Author name cluster

Liang Ma

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.

11 papers

2 author rows

JBHI Journal 2026 Journal Article

scProca: A Cross-Attention-Enhanced Deep Generative Model for Single-Cell Transcriptomics and Proteomics Integration and Imputation

Jiankang Xiong
Shuqiao Zheng
Fuzhou Gong
Liang Ma
Lin Wan

Understanding the molecular mechanisms of complex diseases requires insight into cellular interactions and protein expression. While large-scale sequencing enables disease subtyping and patient stratification, integrating proteomics and transcriptomics data offers a deeper view of cellular states. Recent methods combine scRNA-seq, which provides broad cellular coverage, with transcriptomics and proteomics co-profiling, which provides more comprehensive molecular measurements. However, many models adopt simplistic strategies for joint analysis. We introduce scProca, a deep generative model that incorporates inter-cellular relationships via cross-attention mechanisms to handle heterogeneous inputs, whether from RNA-seq or co-profiling datasets. scProca achieves state-of-the-art integration and imputation, remains robust under high protein sparsity, generalizes across species and tissues, scales to large datasets, and is compatible with experimental batches, demonstrating strong flexibility for complex experimental settings.

Details DOI

EAAI Journal 2025 Journal Article

A novel multi-source heterogeneous data fusion based fault diagnosis framework for manufacturing processes

Liang Ma
Qikai Yang
Orestes Llanes-Santiago
Kaixiang Peng

Fault diagnosis technologies are important means to ensure the production operation safety and product quality stability for manufacturing processes. After a fault occurs in the manufacturing processes, it may be characterized by high-frequency and high-dimensional structured time series data anomalies such as sensor data, or by unstructured data anomalies such as images. Traditionally, single structured sensor data is often used for constructing diagnosis models, the fault characteristics may not be adequately characterized, thus affecting the diagnosis performance. Therefore, in this paper, in order to make full use of multi-source data and obtain more comprehensive and accurate diagnosis results, a new multi-source heterogeneous data fusion based fault diagnosis framework is designed for manufacturing processes. Specifically, to solve the problem that the important information is ignored during data level fusion of multi-source data, an adaptive weight multi-source data fusion method is proposed. Furthermore, in response to the problem of feature redundancy in feature level fusion of heterogeneous data, a feature differentiation extraction and heterogeneous feature fusion method is proposed, of which a feature source discriminator is constructed for enhancing the complementarity of the extracted heterogeneous features, and feature concatenation is performed to improve the feature expression ability. Finally, the effectiveness and feasibility of the proposed framework is verified on actual datasets from the hot rolling process and the Tennessee Eastman process. Experimental results show that the proposed framework is both effective and feasible in fault diagnosis with multi-source heterogeneous data.

Details DOI

YNIMG Journal 2025 Journal Article

Iterative prior-guided parcellation (iPGP) for capturing inter-subject and inter-nuclei variability in thalamic mapping

Chaohong Gao
Xia Wu
Liang Ma
Deying Li
Yufan Wang
Changlu Guo
Wen Li
Haiyan Wang

The thalamus, a critical relay station in the brain, consists of multiple nuclei that play essential roles in various brain circuits. Identifying these nuclei is crucial for understanding how thalamic structures influence cognitive functions. However, genetic and environmental factors introduce substantial variability in thalamic parcellation patterns, posing both challenges and opportunities for individualized mapping of thalamic function. This study proposes an iterative prior-guided parcellation (iPGP) framework to construct individualized thalamic parcellations. The iPGP method utilizes the Morel histological atlas as prior guidance, incorporates spatially constrained local diffusion characteristics as features, and employs an iterative framework to optimize an individual-specific parcellation model. As a result, iPGP automatically adapts to individual thalamic contrast variations, producing personalized and anatomically consistent parcellations. Through test-retest assessments, iPGP demonstrated a high degree of intra-subject reproducibility. By evaluating inter-subject and inter-nuclei variability, iPGP exhibited strong adaptability across different age groups while capturing subject-specific and region-specific variability. Furthermore, thalamic parcellations generated by iPGP showed significant associations with adolescent age and adult behavioral-cognitive scores. Our findings suggest that iPGP effectively captures inter-subject and inter-nuclei variability in thalamic parcellation, highlighting its potential for advancing thalamic mapping in exploring brain function.

Details DOI

NeurIPS Conference 2025 Conference Paper

MineAnyBuild: Benchmarking Spatial Planning for Open-world AI Agents

Ziming Wei
Bingqian Lin
Zijian Jiao
Yunshuang Nie
Liang Ma
Yuecheng Liu
Yuzheng Zhuang
Xiaodan Liang

Spatial Planning is a crucial part in the field of spatial intelligence, which requires the understanding and planning about object arrangements in space perspective. AI agents with the spatial planning ability can better adapt to various real-world applications, including robotic manipulation, automatic assembly, urban planning etc. Recent works have attempted to construct benchmarks for evaluating the spatial intelligence of Multimodal Large Language Models (MLLMs). Nevertheless, these benchmarks primarily focus on spatial reasoning based on typical Visual Question-Answering (VQA) forms, which suffers from the gap between abstract spatial understanding and concrete task execution. In this work, we take a step further to build a comprehensive benchmark called MineAnyBuild, aiming to evaluate the spatial planning ability of open-world AI agents in the Minecraft game. Specifically, MineAnyBuild requires an agent to generate executable architecture building plans based on the given multi-modal human instructions. It involves 4, 000 curated spatial planning tasks and also provides a paradigm for infinitely expandable data collection by utilizing rich player-generated content. MineAnyBuild evaluates spatial planning through four core supporting dimensions: spatial understanding, spatial reasoning, creativity, and spatial commonsense. Based on MineAnyBuild, we perform a comprehensive evaluation for existing MLLM-based agents, revealing the severe limitations but enormous potential in their spatial planning abilities. We believe our MineAnyBuild will open new avenues for the evaluation of spatial intelligence and help promote further development for open-world AI agents capable of spatial planning.

PDF Details

NeurIPS Conference 2025 Conference Paper

PhyBlock: A Progressive Benchmark for Physical Understanding and Planning via 3D Block Assembly

Liang Ma
Jiajun Wen
Min Lin
Rongtao Xu
Xiwen Liang
Bingqian Lin
Jun Ma
Yongxin Wang

While vision-language models (VLMs) have demonstrated promising capabilities in reasoning and planning for embodied agents, their ability to comprehend physical phenomena, particularly within structured 3D environments, remains severely limited. To close this gap, we introduce PhyBlock, a progressive benchmark designed to assess VLMs on physical understanding and planning through robotic 3D block assembly tasks. PhyBlock integrates a novel four-level cognitive hierarchy assembly task alongside targeted Visual Question Answering (VQA) samples, collectively aimed at evaluating progressive spatial reasoning and fundamental physical comprehension, including object properties, spatial relationships, and holistic scene understanding. PhyBlock includes 2600 block tasks (400 assembly tasks, 2200 VQA tasks) and evaluates models across three key dimensions: partial completion, failure diagnosis, and planning robustness. We benchmark 23 state-of-the-art VLMs, highlighting their strengths and limitations in physically grounded, multi-step planning. Our empirical findings indicate that the performance of VLMs exhibits pronounced limitations in high-level planning and reasoning capabilities, leading to a notable decline in performance for the growing complexity of the tasks. Error analysis reveals persistent difficulties in spatial orientation and dependency reasoning. We position PhyBlock as a unified testbed to advance embodied reasoning, bridging vision-language understanding and real-world physical problem-solving.

PDF Details

ICML Conference 2025 Conference Paper

Unbiased Evaluation of Large Language Models from a Causal Perspective

Meilin Chen
Jian Tian
Liang Ma
Di Xie
Weijie Chen 0006
Jiang Zhu

Benchmark contamination has become a significant concern in the LLM evaluation community. Previous Agents-as-an-Evaluator address this issue by involving agents in the generation of questions. Despite their success, the biases in Agents-as-an-Evaluator methods remain largely unexplored. In this paper, we present a theoretical formulation of evaluation bias, providing valuable insights into designing unbiased evaluation protocols. Furthermore, we identify two type of bias in Agents-as-an-Evaluator through carefully designed probing tasks on a minimal Agents-as-an-Evaluator setup. To address these issues, we propose the Unbiased Evaluator, an evaluation protocol that delivers a more comprehensive, unbiased, and interpretable assessment of LLMs. Extensive experiments reveal significant room for improvement in current LLMs. Additionally, we demonstrate that the Unbiased Evaluator not only offers strong evidence of benchmark contamination but also provides interpretable evaluation results.

Details

EAAI Journal 2023 Journal Article

Green location routing problem with flexible multi-compartment for source-separated waste: A Q-learning and multi-strategy-based hyper-heuristic algorithm

Chunjian Shang
Liang Ma
Yong Liu

In this paper, we extend a novel model for source-separated waste collection and transportation, the green location routing problem with multi-compartment (GLRPFMC), for which we design a Q-learning and multi-strategy-based hyper-heuristic algorithm (QLMSHH). The remarkable merits of this paper can be highlighted as the following threefold: (1) The GLRPFMC is novel in that it constructs a variant of the location routing problem with carbon emissions and flexible multi-compartment sizes that occurs in a source-separated waste transportation context. (2) For the methodological contribution, the QLMSHH is presented to design a hyper-heuristic model by intelligently selecting appropriate high-level heuristic components during different stages of the optimization process. (3) The proposed method incorporates the design of solution representations, evolution-acceptance pairs for high-level heuristic construction, the repairing solution scheme, and the local search strategy. Finally, sufficient experiments are conducted on the benchmark, new instances, and simulation data of GLRPFMC and draw some managerial insights. The satisfactory results highlight the efficiency and universality of the proposed model and method.

Details DOI

YNIMG Journal 2022 Journal Article

A model-based approach to assess reproducibility for large-scale high-throughput MRI-based studies

Zeyu Jiao
Yinglei Lai
Jujiao Kang
Weikang Gong
Liang Ma
Tianye Jia
Chao Xie
Shitong Xiang

Magnetic Resonance Imaging (MRI) technology has been increasingly used in neuroscience studies. Reproducibility of statistically significant findings generated by MRI-based studies, especially association studies (phenotype vs. MRI metric) and task-induced brain activation, has been recently heavily debated. However, most currently available reproducibility measures depend on thresholds for the test statistics and cannot be use to evaluate overall study reproducibility. It is also crucial to elucidate the relationship between overall study reproducibility and sample size in an experimental design. In this study, we proposed a model-based reproducibility index to quantify reproducibility which could be used in large-scale high-throughput MRI-based studies including both association studies and task-induced brain activation. We performed the model-based reproducibility assessments for a few association studies and task-induced brain activation by using several recent large sMRI/fMRI databases. For large sample size association studies between brain structure/function features and some basic physiological phenotypes (i.e. Sex, BMI), we demonstrated that the model-based reproducibility of these studies is more than 0.99. For MID task activation, similar results could be observed. Furthermore, we proposed a model-based analytical tool to evaluate minimal sample size for the purpose of achieving a desirable model-based reproducibility. Additionally, we evaluated the model-based reproducibility of gray matter volume (GMV) changes for UK Biobank (UKB) vs. Parkinson Progression Marker Initiative (PPMI) and UK Biobank (UKB) vs. Human Connectome Project (HCP). We demonstrated that both sample size and study-specific experimental factors play important roles in the model-based reproducibility assessments for different experiments. In summary, a systematic assessment of reproducibility is fundamental and important in the current large-scale high-throughput MRI-based studies.

Details DOI

YNICL Journal 2021 Journal Article

Neurological effects of hemodialysis on white matter microstructure in end-stage renal disease

Junya Mu
Liang Ma
Shaohui Ma
Dun Ding
Peng Li
Xueying Ma
Ming Zhang
Jixin Liu

OBJECTIVES: To detect the effects of hemodialysis (HD) on the central nervous system (CNS), the present study forces the memory storage capacity and the difference in white matter (WM) microstructure characteristics among end-stage renal disease (ESRD) participants before HD initiation (ESRD-BHD), ESRD participants with maintenance HD (ESRD-MHD), and healthy participants (HCs). METHODS: Between 2016 and 2018, 56 ESRD-BHD, 39 ESRD-MHD, and 56 HCs were recruited for this study. The fractional anisotropy (FA) of tractography streamlines within the working memory network was investigated using a novel along-tracts analysis method. The relationship between WM microstructure and working memory scores, measured from an n-back task, were detected by multiple correlation analysis. RESULTS: As compared with HCs, a significantly lower FA was found along part of the WM in the working memory network in ESRD-BHD. In the group-difference location of ESRD-BHD and HCs, the FA of ESRD-MHD was reversed to normal levels in HCs. However, the FA in a new location was differentially reduced across groups: highest in HCs, intermediate in ESRD-BHD, and lowest in ESRD-MHD. Correlation analysis showed that a longer reaction time correlated to a lower FA, according to the following pattern: ESRD-BHD > ESRD-MHD > HCs. CONCLUSION: Despite the persisting abnormal brain structure, our findings suggest HD has a neuroprotective effect in ESRD patients.

Details DOI

AAAI Conference 2019 Conference Paper

Learning Incremental Triplet Margin for Person Re-Identification

Yingying Zhang
Qiaoyong Zhong
Liang Ma
Di Xie
Shiliang Pu

Person re-identification (ReID) aims to match people across multiple non-overlapping video cameras deployed at different locations. To address this challenging problem, many metric learning approaches have been proposed, among which triplet loss is one of the state-of-the-arts. In this work, we explore the margin between positive and negative pairs of triplets and prove that large margin is beneficial. In particular, we propose a novel multi-stage training strategy which learns incremental triplet margin and improves triplet loss effectively. Multiple levels of feature maps are exploited to make the learned features more discriminative. Besides, we introduce global hard identity searching method to sample hard identities when generating a training batch. Extensive experiments on Market-1501, CUHK03, and DukeMTMCreID show that our approach yields a performance boost and outperforms most existing state-of-the-art methods.

PDF Details

YNIMG Journal 2017 Journal Article

Functional connectivity decreases in autism in emotion, self, and face circuits identified by Knowledge-based Enrichment Analysis

Wei Cheng
Edmund T. Rolls
Jie Zhang
Wenbo Sheng
Liang Ma
Lin Wan
Qiang Luo
Jianfeng Feng

A powerful new method is described called Knowledge based functional connectivity Enrichment Analysis (KEA) for interpreting resting state functional connectivity, using circuits that are functionally identified using search terms with the Neurosynth database. The method derives its power by focusing on neural circuits, sets of brain regions that share a common biological function, instead of trying to interpret single functional connectivity links. This provides a novel way of investigating how task- or function-related networks have resting state functional connectivity differences in different psychiatric states, provides a new way to bridge the gap between task and resting-state functional networks, and potentially helps to identify brain networks that might be treated. The method was applied to interpreting functional connectivity differences in autism. Functional connectivity decreases at the network circuit level in 394 patients with autism compared with 473 controls were found in networks involving the orbitofrontal cortex, anterior cingulate cortex, middle temporal gyrus cortex, and the precuneus, in networks that are implicated in the sense of self, face processing, and theory of mind. The decreases were correlated with symptom severity.

Details DOI