Author name cluster

Yu Fu

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

13 papers

2 author rows

AAAI Conference 2026 Conference Paper

Harnessing the Unseen: The Hidden Influence of Intrinsic Knowledge in Long-Context Language Models

Yu Fu
Haz Sameen Shahgir
Hui Liu
Xianfeng Tang
Qi He
Yue Dong

Recent advances in long-context language models (LCLMs), designed to handle extremely long contexts, primarily focus on utilizing external contextual information, often leaving the influence of language models' parametric knowledge underexplored. In this work, we firstly investigate how this parametric knowledge affects content generation and demonstrate that its impact becomes increasingly pronounced as context length extends. Furthermore, we show that the model’s ability to utilize parametric knowledge, which we call parametric recall ability, does not improve simultaneously with its ability to leverage contextual knowledge through extrinsic retrieval ability. Moreover, better extrinsic retrieval ability can interfere with the model’s parametric recall ability, limiting its full potential. To bridge this gap, we design a simple yet effective Hybrid Needle-in-a-Haystack test that evaluates models based on their capabilities across both abilities, rather than solely emphasizing extrinsic retrieval ability. Our experimental results reveal that Qwen-2.5 models significantly outperform Llama-3.1 models, demonstrating superior potential to combine various abilities. Moreover, even the more powerful Llama-3.1-70B-Instruct model fails to exhibit better performance, highlighting the importance of evaluating models from a dual-ability perspective.

PDF Details DOI

ICLR Conference 2025 Conference Paper

XAIguiFormer: explainable artificial intelligence guided transformer for brain disorder identification

Hanning Guo
Farah Abdellatif
Yu Fu
N. Jon Shah
Abigail Morrison
Jürgen Dammers

EEG-based connectomes offer a low-cost and portable method to identify brain disorders using deep learning. With the growing interest in model interpretability and transparency, explainable artificial intelligence (XAI) is widely applied to understand the decision of deep learning models. However, most research focuses solely on interpretability analysis based on the insights from XAI, overlooking XAI’s potential to improve model performance. To bridge this gap, we propose a dynamical-system-inspired architecture, XAI guided transformer (XAIguiFormer), where XAI not only provides explanations but also contributes to enhancing the transformer by refining the originally coarse information in self-attention mechanism to capture more relevant dependency relationships. In order not to damage the connectome’s topological structure, the connectome tokenizer treats the single-band graphs as atomic tokens to generate a sequence in the frequency domain. To address the limitations of conventional positional encoding in understanding the frequency and mitigating the individual differences, we integrate frequency and demographic information into tokens via a rotation matrix, resulting in a richly informative representation. Our experiment demonstrates that XAIguiFormer achieves superior performance over all baseline models. In addition, XAIguiFormer provides valuable interpretability through visualization of the frequency band importance. Our code is available at https://github.com/HanningGuo/XAIguiFormer.

Details

NeurIPS Conference 2024 Conference Paper

Free Lunch in Pathology Foundation Model: Task-specific Model Adaptation with Concept-Guided Feature Enhancement

Yanyan Huang
Weiqin Zhao
Yihang Chen
Yu Fu
Lequan Yu

Whole slide image (WSI) analysis is gaining prominence within the medical imaging field. Recent advances in pathology foundation models have shown the potential to extract powerful feature representations from WSIs for downstream tasks. However, these foundation models are usually designed for general-purpose pathology image analysis and may not be optimal for specific downstream tasks or cancer types. In this work, we present Concept Anchor-guided Task-specific Feature Enhancement (CATE), an adaptable paradigm that can boost the expressivity and discriminativeness of pathology foundation models for specific downstream tasks. Based on a set of task-specific concepts derived from the pathology vision-language model with expert-designed prompts, we introduce two interconnected modules to dynamically calibrate the generic image features extracted by foundation models for certain tasks or cancer types. Specifically, we design a Concept-guided Information Bottleneck module to enhance task-relevant characteristics by maximizing the mutual information between image features and concept anchors while suppressing superfluous information. Moreover, a Concept-Feature Interference module is proposed to utilize the similarity between calibrated features and concept anchors to further generate discriminative task-specific features. The extensive experiments on public WSI datasets demonstrate that CATE significantly enhances the performance and generalizability of MIL models. Additionally, heatmap and umap visualization results also reveal the effectiveness and interpretability of CATE.

PDF Details DOI

ICRA Conference 2024 Conference Paper

Phasic Diversity Optimization for Population-Based Reinforcement Learning

Jingcheng Jiang
Haiyin Piao
Yu Fu
Yihang Hao
Chuanlu Jiang
Ziqi Wei 0001
Xin Yang 0011

Reviewing the previous work of diversity Reinforcement Learning, diversity is often obtained via an augmented loss function, which requires a balance between reward and diversity. Generally, diversity optimization algorithms use Multi-armed Bandits algorithms to select the coefficient in the pre-defined space. However, the dynamic distribution of reward signals for MABs or the conflict between quality and diversity limits the performance of these methods. We introduce the Phasic Diversity Optimization (PDO) algorithm, a Population-Based Training framework that separates reward and diversity training into distinct phases instead of optimizing a multi-objective function. In the auxiliary phase, agents with poor performance diversified via determinants will not replace the better agents in the archive. The decoupling of reward and diversity allows us to use an aggressive diversity optimization in the auxiliary phase without performance degradation. Furthermore, we construct a dogfight scenario for aerial agents to demonstrate the practicality of the PDO algorithm. We introduce two implementations of PDO archive and conduct tests in the newly proposed adversarial dogfight and MuJoCo simulations. The results show that our proposed algorithm achieves better performance than baselines.

Details

AAAI Conference 2024 Conference Paper

Watermarking Conditional Text Generation for AI Detection: Unveiling Challenges and a Semantic-Aware Watermark Remedy

Yu Fu
Deyi Xiong
Yue Dong

To mitigate potential risks associated with language models (LMs), recent AI detection research proposes incorporating watermarks into machine-generated text through random vocabulary restrictions and utilizing this information for detection. In this paper, we show that watermarking algorithms designed for LMs cannot be seamlessly applied to conditional text generation (CTG) tasks without a notable decline in downstream task performance. To address this issue, we introduce a simple yet effective semantic-aware watermarking algorithm that considers the characteristics of conditional text generation with the input context. Compared to the baseline watermarks, our proposed watermark yields significant improvements in both automatic and human evaluations across various text generation models, including BART and Flan-T5, for CTG tasks such as summarization and data-to-text generation. Meanwhile, it maintains detection ability with higher z-scores but lower AUC scores, suggesting the presence of a detection paradox that poses additional challenges for watermarking CTG.

PDF Details DOI

YNIMG Journal 2023 Journal Article

Functional activity of the caudate mediates the relation between early childhood microstructural variations and elevated metabolic syndrome scores

Pei Huang
Mya Thway Tint
Marissa Lee
Zhen Ming Ngoh
Peter Gluckman
Yap Seng Chong
Weiping Han
Yu Fu

BACKGROUND: Metabolic syndrome score in children assesses the risk of developing cardiovascular disease in future. We aim to probe the role of the caudate in relation to the metabolic syndrome score. Furthermore, using both functional and structural neuroimaging, we aim to examine the interplay between functional and structural measures. METHODS: A longitudinal birth cohort study with functional and structural neuroimaging data obtained at 4.5, 6.0 and 7.5 years and metabolic syndrome scores at 8.0 years was used. Pearson correlation and linear regression was used to test for correlation fractional anisotropy (FA) and fractional amplitude of low frequency fluctuations (fALFF) of the caudate with metabolic syndrome scores. Mediation analysis was used to test if later brain measures mediated the relation between earlier brain measures and metabolic syndrome scores. Inhibitory control was also tested as a mediator of the relation between caudate brain measures and metabolic syndrome scores. RESULTS: FA at 4.5 years and fALFF at 7.5 years of the left caudate was significantly correlated with metabolic syndrome scores. Post-hoc mediation analysis showed that fALFF at 7.5 years fully mediated the relation between FA at 4.5 years and metabolic syndrome scores. Inhibitory control was significantly correlated with fALFF at 7.5 years, but did not mediate the relation between fALFF at 7.5 years and metabolic syndrome scores. CONCLUSIONS: We found that variations in caudate microstructure at 4.5 years predict later variation in functional activity at 7.5 years. This later variation in functional activity fully mediates the relation between microstructural changes in early childhood and metabolic syndrome scores at 8.0 years.

Details DOI

JBHI Journal 2023 Journal Article

PKA 2 -Net: Prior Knowledge-Based Active Attention Network for Accurate Pneumonia Diagnosis on Chest X-Ray Images

Yu Fu
Peng Xue
Zhili Zhang
Enqing Dong

To accurately diagnose pneumonia patients on a limited annotated chest X-ray image dataset, a prior knowledge-based active attention network (PKA 2 -Net 1 ) was constructed. The PKA 2 -Net uses improved ResNet as the backbone network and consists of residual blocks, novel subject enhancement and background suppression (SEBS) blocks and candidate template generators, where template generators are designed to generate candidate templates for characterizing the importance of different spatial locations in feature maps. The core of PKA 2 -Net is SEBS block, which is proposed based on the prior knowledge that highlighting distinctive features and suppressing irrelevant features can improve the recognition effect. The purpose of SEBS block is to generate active attention features without any high-level features and enhance the ability of the model to localize lung lesions. In SEBS block, first, a series of candidate templates T with different spatial energy distributions are generated and the controllability of the energy distribution in T enables active attention features to maintain the continuity and integrity of the feature space distributions. Second, Top- n templates are selected from T according to certain learning rules, which are then operated by a convolution layer for generating supervision information that can guide the inputs of SEBS block to form active attention features. We evaluated the PKA 2 -Net on the binary classification problem of identifying pneumonia and healthy controls on a dataset containing 5856 chest X-ray images (ChestXRay2017), the results showed that our method can achieve 97. 63% accuracy and 0. 9872 sensitivity.

Details DOI

JBHI Journal 2023 Journal Article

Semi-Supervised Adversarial Learning for Improving the Diagnosis of Pulmonary Nodules

Yu Fu
Peng Xue
Taohui Xiao
Zhili Zhang
Youren Zhang
Enqing Dong

Achieving the pathological type diagnosis of pulmonary nodules on chest CT is a critical step in the early detection of lung cancer and treatment of patients. Based on a small and unbalanced self-constructed dataset, we achieved intelligent diagnosis of five pathological types including adenocarcinoma, squamous cell carcinoma, small cell carcinoma, inflammatory and other benign diseases for the first time. In order to reduce the dependence of deep convolutional neural network (DCNN) on a large amount of training data, a reverse adversarial classification network (RACN) was proposed based on semi-supervised learning, which consists of a reverse generative adversarial network (RGAN) for unsupervised regression and a supervised classification network (CN). In RGAN, five specific normal distributions P with different means and variances were assigned to represent the five pathological types, and then a special regression task was designed by mapping pulmonary nodules to the random sampling Z of P. The input of generator in RGAN is set to 3D nodule volume data, the inputs of discriminator are set to Z and the output of generator. The regression task enables RGAN to extract specific features, which will be deeply integrate into CN to improve the classification performance. Experiments showed that the average sensitivity of RACN in detecting malignant nodules was 0. 6525, where the sensitivity of adenocarcinoma, small cell carcinoma and squamous cell carcinoma was 0. 8426, 0. 5604 and 0. 5543. Besides, the RACN can achieve 93. 21% accuracy for diagnosing malignant nodules on the public LIDC-IDRI dataset, obtaining the state-of-the-art results.

Details DOI

JBHI Journal 2022 Journal Article

Harmony Loss for Unbalanced Prediction

Yu Fu
Peng Xue
Meirong Ren
Enqing Dong

In medical image analysis, in order to reduce the impact of unbalanced data sets on data-driven deep learning models, according to the characteristic that the area under the Precision-Recall curve (AUC PR ) is sensitive to each category of samples, a novel Harmony loss function with fast convergence speed and high stability was constructed. Since AUC PR needs to be calculated in discrete domain, in order to ensure the continuous differentiability and gradient existence of the Harmony loss, first, the Logistic function was used to approximate the Logical function in AUC PR. Then, to improve the optimization speed of the Harmony loss during model training, a method of manually setting a certain number of classification thresholds was proposed to further approximate the calculation of AUC PR. After the above two approximate calculation processes, the Harmony loss with stable gradient and high computational efficiency was designed. In the optimization process of the model, since Harmony loss can reconcile recall and precision of each category under different classification thresholds, thereby, it can not only improve the model's ability to recognize categories with less samples, but also maintain the stability of the training curve. To comprehensively evaluate the effects of Harmony loss function, we performed experiments on image 3D reconstruction, 2D segmentation, and unbalanced classification tasks. Experimental results showed that the Harmony loss achieved the state-of-the-art results on four unbalanced data sets. Moreover, the Harmony loss can be easily combined with existing loss functions, and is suitable for most common deep learning models.

Details DOI

YNIMG Journal 2022 Journal Article

Uncovering the global task-modulated brain network in chunk decomposition with Chinese characters

Jiaoyan Pang
Hanning Guo
Xiaochen Tang
Yu Fu
Zhengwu Yang
Yongchao Li
Na An
Jing Luo

Chunk decomposition, which requires the mental representation transformation in accordance with behavioral goals, is of vital importance to problem solving and creative thinking. Previous studies have identified that the frontal, parietal, and occipital cortex in the cognitive control network selectively activated in response to chunk tightness, however, functional localization strategy may overlook the interaction brain regions. Based on the notion of a global brain network, we proposed that multiple specialized regions have to be interconnected to maintain goal representation during the course of chunk decomposition. Therefore, the present study applied a beta-series correlation method to investigate interregional functional connectivity in the event-related design of chunk decomposition tasks using Chinese characters, which would highlight critical nodes irrespective to chunk tightness. The results reveal a network of functional hubs with highly within or between module connections, including the orbitofrontal cortex, superior/inferior parietal lobule, hippocampus, and thalamus. We speculate that the thalamus integrates information across modular as an integrative hub while the orbitofrontal cortex tracks the mental states of chunk decomposition on a moment-to-moment basis. The superior and inferior parietal lobule collaborate to manipulate the mental representation of chunk decomposition and the hippocampus associates the relationship between elements in the question and solution phase. Furthermore, the tightness of chunks is not only associated with different processors in visual systems but also leads to increased intermodular connections in right superior frontal gyrus and left precentral gyrus. To summary up, the present study first reveals the task-modulated brain network of chunk decomposition in addition to the tightness-related nodes in the frontal and occipital cortex.

Details DOI

JBHI Journal 2021 Journal Article

Lung Respiratory Motion Estimation Based on Fast Kalman Filtering and 4D CT Image Registration

Peng Xue
Yu Fu
Huizhong Ji
Wentao Cui
Enqing Dong

Respiratory motion estimation is an important part in image-guided radiation therapy and clinical diagnosis. However, most of the respiratory motion estimation methods rely on indirect measurements of external breathing indicators, which will not only introduce great estimation errors, but also bring invasive injury for patients. In this paper, we propose a method of lung respiratory motion estimation based on fast Kalman filtering and 4D CT image registration (LRME-4DCT). In order to perform dynamic motion estimation for continuous phases, a motion estimation model is constructed by combining two kinds of GPU-accelerated 4D CT image registration methods with fast Kalman filtering method. To address the high computational requirements of 4D CT image sequences, a multi-level processing strategy is adopted in the 4D CT image registration methods, and respiratory motion states are predicted from three independent directions. In the DIR-lab dataset and POPI dataset with 4D CT images, the average target registration error (TRE) of the LRME-4DCT method can reach 0. 91 mm and 0. 85 mm respectively. Compared with traditional estimation methods based on pair-wise image registration, the proposed LRME-4DCT method can estimate the physiological respiratory motion more accurately and quickly. Our proposed LRME-4DCT method fully meets the practical clinical requirements for rapid dynamic estimation of lung respiratory motion.

Details DOI

JBHI Journal 2014 Journal Article

A Novel MRI Tracer-Based Method for Measuring Water Diffusion in the Extracellular Space of the Rat Brain (December2013)

Hongbin Han
Chunyan Shi
Yu Fu
Long Zuo
Kejia Lee
Qingyuan He
Haojun Han

We proposed a novel MRI tracer-based method for the determination of water diffusion in the brain extracellular space (ECS). The measuring system was validated in 32 Sprague Dawley rats. The rats were randomly divided into four groups with different injection sites: 1) caudate nucleus (Cn.); 2) thalamus (T.); 3) cortex (Cor.); and 4) substantia nigra (Sn.). The spin-lattice relaxation time of hydrogen nuclei in water molecules were shortened, which presented as high signal on MRI after the injection of gadolinium-diethylene triamine pentaacetic acid (Gd-DTPA) into the rat brain ECS. The enhancement on MRI decreased over time due to the water diffusion and clearance process within the brain ECS. The process was dynamically recorded on a series of magnetic resonance (MR) images. As the increment in signal intensity (ΔSI) could be converted to local Gd-DTPA concentration, the water diffusion parameters were further calculated voxel by voxel based on a modified diffusion model. The most tortuous ECS (λ = 1. 77 ± 0. 71) was found in Sn. with D*Sn of (2. 06 ± 1. 01) × 10 -4 mm 2 ·s -1 ( P -4 mm 2 ·s -1 ( F = 0. 18, P > 0. 05). By using the tracer-based MRI method, the local diffusion parameters of the brain ECS can be quantitatively measured. The different distribution territories and clearance rates of the tracer in four brain areas indicated that the brain ECS is a physiologically partitioned system.

Details DOI

IROS Conference 2011 Conference Paper

Monte Carlo Localization using 3D texture maps

Yu Fu
Stephen Tully
George Kantor
Howie Choset

This paper uses KLD-based (Kullback-Leibler Divergence) Monte Carlo Localization (MCL) to localize a mobile robot in an indoor environment represented by 3D texture maps. A 3D texture map is a simplified model that includes vertical planes with colored texture information associated with each vertical plane. At each time step, a distance measurement and an observed texture from an omnidirectional camera are compared to the expected distance measurement and the expected texture according to each hypothesis of the robot's pose in an MCL framework. Compared to previous implementations of MCL, our proposed approach converges faster than distance-only MCL and localizes the robot more precisely than SIFT-based MCL. We demonstrate this new MCL algorithm for robot localization with experiments in several hallways.

Details