arXiv subjects

Yi Jia

Publications and source records attributed to Yi Jia.

3 recordsLinked to original sources

arXiv2026

From Data Completeness to Data Sufficiency: A Task-Driven Imaging Framework for Intraoperative CBCT under Quality-Time-Dose Trade-offs

Mobile C-arm cone-beam computed tomography (CBCT) has been widely used for real-time intraoperative 3D imaging. However, current practice often mechanically applies the fan-beam CT criterion of "180{\deg} plus fan angle" in pursuit of "data completeness" in reconstruction. This review argues that, under the single circular trajectory of three-dimensional cone-beam geometry, complete data are mathematically unattainable; moreover, blindly increasing sampling may exacerbate the trade-off among intraoperative image quality (Q), imaging time (T), and radiation dose (D). Against this background, this review reframes the evaluation of intraoperative CBCT around "data sufficiency" rather than "data completeness." This perspective moves beyond the excessive pursuit of absolute mathematical and analytic accuracy, and instead emphasizes task-specific minimum image-quality thresholds required for clinical decision-making. By synthesizing evidence from multiple clinical scenarios, this review suggests that approximation errors can be acceptable when clinical decision-making requirements are satisfied, thereby achieving a Q-T-D balance.

eess.IV↗

arXiv2025

Searching for GeV Gamma-Ray Polarization and Axion-Like Particles with AMS-02

We study the detectability of GeV-band gamma-ray polarization with the AMS-02 experiment and its proposed successor AMS-100, from Galactic and extragalactic sources. Characterizing gamma-ray polarization in this energy range could shed light on gamma-ray emission mechanisms in the sources; physics beyond the Standard Model, such as the presence of axion-like particles (ALPs), could also induce a distinctive energy-dependent polarization signal due to propagation effects in magnetic fields. We present estimates for the minimum detectable polarization from bright sources and the forecast reach for axion-like particles (ALPs). We show that AMS-02 will have sensitivity to gamma-ray polarization only for the brightest steady-state Galactic sources, such as the Vela and Geminga pulsars; it is not expected to be capable of detecting polarization in Galactic or extragalactic sources that have been previously proposed as good targets for ALP searches with gamma-ray intensity measurements. However, AMS-100 observing the extragalactic source NGC1275 would be expected to probe new parameter space even for unfavorable B-field models, with prospects to measure the energy-dependence of such a signal. For Galactic sources, polarization measurements could provide a unique test of scenarios where ALPs induce energy-dependent features in the photon intensity. However, in the absence of a bright transient source (such as a Galactic supernova), the parameter space that would be probed by this approach with ten years of AMS-100 data is already nominally excluded by other experiments, although this conflict may be avoided in specific ALP models.

hep-ph↗

arXiv2025

RLFactory: A Plug-and-Play Reinforcement Learning Post-Training Framework for LLM Multi-Turn Tool-Use

Large language models excel at basic reasoning but struggle with tasks that require interaction with external tools. We present RLFactory, a plug-and-play reinforcement learning post-training framework for multi-round tool use. RLFactory tackles (i) tool-call stability and adaptability amid tool heterogeneity and interface issues via an asyncio-based asynchronous caller and a decoupled tool/training architecture, and (ii) diverse evaluation needs via a reward layer supporting rule-based, model-judgment, and tool-verification signals. It reconstructs the MDP by introducing observation markers from tool feedback, closing the loop among model, tools, and environment, and implements a generate-parse-invoke-update workflow for dynamic policy optimization. On Search-R1 with Qwen3-4B, RLFactory achieves a 0.486 test score on the Natural Questions (NQ) dataset, surpassing larger models trained with similar techniques (e.g., Qwen2.5-7B-Instruct-GRPO at 0.473), and increases training throughput by 6.8x. RLFactory provides a low-barrier, highly adaptable framework for strengthening multi-round tool use of LLMs in real-world scenarios. Code: https://github.com/Simple-Efficient/RL-Factory.

cs.LG↗