arXiv subjects

Yifan Ge

Publications and source records attributed to Yifan Ge.

7 recordsLinked to original sources

arXiv2026

UniToolCall: Unifying Tool-Use Representation, Data, and Evaluation for LLM Agents

Tool-use capability is a fundamental component of LLM agents, enabling them to interact with external systems through structured function calls. However, existing research exhibits inconsistent interaction representations, largely overlooks the structural distribution of tool-use trajectories, and relies on incompatible evaluation benchmarks. We present UniToolCall, a unified framework for tool learning that standardizes the entire pipeline from toolset construction and dataset generation to evaluation. The framework curates a large tool pool of 22k+ tools and constructs a hybrid training corpus of 390k+ instances by combining 10 standardized public datasets with structurally controlled synthetic trajectories. It explicitly models diverse interaction patterns, including single-hop vs. multi-hop and single-turn vs. multi-turn, while capturing both serial and parallel execution structures. To support coherent multi-turn reasoning, we further introduce an Anchor Linkage mechanism that enforces cross-turn dependencies. Furthermore, we convert 7 public benchmarks into a unified Query--Action--Observation--Answer (QAOA) representation with fine-grained evaluation at the function-call, turn, and conversation levels. Experiments show that fine-tuning Qwen3-8B on our dataset substantially improves tool-use performance. Under the distractor-heavy Hybrid-20 setting, achieves 93.0% single-turn Strict Precision, outperforming commercial models including GPT, Gemini, and Claude.

cs.AI↗

arXiv2025

Decoupling Perception from Reasoning for Hallucination-Resistant Video Understanding

Video Large Language Models improve reasoning over complex videos by generating intermediate reasoning text. However, reliable reasoning depends on accurate video perception. In existing approaches, perception evidence is intertwined with reasoning text, making it difficult to directly supervise the perception process. We argue that reliable supervision requires explicitly separating perception evidence from reasoning so that perception can be verified independently. To supervise perception directly, we propose Decoupled Perception and Logic (DPL), which represents perception as fixed-format evidence units containing timestamps and visual descriptions. This structured representation enables direct extraction of perception content and simplifies alignment between video segments and reward evaluation. Building on DPL, we introduce a perception reward that encourages both hallucination resistance and perception-based reasoning. An Factual-Aware Evaluator (FAE) provides anti-hallucination scores and achieves hallucination evaluation performance comparable to GPT-4o. In addition, we validate reasoning consistency by feeding perception results and questions into a reference model. Experiments show that, by providing reliable process rewards, Video-DPL consistently improves post-training performance at both 3B and 7B scales, while delivering higher data efficiency.

cs.CV↗

arXiv2025

Medical Report Generation: A Hierarchical Task Structure-Based Cross-Modal Causal Intervention Framework

Medical Report Generation (MRG) is a key part of modern medical diagnostics, as it automatically generates reports from radiological images to reduce radiologists' burden. However, reliable MRG models for lesion description face three main challenges: insufficient domain knowledge understanding, poor text-visual entity embedding alignment, and spurious correlations from cross-modal biases. Previous work only addresses single challenges, while this paper tackles all three via a novel hierarchical task decomposition approach, proposing the HTSC-CIF framework. HTSC-CIF classifies the three challenges into low-, mid-, and high-level tasks: 1) Low-level: align medical entity features with spatial locations to enhance domain knowledge for visual encoders; 2) Mid-level: use Prefix Language Modeling (text) and Masked Image Modeling (images) to boost cross-modal alignment via mutual guidance; 3) High-level: a cross-modal causal intervention module (via front-door intervention) to reduce confounders and improve interpretability. Extensive experiments confirm HTSC-CIF's effectiveness, significantly outperforming state-of-the-art (SOTA) MRG methods. Code will be made public upon paper acceptance.

cs.CV↗

arXiv2024

High SNR 3D Imaging from Millimeter-scale Thick Tissues to Cellular Dynamics via Structured Illumination Microscopy

Three-dimensional (3D) fluorescence imaging provides a vital approach for study of biological tissues with intricate structures, and optical sectioning structured illumination microscopy (OS-SIM) stands out for its high imaging speed, low phototoxicity and high spatial resolution. However, OS-SIM faces the problem of low signal-to-noise ratio (SNR) when using traditional decoding algorithms, especially in thick tissues. Here we propose a Hilbert-transform decoding and space domain based high-low (HT-SHiLo) algorithm for noise suppression in OS-SIM. We demonstrate HT-SHiLo algorithm can significantly improve the SNR of optical sectioning images at rapid processing speed, and double the imaging depth in thick tissues. With our OS-SIM system, we achieve high quality 3D images of various biological samples including mouse brains, Drosophila clock neurons, organoids, and live cells. We anticipate that this approach will render OS-SIM a powerful technique for research of cellular organelles or thick tissues in 3D morphology.

physics.optics↗

arXiv2023

Reciprocal polarization imaging of complex media

The vectorial evolution of polarized light interaction with a medium can reveal its microstructure and anisotropy beyond what can be obtained from scalar light interaction. Anisotropic properties (diattenuation, retardance, and depolarization) of a complex medium can be quantified by polarization imaging by measuring the Mueller matrix. However, polarization imaging in the reflection geometry, ubiquitous and often preferred in diverse applications, has suffered a poor recovery of the medium's anisotropic properties due to the lack of suitable decomposition of the Mueller matrices measured inside a backward geometry. Here, we present reciprocal polarization imaging of complex media after introducing reciprocal polar decomposition for backscattering Mueller matrices. Based on the reciprocity of the optical wave in its forward and backward scattering paths, the anisotropic diattenuation, retardance, and depolarization of a complex medium are determined by measuring the backscattering Mueller matrix. We demonstrate reciprocal polarization imaging in various applications for quantifying complex non-chiral and chiral media (birefringence resolution target, tissue sections, and glucose suspension), uncovering their anisotropic microstructures with remarkable clarity and accuracy. We also highlight types of complex media that Lu-Chipman and differential decompositions of backscattering Mueller matrices lead to erroneous medium polarization properties, whereas reciprocal polar decomposition recovers properly. Reciprocal polarization imaging will be instrumental in imaging complex media from remote sensing to biomedicine and will open new applications of polarization optics in reflection geometry.

physics.optics↗

arXiv2021

Friends and Foes in Learning from Noisy Labels

Learning from examples with noisy labels has attracted increasing attention recently. But, this paper will show that the commonly used CIFAR-based datasets and the accuracy evaluation metric used in the literature are both inappropriate in this context. An alternative valid evaluation metric and new datasets are proposed in this paper to promote proper research and evaluation in this area. Then, friends and foes are identified from existing methods as technical components that are either beneficial or detrimental to deep learning from noisy labeled examples, respectively, and this paper improves and combines technical components from the friends category, including self-supervised learning, new warmup strategy, instance filtering and label correction. The resulting F&F method significantly outperforms existing methods on the proposed nCIFAR datasets and the real-world Clothing1M dataset.

cs.CV↗

arXiv2020

Distilling Knowledge by Mimicking Features

Knowledge distillation (KD) is a popular method to train efficient networks ("student") with the help of high-capacity networks ("teacher"). Traditional methods use the teacher's soft logits as extra supervision to train the student network. In this paper, we argue that it is more advantageous to make the student mimic the teacher's features in the penultimate layer. Not only the student can directly learn more effective information from the teacher feature, feature mimicking can also be applied for teachers trained without a softmax layer. Experiments show that it can achieve higher accuracy than traditional KD. To further facilitate feature mimicking, we decompose a feature vector into the magnitude and the direction. We argue that the teacher should give more freedom to the student feature's magnitude, and let the student pay more attention on mimicking the feature direction. To meet this requirement, we propose a loss term based on locality-sensitive hashing (LSH). With the help of this new loss, our method indeed mimics feature directions more accurately, relaxes constraints on feature magnitudes, and achieves state-of-the-art distillation accuracy. We provide theoretical analyses of how LSH facilitates feature direction mimicking, and further extend feature mimicking to multi-label recognition and object detection.

cs.CV↗