Author name cluster

Yin Liu

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.

7 papers

2 author rows

JBHI Journal 2026 Journal Article

A Graph Convolutional Network with Pretrained Features and Iterative Polar Coordinate Attention for Cross-Subject EEG-based Neuropsychiatric Diagnosis

Yin Liu
Jiaojiao Deng
Runyi Xu
Yi Zhou

Electroencephalography (EEG)-based diagnosis of neuropsychiatric disorders offers a non-invasive and cost-effective solution for early detection. However, robust cross-subject generalization remains a major challenge due to substantial inter-individual variability in EEG signals. To address this, we propose PreIPCA-GCN, a novel graph convolutional network that integrates pretrained temporal features and Iterative Polar Coordinate Attention (IPCA)-based brain connectivity modeling. Specifically, we utilize a modified version of LaBraM, a large-scale pretrained EEG model, to extract subject-invariant node representations. Functional brain connectivity is then characterized using Pearson correlation and cosine similarity in polar space, capturing both connectivity strength (radius) and phase synchronization (angle). To fuse these complementary cues, we introduce a dual-path IPCA mechanism, refining the adjacency matrix across iterations. PreIPCA-GCN is evaluated on six public EEG datasets covering five neuropsychiatric disorders (e. g. , attention-deficit/hyperactivity disorder, Alzheimer's disease, schizophrenia), consistently demonstrating strong cross-subject accuracies under both hold-out (86. 77%-95. 69%) and leave-one-subject-out cross-validation (88. 81%-97. 49%). Comprehensive comparative results show that PreIPCA-GCN outperforms several state-of-the-art methods. Ablation studies further confirm the effectiveness of both the pretrained node features and IPCA-based fused adjacency matrix in improving cross-subject generalization. These findings suggest PreIPCA-GCN as a robust and generalizable framework for cross-subject EEG-based neuropsychiatric diagnosis, offering strong potential for future clinical applications.

Details DOI

YNIMG Journal 2026 Journal Article

Multimodal MRI data fusion reveals distinct structural, functional and neurochemical correlates of depression in patients with Parkinson's disease

Xiaorong Hou
Jiajian Zhang
Junhong Duan
Yafei Song
Xuxiong Tang
Ziwei Gong
Ziwen Li
Zhineng Kang

Depression in Parkinson's disease (PD) involves complex structural, functional and multiple neurotransmitter systems alterations. So far, the precise interplay between structural and functional brain alterations and their underlying neurotransmitter processes remains largely unexplored. The advent of parallel independent component analysis (pICA) and the JuSpace toolbox provide a possible clue to elucidate their interrelationships and underlying mechanisms. In this study, we employed pICA to examine co-varying components interaction between gray matter volume (GMV) and fractional amplitude of low-frequency fluctuations (fALFF) in a cohort of 142 PD patients, comprising 53 PD patients with depression (PDD) and 89 PD patients without depression (PDND). Furthermore, we examined the spatial correlations between the GMV/fALFF components identified by pICA and neurotransmitter system maps using the JuSpace toolbox. Our analysis revealed significant negative correlations between one fMRI component (fALFF_IC6, frontoparietal, temporal and cerebellar regions) and two sMRI components (GMV_IC1 and GMV_IC4, basal ganglia, thalamocortical circuits, cerebellum and sensorimotor networks), which was significantly different between PDD and PDND group. Meanwhile, we found that alterations in both fALFF and GMV were widely associated with multiple neurotransmitter systems, primarily the dopaminergic and serotonergic systems. Notably, the severity of depression in PD was significantly correlated with these two distinct structural networks, independent of disease duration, motor symptoms and cognitive performance. These findings suggest the PD related depression-specific interrelationships between intrinsic network activity and GMV, potentially elucidating the multimodal neural circuitry and potential neurotransmitter patterns underlying depression in PD.

Details DOI

ICML Conference 2025 Conference Paper

A Memory Efficient Randomized Subspace Optimization Method for Training Large Language Models

Yiming Chen
Yuan Zhang
Yin Liu
Kun Yuan
Zaiwen Wen

The memory challenges associated with training Large Language Models (LLMs) have become a critical concern, particularly when using the Adam optimizer. To address this issue, numerous memory-efficient techniques have been proposed, with GaLore standing out as a notable example designed to reduce the memory footprint of optimizer states. However, these approaches do not alleviate the memory burden imposed by activations, rendering them unsuitable for scenarios involving long context sequences or large mini-batches. Moreover, their convergence properties are still not well-understood in the literature. In this work, we introduce a Randomized Subspace Optimization framework for pre-training and fine-tuning LLMs. Our approach decomposes the high-dimensional training problem into a series of lower-dimensional subproblems. At each iteration, a random subspace is selected, and the parameters within that subspace are optimized. This structured reduction in dimensionality allows our method to simultaneously reduce memory usage for both activations and optimizer states. We establish comprehensive convergence guarantees and derive rates for various scenarios, accommodating different optimization strategies to solve the subproblems. Extensive experiments validate the superior memory and communication efficiency of our method, achieving performance comparable to GaLore and Adam.

Details

EAAI Journal 2025 Journal Article

Local interpretation of deep learning models for Aspect-Based Sentiment Analysis

Stefan Lam
Yin Liu
Max Broers
Jasper van der Vos
Flavius Frasincar
David Boekestijn
Finn van der Knaap

Currently, deep learning models are commonly used for Aspect-Based Sentiment Analysis (ABSA). These deep learning models are often seen as black boxes, meaning that they are inherently difficult to interpret. To improve deep learning models, it is crucial to understand their inner workings. We aim to interpret black box models by implementing model-agnostic local interpretation methods. Inspired by Local Interpretable Model-agnostic Explanations (LIME) and Local Rule-based Explanations (LORE) and combined with a Similarity-based Sampling (SS) method, we propose SS-LIME and SS-LORE, and use Anchor to explain two state-of-the-art ABSA deep learning models. The deep learning models build upon the Left-Center-Right separated neural network with Rotatory attention (LCR-Rot) model, extended by iterating multiple times over the rotatory attention mechanism (LCR-Rot-hop) and hierarchical attention and context-dependent word embeddings (LCR-Rot-hop++). We evaluate the proposed models in terms of fidelity, hit rate, and user interpretability using the SemEval 2016 dataset consisting of restaurant reviews for ternary sentiment classification. Results show that the LCR-Rot-hop and LCR-Rot-hop++ models are best explained by SS-LIME and SS-LORE, respectively. Furthermore, we conclude that the LCR-Rot-hop++ model can be better interpreted than the LCR-Rot-hop model.

Details DOI

EAAI Journal 2025 Journal Article

Prediction and evaluation of key parameters in coalbed methane pre-extraction based on transformer and inversion model

Li Yan
Hu Wen
Zhenping Wang
Yongfei Jin
Jun Guo
Yin Liu
Shixing Fan

Accurate parameter prediction in the coalbed methane (CBM) pre-extraction process is crucial for formulating effective control measures and preventing CBM-related accidents. Traditional prediction methods rely on feature extraction or complex physical model parameter calculations, which require extensive manual intervention and have limited practical applicability. Additionally, simple neural network methods are prone to overfitting and gradient vanishing when handling parameters, and they lack the capability to dynamically monitor gas pressure during extraction, leading to inefficient and blind extraction operations. This study proposes a CBM pre-extraction parameter and completion time prediction method based on the Transformer model. By integrating autoregressive models and wavelet denoising techniques, the approach effectively captures temporal features and long-term dependencies in CBM data. Experimental results demonstrate that the proposed model outperforms traditional methods in short-, medium-, and long-term predictions, with a median R2 value of 0. 99072, and 76% of the training results exceeding 0. 9. Furthermore, a CBM pressure inversion model was developed, combining dimensional analysis and physical similarity principles with the Transformer model, enabling the dynamic detection of high- and low-pressure regions in coal seams. In single borehole compliance time predictions, the median compliance time for the first stage is 4 days, with an average of 49 days and a maximum of 277 days, providing adjustment guidance for boreholes with extended compliance times. The proposed model significantly improves prediction accuracy and stability, offering critical support for developing scientifically sustainable pre-extraction plans and advancing intelligent CBM management.

Details DOI

TCS Journal 2025 Journal Article

Restricted holant dichotomy on domain sizes 3 and 4

Yin Liu
Austen Z. Fan
Jin-Yi Cai

Holant ⁎ ( f ) denotes a class of counting problems specified by a constraint function f. We prove complexity dichotomy theorems for Holant ⁎ ( f ) in two settings: (1) f is any symmetric arity-3 real-valued function on input of domain size 3. (2) f is any symmetric arity-3 { 0, 1 } -valued function on input of domain size 4.

Details DOI

YNIMG Journal 2024 Journal Article

Quantitative susceptibility mapping through model-based deep image prior (MoDIP)

Zhuang Xiong
Yang Gao
Yin Liu
Amir Fazlollahi
Peter Nestor
Feng Liu
Hongfu Sun

The data-driven approach of supervised learning methods has limited applicability in solving dipole inversion in Quantitative Susceptibility Mapping (QSM) with varying scan parameters across different objects. To address this generalization issue in supervised QSM methods, we propose a novel training-free model-based unsupervised method called MoDIP (Model-based Deep Image Prior). MoDIP comprises a small, untrained network and a Data Fidelity Optimization (DFO) module. The network converges to an interim state, acting as an implicit prior for image regularization, while the optimization process enforces the physical model of QSM dipole inversion. Experimental results demonstrate MoDIP's excellent generalizability in solving QSM dipole inversion across different scan parameters. It exhibits robustness against pathological brain QSM, achieving over 32 % accuracy improvement than supervised deep learning methods. It is also 33 % more computationally efficient and runs 4 times faster than conventional DIP-based approaches, enabling 3D high-resolution image reconstruction in under 4.5 min.

Details DOI