Tootfinder

Benefits and Pitfalls of Reinforcement Learning for Language Model Planning: A Theoretical Perspective
Siwei Wang, Yifei Shen, Haoran Sun, Shi Feng, Shang-Hua Teng, Li Dong, Yaru Hao, Wei Chen
https://arxiv.org/abs/2509.22613

Benefits and Pitfalls of Reinforcement Learning for Language Model Planning: A Theoretical Perspective
Recent reinforcement learning (RL) methods have substantially enhanced the planning capabilities of Large Language Models (LLMs), yet the theoretical basis for their effectiveness remains elusive. In this work, we investigate RL's benefits and limitations through a tractable graph-based abstraction, focusing on policy gradient (PG) and Q-learning methods. Our theoretical analyses reveal that supervised fine-tuning (SFT) may introduce co-occurrence-based spurious solutions, whereas RL achieves c…

Decision Transformer-Based Drone Trajectory Planning with Dynamic Safety-Efficiency Trade-Offs
Chang-Hun Ji, SiWoon Song, Youn-Hee Han, SungTae Moon
https://arxiv.org/abs/2507.21506

Decision Transformer-Based Drone Trajectory Planning with Dynamic Safety-Efficiency Trade-Offs
A drone trajectory planner should be able to dynamically adjust the safety-efficiency trade-off according to varying mission requirements in unknown environments. Although traditional polynomial-based planners offer computational efficiency and smooth trajectory generation, they require expert knowledge to tune multiple parameters to adjust this trade-off. Moreover, even with careful tuning, the resulting adjustment may fail to achieve the desired trade-off. Similarly, although reinforcement le…

ReCon-GS: Continuum-Preserved Guassian Streaming for Fast and Compact Reconstruction of Dynamic Scenes
Jiaye Fu, Qiankun Gao, Chengxiang Wen, Yanmin Wu, Siwei Ma, Jiaqi Zhang, Jian Zhang
https://arxiv.org/abs/2509.24325

ReCon-GS: Continuum-Preserved Guassian Streaming for Fast and Compact Reconstruction of Dynamic Scenes
Online free-viewpoint video (FVV) reconstruction is challenged by slow per-frame optimization, inconsistent motion estimation, and unsustainable storage demands. To address these challenges, we propose the Reconfigurable Continuum Gaussian Stream, dubbed ReCon-GS, a novel storage-aware framework that enables high fidelity online dynamic scene reconstruction and real-time rendering. Specifically, we dynamically allocate multi-level Anchor Gaussians in a density-adaptive fashion to capture inter-…

Time-resolved protoplanetary disk physics in DQ Tau with JWST
\'A. K\'osp\'al, P. \'Abrah\'am, V. V. Akimkin, L. Chen, J. Forbrich, K. V. Getman, B. Portilla-Revelo, D. Semenov, S. E. van Terwisga, J. Varga, L. Zwicky, G. G. Bal\'azs, Zs. Bora, \'A. Horti-D\'avid, A. P. Jo\'o, W. Og{\l}oza, B. Seli, M. Siwak, \'A. S\'odor, N. Tak\'acs

Time-resolved protoplanetary disk physics in DQ Tau with JWST
Accretion variability is ubiquitous in YSOs. While large outbursts may strongly affect the disk, the effects of moderate bursts are less understood. We aim to study the physical response of the disk around the eccentric binary system DQ Tau to its periodic accretion changes. We organized a multi-wavelength campaign centered on four JWST/MIRI spectra. We targeted three periastrons (high accretion state) and one apastron (quiescence). We used optical and near-infrared spectroscopy and photometry …

A Generalized $\chi_n$-Function
Cheng Lyu, Mu Yuan, Dabin Zheng, Siwei Sun, Shun Li
https://arxiv.org/abs/2509.20880 https://arxiv.org/pdf/2509.20880

A Generalized $χ_n$-Function
The mapping $χ_n$ from $\F_{2}^{n}$ to itself defined by $y=χ_n(x)$ with $y_i=x_i+x_{i+2}(1+x_{i+1})$, where the indices are computed modulo $n$, has been widely studied for its applications in lightweight cryptography. However, $χ_n $ is bijective on $\F_2^n$ only when $n$ is odd, restricting its use to odd-dimensional vector spaces over $\F_2$. To address this limitation, we introduce and analyze the generalized mapping $χ_{n, m}$ defined by $y=χ_{n,m}(x)$ with $y_i=x_i+x_{i+m} (x_{i+m-1…

Conditionally Whitened Generative Models for Probabilistic Time Series Forecasting
Yanfeng Yang, Siwei Chen, Pingping Hu, Zhaotong Shen, Yingjie Zhang, Zhuoran Sun, Shuai Li, Ziqi Chen, Kenji Fukumizu
https://arxiv.org/abs/2509.20928

Conditionally Whitened Generative Models for Probabilistic Time Series Forecasting
Probabilistic forecasting of multivariate time series is challenging due to non-stationarity, inter-variable dependencies, and distribution shifts. While recent diffusion and flow matching models have shown promise, they often ignore informative priors such as conditional means and covariances. In this work, we propose Conditionally Whitened Generative Models (CW-Gen), a framework that incorporates prior information through conditional whitening. Theoretically, we establish sufficient condition…

Code Difference Guided Fuzzing for FPGA Logic Synthesis Compilers via Bayesian Optimization
Zhihao Xu, Shikai Guo, Guilin Zhao, Peiyu Zou, Siwen Wang, Qian Ma, Hui Li, Furui Zhan
https://arxiv.org/abs/2508.17713

Code Difference Guided Fuzzing for FPGA Logic Synthesis Compilers via Bayesian Optimization
Field Programmable Gate Arrays (FPGAs) play a crucial role in Electronic Design Automation (EDA) applications, which have been widely used in safety-critical environments, including aerospace, chip manufacturing, and medical devices. A critical step in FPGA development is logic synthesis, which enables developers to translate their software designs into hardware net lists, which facilitates the physical implementation of the chip, detailed timing and power analysis, gate-level simulation, test …

Where Non-Invertible Symmetries End: Twist Defects for Electromagnetic Duality
Shu-Heng Shao, Siwei Zhong
https://arxiv.org/abs/2509.21279 https://arxiv.or…

Where Non-Invertible Symmetries End: Twist Defects for Electromagnetic Duality
We study novel conformal twist defects in 4d Maxwell theory, around which electric and magnetic fields are exchanged. These are codimension-2 defects living at the end of topological defects for certain non-invertible global symmetries. We determine the operator spectrum of the twist defect by solving classical electromagnetic wave equations subject to a twisted boundary condition. Using techniques from defect CFT, we show that correlation functions of these defect operators factorize into two …

Vier Tote bei russischem Angriff auf Kleinstadt
Bei dem russischen Drohnenangriff auf die nordukrainische Region Tschernihiw sind nach Angaben der örtlichen Behörden mindestens vier Zivilisten getötet worden. Bei gut 20 Einschlägen in der Stadt Nowhorod-Siwerskyj seien weitere sieben Menschen verletzt worden, teilte der Gouverneur des Gebiets Tschernihiw, Wjatscheslaw Tschaus, bei Telegram mit. Es g…
📷

Generalized Deep Multi-view Clustering via Causal Learning with Partially Aligned Cross-view Correspondence
Xihong Yang, Siwei Wang, Jiaqi Jin, Fangdi Wang, Tianrui Liu, Yueming Jin, Xinwang Liu, En Zhu, Kunlun He
https://arxiv.org/abs/2509.16022

Generalized Deep Multi-view Clustering via Causal Learning with Partially Aligned Cross-view Correspondence
Multi-view clustering (MVC) aims to explore the common clustering structure across multiple views. Many existing MVC methods heavily rely on the assumption of view consistency, where alignments for corresponding samples across different views are ordered in advance. However, real-world scenarios often present a challenge as only partial data is consistently aligned across different views, restricting the overall clustering performance. In this work, we consider the model performance decreasing …