Author name cluster

Yu Xia

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.

8 papers

1 author row

EAAI Journal 2026 Journal Article

FEAN: A Fragments Embedding and Aligning Network for image-text matching

Xianlun Tang
Lin Jiang
Yu Xia
Wuquan Deng
Bo Tang
Tianyu Xiang
Hao Zhu

Most recent image-text matching methods use fragments alignment structures to achieve cross modal interaction. Although they compensate for the lack of cross modal interaction caused by overall embedding, the accuracy of image-text matching is compromised by the cross modal semantic gap and redundant alignment. To address these issues, we propose a Fragments Embedding and Aligning Network (FEAN) to achieve accurate image-text matching, which focuses on the representation of global and local image-text features and their similarity. Specifically, we embed local image and textual features into fragments to obtain global image and textual features representations and global similarity scores, and use cross modal weight calculation to obtain fragments aligned local image and textual features and local similarity matrices, in order to mitigate the cross modal semantic gap. To avoid redundant alignment in image and textual fragments, a Similarity Pooling (SP) strategy is proposed to aggregate global similarity scores and local similarity matrices. In addition, to compensate for the missing contextual semantic information on the image region features, we add the Position Weight Feature Reinforcement (PWFR) module before embedding alignment to achieve consistency with the text semantic information. Extensive experiments on two publicly available datasets for image-text matching, Flickr30K and MSCOCO, have shown that our fragments embedding and aligning network achieve the best R@1 and RSUM values in bidirectional matching with optimal accuracy.

Details DOI

NeurIPS Conference 2025 Conference Paper

Multimodal Tabular Reasoning with Privileged Structured Information

Jun-Peng Jiang
Yu Xia
Hai-Long Sun
Shiyin Lu
Qingguo Chen
Weihua Luo
Kaifu Zhang
De-Chuan Zhan

Tabular reasoning requires complex, multi-step information extraction and logical inference, such as aggregation, comparison, or calculation over tabular data. While recent advances have leveraged large language models (LLMs) for reasoning over structured text tables, such high-quality textual representations are often unavailable in real-world settings, where tables typically appear as images. In this paper, we tackle the task of tabular reasoning directly from table images. Our core strategy is to leverage privileged structured information---specifically, the ground-truth structured table data available during training but inaccessible at test time---to enhance multimodal large language models (MLLMs). The key challenges lie in: accurately aligning visual representations with the structured information, particularly mapping the visual evidence to logical steps; and effectively transferring the reasoning skills learned during training to the MLLM for visual inference. To address these, we introduce {\sc Turbo} (TabUlar Reasoning with Bridged infOrmation), a new framework for multimodal tabular reasoning using privileged information. {\sc Turbo} benefits from a structure-aware reasoning trace generator based on DeepSeek-R1, which contributes to high-quality modality-bridged information. On this basis, {\sc Turbo} repeatedly generates and selects advantageous reasoning traces, further enhancing the model's tabular reasoning ability. Experimental results demonstrate that, with limited (9k) data, {\sc Turbo} achieves state-of-the-art performance ($+7. 2\%$ vs. previous SOTA) across multiple datasets.

PDF Details

EAAI Journal 2024 Journal Article

Multi-colour sketch-based image retrieval with an explicable feature embedding

Shuangbu Wang
Yu Xia
Nan Xiang
Kun Qian
Xiaosong Yang
Lihua You
Jianjun Zhang

Sketch-based image retrieval (SBIR) is a cross-domain matching problem that has been gaining continuous attention in the computer vision community. Currently, SBIR techniques mainly focus on dealing with black-and-white sketches and ignore the utilization of multi-colour information of images. This leads to insufficient retrieval performance since they cannot distinguish between images that have the same shape but different colours. To address this problem, a multi-colour sketch-based image retrieval (MCSBIR) method using a two-stage network architecture is proposed. A novel feature embedding for explicably describing the shape and colour information is designed. A triplet loss function is developed to learn the feature embedding, in which a new distance metric is proposed to separate the shape and colour features. In addition, the first multi-colour sketch-image dataset is built to achieve the MCSBIR task and a user interface is designed to visually present the MCSBIR method. The effectiveness of the MCSBIR method is demonstrated by comprehensive experiments.

Details DOI

YNICL Journal 2024 Journal Article

Predicting depression risk in early adolescence via multimodal brain imaging

Zeus Gracia-Tabuenca
Elise B. Barbeau
Yu Xia
Xiaoqian Chai

Depression is an incapacitating psychiatric disorder with increased risk through adolescence. Among other factors, children with family history of depression have significantly higher risk of developing depression. Early identification of pre-adolescent children who are at risk of depression is crucial for early intervention and prevention. In this study, we used a large longitudinal sample from the Adolescent Brain Cognitive Development (ABCD) Study (2658 participants after imaging quality control, between 9-10 years at baseline), we applied advanced machine learning methods to predict depression risk at the two-year follow-up from the baseline assessment, using a set of comprehensive multimodal neuroimaging features derived from structural MRI, diffusion tensor imaging, and task and rest functional MRI. Prediction performance underwent a rigorous cross-validation method of leave-one-site-out. Our results demonstrate that all brain features had prediction scores significantly better than expected by chance, with brain features from rest-fMRI showing the best classification performance in the high-risk group of participants with parental history of depression (N = 625). Specifically, rest-fMRI features, which came from functional connectomes, showed significantly better classification performance than other brain features. This finding highlights the key role of the interacting elements of the connectome in capturing more individual variability in psychopathology compared to measures of single brain regions. Our study contributes to the effort of identifying biological risks of depression in early adolescence in population-based samples.

Details DOI

EAAI Journal 2024 Journal Article

Skeleton-based action recognition based on multidimensional adaptive dynamic temporal graph convolutional network

Yu Xia
Qingyuan Gao
Weiguan Wu
Yi Cao

Due to the superior capability to process the topology of graphs, graph convolutional networks are gaining popularity in the field of action recognition based on skeleton data. However, it remains difficult to effectively extract features with more distinguishing information for both spatial and temporal dimension. A novel multidimensional adaptive dynamic temporal graph convolutional network (MADT-GCN) model for skeleton-based action recognition is proposed in this work. It consists of two modules, one multidimensional adaptive graph convolutional network (MD-AGCN) module and one dynamic temporal convolutional network (DY-TCN) module. Firstly, MD-AGCN has the ability to adaptively change the graph topology in accordance with varieties of the layers and multidimensional information of spatial, temporal, and channel dimensions that are contained in various action samples to capture the complex connections of each couple of joints. Then, DY-TCN is proposed in order to boost the representation capability to capture expressive temporal features. Moreover, the information of both the joints and bones, together with their motion information, are simultaneously modeled in a multi-stream framework, which shows notable improvements in recognition accuracy. Finally, extensive experiments are conducted on two standard datasets, NTU-RGB+D and NTU-RGB+D 120. The experimental results demonstrate the effectiveness of the proposed method.

Details DOI

NeurIPS Conference 2024 Conference Paper

The Closeness of In-Context Learning and Weight Shifting for Softmax Regression

Shuai Li
Zhao Song
Yu Xia
Tong Yu
Tianyi Zhou

Large language models (LLMs) are known for their exceptional performance in natural language processing, making them highly effective in many human life-related tasks. The attention mechanism in the Transformer architecture is a critical component of LLMs, as it allows the model to selectively focus on specific input parts. The softmax unit, which is a key part of the attention mechanism, normalizes the attention scores. Hence, the performance of LLMs in various NLP tasks depends significantly on the crucial role played by the attention mechanism with the softmax unit. In-context learning is one of the celebrated abilities of recent LLMs. Without further parameter updates, Transformers can learn to predict based on few in-context examples. However, the reason why Transformers becomes in-context learners is not well understood. Recently, in-context learning has been studied from a mathematical perspective with simplified linear self-attention without softmax unit. Based on a linear regression formulation $\min_x\| Ax - b \|_2$, existing works show linear Transformers' capability of learning linear functions in context. The capability of Transformers with softmax unit approaching full Transformers, however, remains unexplored. In this work, we study the in-context learning based on a softmax regression formulation $\min_{x} \| \langle \exp(Ax), {\bf 1}_n \rangle^{-1} \exp(Ax) - b \|_2$. We show the upper bounds of the data transformations induced by a single self-attention layer with softmax unit and by gradient-descent on a $\ell_2$ regression loss for softmax prediction function. Our theoretical results imply that when training self-attention-only Transformers for fundamental regression tasks, the models learned by gradient-descent and Transformers show great similarity.

PDF Details DOI

AAAI Conference 2016 Conference Paper

Morphological Segmentation with Window LSTM Neural Networks

Linlin Wang
Zhu Cao
Yu Xia
Gerard de Melo

Morphological segmentation, which aims to break words into meaning-bearing morphemes, is an important task in natural language processing. Most previous work relies heavily on linguistic preprocessing. In this paper, we instead propose novel neural network architectures that learn the structure of input sequences directly from raw input words and are subsequently able to predict morphological boundaries. Our architectures rely on Long Short Term Memory (LSTM) units to accomplish this, but exploit windows of characters to capture more contextual information. Experiments on multiple languages conﬁrm the effectiveness of our models on this task.

PDF Details

JMLR Journal 2012 Journal Article

Active Clustering of Biological Sequences

Konstantin Voevodski
Maria-Florina Balcan
Heiko Röglin
Shang-Hua Teng
Yu Xia

Given a point set S and an unknown metric d on S, we study the problem of efficiently partitioning S into k clusters while querying few distances between the points. In our model we assume that we have access to one versus all queries that given a point s ∈ S return the distances between s and all other points. We show that given a natural assumption about the structure of the instance, we can efficiently find an accurate clustering using only O(k) distance queries. Our algorithm uses an active selection strategy to choose a small set of points that we call landmarks, and considers only the distances between landmarks and other points to produce a clustering. We use our procedure to cluster proteins by sequence similarity. This setting nicely fits our model because we can use a fast sequence database search program to query a sequence against an entire data set. We conduct an empirical study that shows that even though we query a small fraction of the distances between the points, we produce clusterings that are close to a desired clustering given by manual classification. [abs] [ pdf ][ bib ] &copy JMLR 2012. ( edit, beta )

PDF Details