvin Kiu-Wai Tam, Newt Hue-Nam K. Nguyen, Baixuan Xu, Zhaowei Wang, Jiayang Cheng, Hong Ting Tsang, Weiqi Wang, Jiaxin Bai, Tianqing Fang, Yangqiu Song, Ginny Y. Wong, Simon See
https://arxiv.org/abs/2510.07172

NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents
Large language models are emerging as powerful tools for scientific law discovery, a foundational challenge in AI-driven science. However, existing benchmarks for this task suffer from a fundamental methodological trilemma, forcing a trade-off between scientific relevance, scalability, and resistance to memorization. Furthermore, they oversimplify discovery as static function fitting, failing to capture the authentic scientific process of uncovering embedded laws through the interactive explora…