-
SeePhys: Does Seeing Help Thinking? -- Benchmarking Vision-Based Physics Reasoning
Authors:
Kun Xiang,
Heng Li,
Terry Jingchen Zhang,
Yinya Huang,
Zirong Liu,
Peixin Qu,
Jixi He,
Jiaqi Chen,
Yu-Jie Yuan,
Jianhua Han,
Hang Xu,
Hanhui Li,
Mrinmaya Sachan,
Xiaodan Liang
Abstract:
We present SeePhys, a large-scale multimodal benchmark for LLM reasoning grounded in physics questions ranging from middle school to PhD qualifying exams. The benchmark covers 7 fundamental domains spanning the physics discipline, incorporating 21 categories of highly heterogeneous diagrams. In contrast to prior works where visual elements mainly serve auxiliary purposes, our benchmark features a…
▽ More
We present SeePhys, a large-scale multimodal benchmark for LLM reasoning grounded in physics questions ranging from middle school to PhD qualifying exams. The benchmark covers 7 fundamental domains spanning the physics discipline, incorporating 21 categories of highly heterogeneous diagrams. In contrast to prior works where visual elements mainly serve auxiliary purposes, our benchmark features a substantial proportion of vision-essential problems (75%) that mandate visual information extraction for correct solutions. Through extensive evaluation, we observe that even the most advanced visual reasoning models (e.g., Gemini-2.5-pro and o4-mini) achieve sub-60% accuracy on our benchmark. These results reveal fundamental challenges in current large language models' visual understanding capabilities, particularly in: (i) establishing rigorous coupling between diagram interpretation and physics reasoning, and (ii) overcoming their persistent reliance on textual cues as cognitive shortcuts.
△ Less
Submitted 17 June, 2025; v1 submitted 25 May, 2025;
originally announced May 2025.
-
Dark Current and Multipacting Capabilities in OPAL: Model Benchmarks and Applications
Authors:
C. Wang,
A. Adelmann,
Z. G. Yin,
T. J. Zhang
Abstract:
Dark current and multiple electron impacts (multipacting), as for example observed in radio frequency (RF) structures of accelerators, are usually harmful to the equipment and the beam quality. These effects need to be suppressed to guarantee efficient and stable operation. Large scale simulations can be used to understand causes and develop strategies to suppress these phenomenas.
We extend \op…
▽ More
Dark current and multiple electron impacts (multipacting), as for example observed in radio frequency (RF) structures of accelerators, are usually harmful to the equipment and the beam quality. These effects need to be suppressed to guarantee efficient and stable operation. Large scale simulations can be used to understand causes and develop strategies to suppress these phenomenas.
We extend \opal, a parallel framework for charged particle optics in accelerator structures and beam lines, with the necessary physics models to efficiently and precisely simulate multipacting phenomenas. We added a Fowler-Nordheim field emission model, two secondary electron emission models, developed by Furman-Pivi and Vaughan respectively, as well as efficient 3D boundary geometry handling capabilities. The models and their implementation are carefully benchmark against a non-stationary multipacting theory for the classic parallel plate geometry. A dedicated, parallel plate experiment is sketched.
△ Less
Submitted 31 August, 2012;
originally announced August 2012.
-
Towards Quantitative Simulations of High Power Proton Cyclotrons
Authors:
Y. J. Bi,
A. Adelmann,
R. Dölling,
M. Humbel,
W. Joho,
M. Seidel,
T. J. Zhang
Abstract:
PSI operates a cyclotron based high intensity proton accelerator routinely at an average beam power of 1.3MW. With this power the facility is at the worldwide forefront of high intensity proton accelerators. The beam current is practically limited by losses at extraction and the resulting activation of accelerator components. Further intensity upgrades and new projects aiming at an even higher ave…
▽ More
PSI operates a cyclotron based high intensity proton accelerator routinely at an average beam power of 1.3MW. With this power the facility is at the worldwide forefront of high intensity proton accelerators. The beam current is practically limited by losses at extraction and the resulting activation of accelerator components. Further intensity upgrades and new projects aiming at an even higher average beam power, are only possible if the relative losses can be lowered in proportion, thus keeping absolute losses at a constant level. Maintaining beam losses at levels allowing hands-on maintenance is a primary challenge in any high power proton machine design and operation. In consequence, predicting beam halo at these levels is a great challenge and will be addressed in this paper. High power hadron driver have being used in many disciplines of science and, a growing interest in the cyclotron technology for high power hadron drivers are being observed very recently. This report will briefly introduce OPAL, a tool for precise beam dynamics simulations including 3D space charge. One of OPAL's flavors (OPAL-cycl) is dedicated to high power cyclotron modeling and is explained in greater detail. We then explain how to obtain initial conditions for our PSI Ring cyclotron which still delivers the world record in beam power of 1.3 MW continuous wave (cw). Several crucial steps are explained necessary to be able to predict tails at the level of 3σ... 4σin the PSI Ring cyclotron. We compare our results at the extraction with measurements, obtained with a 1.18 MW cw production beam. Based on measurement data, we develop a simple linear model to predict beam sizes of the extracted beam as a function of intensities and confirm the model with simulations.
△ Less
Submitted 21 December, 2010; v1 submitted 3 December, 2010;
originally announced December 2010.
-
Beam Dynamics in High Intensity Cyclotrons Including Neighboring Bunch Effects: Model, Implementation and Application
Authors:
J. J. Yang,
A. Adelmann,
M. Humbel,
M. Seidel,
T. J. Zhang
Abstract:
Space charge effects, being one of the most significant collective effects, play an important role in high intensity cyclotrons. However, for cyclotrons with small turn separation, other existing effects are of equal importance. Interactions of radially neighboring bunches are also present, but their combined effects has not yet been investigated in any great detail. In this paper, a new particl…
▽ More
Space charge effects, being one of the most significant collective effects, play an important role in high intensity cyclotrons. However, for cyclotrons with small turn separation, other existing effects are of equal importance. Interactions of radially neighboring bunches are also present, but their combined effects has not yet been investigated in any great detail. In this paper, a new particle in cell based self-consistent numerical simulation model is presented for the first time. The model covers neighboring bunch effects and is implemented in the three-dimensional object-oriented parallel code OPAL-cycl, a flavor of the OPAL framework. We discuss this model together with its implementation and validation. Simulation results are presented from the PSI 590 MeV Ring Cyclotron in the context of the ongoing high intensity upgrade program, which aims to provide a beam power of 1.8 MW (CW) at the target destination.
△ Less
Submitted 1 March, 2010;
originally announced March 2010.
