Author name cluster

Feng Tian

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.

28 papers

2 author rows

EAAI Journal 2026 Journal Article

A reinforcement learning-assisted differential evolution with population feature replay

Zijian Cao
Chuhang Qiao
Yanna Wang
Baolong Su
Zhihao Zhang
Feng Tian

As an effective global optimization method, differential evolution (DE) often faces limitations in search capability due to its differential mutation strategy and control parameters. To effectively address diverse problems, adaptively adjusting the control parameter and mutation strategy has become an important research direction. Motivated by this challenge, a reinforcement learning-assisted differential evolution with population feature replay (RLDE-PFR) is proposed. In RLDE-PFR, DE can autonomously adjust mutation strategies and control parameters based on the state of the population. In addition, the proposed population feature replay mechanism (PFR) fully exploits historical successful information to guide parameter generation. To comprehensively validate the performance of RLDE-PFR, we compared it with six state-of-the-art algorithms. The experiments were conducted on the CEC2015 and CEC2017 test suites and a practical Transformer hyperparameter optimization problem. The effectiveness of both reinforcement learning-based autonomous adjustment and the PFR is also validated. The experimental results indicate that RLDE-PFR exhibits competitive and superior performance in terms of resulting accuracy and search efficiency. The RLDE-PFR code is available at https: //github. com/Strive-code/rl-pfr. git.

Details DOI

AAAI Conference 2026 Conference Paper

Enhancing Retrieval-Augmented Large Vision Language Models via Knowledge Conflict Mitigation

Wenbin An
Jiahao Nie
Feng Tian
Mingxiang Cai
Yaqiang Wu
Xiaoqin Zhang
Shijian Lu

Multimodal Retrieval-Augmented Generation (MRAG) has recently been explored to empower Large Vision Language Models (LVLMs) with more comprehensive and up-to-date contextual knowledge, aiming to compensate for their limited and coarse-grained parametric knowledge in knowledge-intensive tasks. However, the retrieved contextual knowledge is usually not aligned with LVLMs’ internal parametric knowledge, leading to knowledge conflicts and further unreliable responses. To tackle this issue, we design KCM, a training-free and plug-and-play framework that can effectively mitigate knowledge conflicts while incorporating MRAG for more accurate LVLM responses. KCM enhances contextual knowledge utilization by modifying the LVLM architecture from three key perspectives. First, KCM adaptively adjusts attention distributions among multiple attention heads, encouraging LVLMs to focus on contextual knowledge with reduced distraction. Second, KCM identifies and prunes knowledge-centric LVLM neurons that encode coarse-grained parametric knowledge, thereby suppressing interferences and enabling more effective integration of contextual knowledge. Third, KCM amplifies the information flow from the input context by injecting supplementary context logits, reinforcing its contribution to the final output. Extensive experiments over multiple LVLMs and benchmarks show that KCM outperforms the state-of-the-art consistently by large margins, incurring neither extra training nor external tools.

PDF Details DOI

AIIM Journal 2026 Journal Article

PreLora: A fine-tuning approach with low-rank matrix decomposition and prefix tuning for pre-hospital emergency text classification

Feng Tian
Xian Wang
Saicong Lu
Jiaxuan Gu
Shitao Zhou
Penghui Li
Zhen Wang
Zengjun Jin

Objective With expanding applications of artificial intelligence technology in the medical field, Large Language Models (LLMs) have achieved substantial success in medical text processing. However, there remain a number of challenges in effectively adapting to specific tasks, such as pre-hospital emergency text classification. Methods We propose a novel fine-tuning method PreLora, which combines prefix tuning with matrix low-rank decomposition. First, this approach incorporates task-specific prompts based on multi-layer perceptron (MLP) encoder into the input. Then, it inject trainable rank-decomposed matrices into every layer of the transformer architecture to compress model parameters, reduce the number of parameters, and capture correlations among the input. To validate its efficacy, we carried out a comparative validation on a pre-hospital emergency text dataset. Results Comparison results indicated that the model fine-tuned with PreLora outperformed the baseline models without fine-tuning, achieving a performance improvement of 45. 4%–75. 4%. Moreover, PreLora ranked first among all fine-tuning methods across each LLM evaluated. An in-depth performance analysis was further conducted on 21 ICD-10 categories with distinct semantic features. The results revealed a negative correlation between model performance and semantic similarity of ICD-10 categories: the low similarity groups performed better, while the high similarity groups performed worse. Notably, PreLora consistently maintained robust performance, with a smaller performance decline in high-similarity categories compared to other fine-tuning methods. In the classifying complex cases with high semantic similarity, PreLora still showed superior adaptability, improving by 68. 6%–95. 8% compared to the baseline model and 0. 4%–8. 4% compared to other fine-tuning methods. Conclusion This study demonstrates PreLora is an effective fine-tuning method to process pre-hospital emergency text classification. It has the potential to expand to other mainstream models for adapting specific tasks in the medical field.

Details DOI

NeurIPS Conference 2025 Conference Paper

Boosting Knowledge Utilization in Multimodal Large Language Models via Adaptive Logits Fusion and Attention Reallocation

Wenbin An
Jiahao Nie
Feng Tian
Haonan Lin
Mingxiang Cai
Yaqiang Wu
Qianying Wang
Xiaoqin Zhang

Despite their recent progress, Multimodal Large Language Models (MLLMs) often struggle in knowledge-intensive tasks due to the limited and outdated parametric knowledge acquired during training. Multimodal Retrieval Augmented Generation addresses this issue by retrieving contextual knowledge from external databases, thereby enhancing MLLMs with expanded knowledge sources. However, existing MLLMs often fail to fully leverage the retrieved contextual knowledge for response generation. We examine representative MLLMs and identify two major causes, namely, attention bias toward different tokens and knowledge conflicts between parametric and contextual knowledge. To this end, we design Adaptive Logits Fusion and Attention Reallocation (ALFAR), a training-free and plug-and-play approach that improves MLLM responses by maximizing the utility of the retrieved knowledge. Specifically, ALFAR tackles the challenges from two perspectives. First, it alleviates attention bias by adaptively shifting attention from visual tokens to relevant context tokens according to query-context relevance. Second, it decouples and weights parametric and contextual knowledge at output logits, mitigating conflicts between the two types of knowledge. As a plug-and-play method, ALFAR achieves superior performance across diverse datasets without requiring additional training or external tools. Extensive experiments over multiple MLLMs and benchmarks show that ALFAR consistently outperforms the state-of-the-art by large margins. Our code and data are available at https: //github. com/Lackel/ALFAR.

PDF Details

TMLR Journal 2025 Journal Article

Distributed Hierarchical Decomposition Framework for Multimodal Timeseries Prediction

Wei Ye
Prashant Khanduri
Jiangweizhi Peng
Feng Tian
Jun Gao
Jie Ding
Zhi-Li Zhang
Mingyi Hong

We consider a distributed time series forecasting problem where multiple distributed nodes each observing a local time series (of potentially different modality) collaborate to make both local and global forecasts. This problem is particularly challenging because each node only observes time series generated from a subset of sources, making it challenging to utilize correlations among different streams for accurate forecasting; and the data streams observed at each node may represent different modalities, leading to heterogeneous computational requirements among nodes. To tackle these challenges, we propose a hierarchical learning framework, consisting of multiple local models and a global model, and provide a suite of efficient training algorithms to achieve high local and global forecasting accuracy. We theoretically establish the convergence of the proposed framework and demonstrate the effectiveness of the proposed approach using several time series forecasting tasks, with the (somewhat surprising) observation that the proposed distributed models can match, or even outperform centralized ones.

PDF Details

JBHI Journal 2025 Journal Article

DoctorPupil: A Virtual Reality System for Parkinson's Diagnosis Through Task-Evoked Pupil Response

Xucheng Zhang
Zhirong Wan
Jing Zhao
Xinjin Li
Anfeng Liu
Xiangmin Fan
Wei Sun
Feng Tian

Parkinson's Disease (PD) is one of the most critical neurodegenerative diseases, yet there is no cure for it, and the state-of-the-art treatment is to slow its progression. Thus, the earlier a patient with PD is recognized, the better he can be treated. Our project joins the research effort that aims to support early PD diagnosis by designing a Virtual Reality (VR)-based system to monitor pupil diameter patterns as new biomarkers (e. g. , Pupil Light Reflex and Task-evoked Pupil Response) and provide early warning of potential PD onset. A follow-up experiment with 55 participants shows that the accuracy of recognizing early PD from healthy controls could reach 0. 8942. Our study shows early results of a promising research direction that leverages VR-based technology to non-intrusively recognize patterns and provide alerts to early PD patients who would otherwise not know their symptoms until much later.

Details DOI

AAAI Conference 2025 Conference Paper

GraphAvatar: Compact Head Avatars with GNN-Generated 3D Gaussians

Xiaobao Wei
Peng Chen
Ming Lu
Hui Chen
Feng Tian

Rendering photorealistic head avatars from arbitrary viewpoints is crucial for various applications like virtual reality. Although previous methods based on Neural Radiance Fields (NeRF) can achieve impressive results, they lack fidelity and efficiency. Recent methods using 3D Gaussian Splatting (3DGS) have improved rendering quality and real-time performance but still require significant storage overhead. In this paper, we introduce a method called GraphAvatar that utilizes Graph Neural Networks (GNN) to generate 3D Gaussians for the head avatar. Specifically, GraphAvatar trains a geometric GNN and an appearance GNN to generate the attributes of the 3D Gaussians from the tracked mesh. Therefore, our method can store the GNN models instead of the 3D Gaussians, significantly reducing the storage overhead to just 10MB. To reduce the impact of face-tracking errors, we also present a novel graph-guided optimization module to refine face-tracking parameters during training. Finally, we introduce a 3D-aware enhancer for post-processing to enhance the rendering quality. We conduct comprehensive experiments to demonstrate the advantages of GraphAvatar, surpassing existing methods in visual fidelity and storage consumption. The ablation study sheds light on the trade-offs between rendering quality and model size.

PDF Details DOI

AAAI Conference 2025 Conference Paper

LLMEmb: Large Language Model Can Be a Good Embedding Generator for Sequential Recommendation

Qidong Liu
Xian Wu
Wanyu Wang
Yejing Wang
Yuanshao Zhu
Xiangyu Zhao
Feng Tian
Yefeng Zheng

Sequential Recommender Systems (SRS), which model a user's interaction history to predict the next item of interest, are widely used in various applications. However, existing SRS often struggle with low-popularity items, a challenge known as the long-tail problem. This issue leads to reduced serendipity for users and diminished profits for sellers, ultimately harming the overall system. Large Language Model (LLM) has the ability to capture semantic relationships between items, independent of their popularity, making them a promising solution to this problem. In this paper, we introduce LLMEmb, a novel method leveraging LLM to generate item embeddings that enhance SRS performance. To bridge the gap between general-purpose LLM and the recommendation domain, we propose a Supervised Contrastive Fine-Tuning (SCFT) approach. This approach includes attribute-level data augmentation and a tailored contrastive loss to make LLM more recommendation-friendly. Additionally, we emphasize the importance of integrating collaborative signals into LLM-generated embeddings, for which we propose Recommendation Adaptation Training (RAT). This further refines the embeddings for optimal use in SRS. The LLMEmb-derived embeddings can be seamlessly integrated with any SRS model, underscoring the practical value. Comprehensive experiments conducted on three real-world datasets demonstrate that LLMEmb significantly outperforms existing methods across multiple SRS models.

PDF Details DOI

ICLR Conference 2025 Conference Paper

PostEdit: Posterior Sampling for Efficient Zero-Shot Image Editing

Feng Tian
Yixuan Li
Yichao Yan
Shanyan Guan
Yanhao Ge
Xiaokang Yang 0001

In the field of image editing, three core challenges persist: controllability, background preservation, and efficiency. Inversion-based methods rely on time-consuming optimization to preserve the features of the initial images, which results in low efficiency due to the requirement for extensive network inference. Conversely, inversion-free methods lack theoretical support for background similarity, as they circumvent the issue of maintaining initial features to achieve efficiency. As a consequence, none of these methods can achieve both high efficiency and background consistency. To tackle the challenges and the aforementioned disadvantages, we introduce PostEdit, a method that incorporates a posterior scheme to govern the diffusion sampling process. Specifically, a corresponding measurement term related to both the initial features and Langevin dynamics is introduced to optimize the estimated image generated by the given target prompt. Extensive experimental results indicate that the proposed PostEdit achieves state-of-the-art editing performance while accurately preserving unedited regions. Furthermore, the method is both inversion- and training-free, necessitating approximately 1.5 seconds and 18 GB of GPU memory to generate high-quality results.

Details

AAAI Conference 2025 Conference Paper

Unleashing the Potential of Model Bias for Generalized Category Discovery

Wenbin An
Haonan Lin
Jiahao Nie
Feng Tian
Wenkai Shi
Yaqiang Wu
Qianying Wang
Ping Chen

Generalized Category Discovery is a significant and complex task that aims to identify both known and undefined novel categories from a set of unlabeled data, leveraging another labeled dataset containing only known categories. The primary challenges stem from model bias induced by pre-training on only known categories and the lack of precise supervision for novel ones, leading to category bias towards known categories and category confusion among different novel categories, which hinders models' ability to identify novel categories effectively. To address these challenges, we propose a novel framework named Self-Debiasing Calibration (SDC). Unlike prior methods that regard model bias towards known categories as an obstacle to novel category identification, SDC provides a novel insight into unleashing the potential of the bias to facilitate novel category learning. Specifically, we utilize the biased pre-trained model to guide the subsequent learning process on unlabeled data. The output of the biased model serves two key purposes. First, it provides an accurate modeling of category bias, which can be utilized to measure the degree of bias and debias the output of the current training model. Second, it offers valuable insights for distinguishing different novel categories by transferring knowledge between similar categories. Based on these insights, SDC dynamically adjusts the output logits of the current training model using the output of the biased model. This approach produces less biased logits to effectively address the issue of category bias towards known categories, and generates more accurate pseudo labels for unlabeled data, thereby mitigating category confusion for novel categories. Experiments on three benchmark datasets show that SDC outperforms SOTA methods, especially in the identification of novel categories.

PDF Details DOI

AAAI Conference 2024 Conference Paper

A Unified Knowledge Transfer Network for Generalized Category Discovery

Wenkai Shi
Wenbin An
Feng Tian
Yan Chen
Yaqiang Wu
Qianying Wang
Ping Chen

Generalized Category Discovery (GCD) aims to recognize both known and novel categories in an unlabeled dataset by leveraging another labeled dataset with only known categories. Without considering knowledge transfer from known to novel categories, current methods usually perform poorly on novel categories due to the lack of corresponding supervision. To mitigate this issue, we propose a unified Knowledge Transfer Network (KTN), which solves two obstacles to knowledge transfer in GCD. First, the mixture of known and novel categories in unlabeled data makes it difficult to identify transfer candidates (i.e., samples with novel categories). For this, we propose an entropy-based method that leverages knowledge in the pre-trained classifier to differentiate known and novel categories without requiring extra data or parameters. Second, the lack of prior knowledge of novel categories presents challenges in quantifying semantic relationships between categories to decide the transfer weights. For this, we model different categories with prototypes and treat their similarities as transfer weights to measure the semantic similarities between categories. On the basis of two treatments, we transfer knowledge from known to novel categories by conducting pre-adjustment of logits and post-adjustment of labels for transfer candidates based on the transfer weights between different categories. With the weighted adjustment, KTN can generate more accurate pseudo-labels for unlabeled data, which helps to learn more discriminative features and boost model performance on novel categories. Extensive experiments show that our method outperforms state-of-the-art models on all evaluation metrics across multiple benchmark datasets. Furthermore, different from previous clustering-based methods that can only work offline with abundant data, KTN can be deployed online conveniently with faster inference speed. Code and data are available at https://github.com/yibai-shi/KTN.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Flipped Classroom: Aligning Teacher Attention with Student in Generalized Category Discovery

Haonan Lin
Wenbin An
Jiahao Wang
Yan Chen
Feng Tian
Mengmeng Wang
Guang Dai
Qianying Wang

Recent advancements have shown promise in applying traditional Semi-Supervised Learning strategies to the task of Generalized Category Discovery (GCD). Typically, this involves a teacher-student framework in which the teacher imparts knowledge to the student to classify categories, even in the absence of explicit labels. Nevertheless, GCD presents unique challenges, particularly the absence of priors for new classes, which can lead to the teacher's misguidance and unsynchronized learning with the student, culminating in suboptimal outcomes. In our work, we delve into why traditional teacher-student designs falter in generalized category discovery as compared to their success in closed-world semi-supervised learning. We identify inconsistent pattern learning as the crux of this issue and introduce FlipClass—a method that dynamically updates the teacher to align with the student's attention, instead of maintaining a static teacher reference. Our teacher-attention-update strategy refines the teacher's focus based on student feedback, promoting consistent pattern recognition and synchronized learning across old and new classes. Extensive experiments on a spectrum of benchmarks affirm that FlipClass significantly surpasses contemporary GCD methods, establishing new standards for the field.

PDF Details DOI

AAAI Conference 2024 Conference Paper

G2L-CariGAN: Caricature Generation from Global Structure to Local Features

Xin Huang
Yunfeng Bai
Dong Liang
Feng Tian
Jinyuan Jia

Existing GAN-based approaches to caricature generation mainly focus on exaggerating a character’s global facial structure. This often leads to the failure in highlighting significant facial features such as big eyes and hook nose. To address this limitation, we propose a new approach termed as G2L-CariGAN, which uses feature maps of spatial dimensions instead of latent codes for geometric exaggeration. G2L-CariGAN first exaggerates the global facial structure of the character on a low-dimensional feature map and then exaggerates its local facial features on a high-dimensional feature map. Moreover, we develop a caricature identity loss function based on feature maps, which well retains the character's identity after exaggeration. Our experiments have demonstrated that G2L-CariGAN outperforms the state-of-arts in terms of the quality of exaggerating a character and retaining its identity.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

LLM-ESR: Large Language Models Enhancement for Long-tailed Sequential Recommendation

Qidong Liu
Xian Wu
Yejing Wang
Zijian Zhang
Feng Tian
Yefeng Zheng
Xiangyu Zhao

Sequential recommender systems (SRS) aim to predict users' subsequent choices based on their historical interactions and have found applications in diverse fields such as e-commerce and social media. However, in real-world systems, most users interact with only a handful of items, while the majority of items are seldom consumed. These two issues, known as the long-tail user and long-tail item challenges, often pose difficulties for existing SRS. These challenges can adversely affect user experience and seller benefits, making them crucial to address. Though a few works have addressed the challenges, they still struggle with the seesaw or noisy issues due to the intrinsic scarcity of interactions. The advancements in large language models (LLMs) present a promising solution to these problems from a semantic perspective. As one of the pioneers in this field, we propose the Large Language Models Enhancement framework for Sequential Recommendation (LLM-ESR). This framework utilizes semantic embeddings derived from LLMs to enhance SRS without adding extra inference load. To address the long-tail item challenge, we design a dual-view modeling framework that combines semantics from LLMs and collaborative signals from conventional SRS. For the long-tail user challenge, we propose a retrieval augmented self-distillation method to enhance user preference representation using more informative interactions from similar users. To verify the effectiveness and versatility of our proposed enhancement framework, we conduct extensive experiments on three real-world datasets using three popular SRS models. The results consistently show that our method surpasses existing baselines. The implementation code is available in Supplementary Material.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Schedule Your Edit: A Simple yet Effective Diffusion Noise Schedule for Image Editing

Haonan Lin
Yan Chen
Jiahao Wang
Wenbin An
Mengmeng Wang
Feng Tian
Yong Liu
Guang Dai

Text-guided diffusion models have significantly advanced image editing, enabling high-quality and diverse modifications driven by text prompts. However, effective editing requires inverting the source image into a latent space, a process often hindered by prediction errors inherent in DDIM inversion. These errors accumulate during the diffusion process, resulting in inferior content preservation and edit fidelity, especially with conditional inputs. We address these challenges by investigating the primary contributors to error accumulation in DDIM inversion and identify the singularity problem in traditional noise schedules as a key issue. To resolve this, we introduce the Logistic Schedule, a novel noise schedule designed to eliminate singularities, improve inversion stability, and provide a better noise space for image editing. This schedule reduces noise prediction errors, enabling more faithful editing that preserves the original content of the source image. Our approach requires no additional retraining and is compatible with various existing editing methods. Experiments across eight editing tasks demonstrate the Logistic Schedule's superior performance in content preservation and edit fidelity compared to traditional noise schedules, highlighting its adaptability and effectiveness. The project page is available at https: //lonelvino. github. io/SYE/.

PDF Details DOI

AAAI Conference 2024 Conference Paper

Transfer and Alignment Network for Generalized Category Discovery

Wenbin An
Feng Tian
Wenkai Shi
Yan Chen
Yaqiang Wu
Qianying Wang
Ping Chen

Generalized Category Discovery (GCD) is a crucial real-world task that aims to recognize both known and novel categories from an unlabeled dataset by leveraging another labeled dataset with only known categories. Despite the improved performance on known categories, current methods perform poorly on novel categories. We attribute the poor performance to two reasons: biased knowledge transfer between labeled and unlabeled data and noisy representation learning on the unlabeled data. The former leads to unreliable estimation of learning targets for novel categories and the latter hinders models from learning discriminative features. To mitigate these two issues, we propose a Transfer and Alignment Network (TAN), which incorporates two knowledge transfer mechanisms to calibrate the biased knowledge and two feature alignment mechanisms to learn discriminative features. Specifically, we model different categories with prototypes and transfer the prototypes in labeled data to correct model bias towards known categories. On the one hand, we pull instances with known categories in unlabeled data closer to these prototypes to form more compact clusters and avoid boundary overlap between known and novel categories. On the other hand, we use these prototypes to calibrate noisy prototypes estimated from unlabeled data based on category similarities, which allows for more accurate estimation of prototypes for novel categories that can be used as reliable learning targets later. After knowledge transfer, we further propose two feature alignment mechanisms to acquire both instance- and category-level knowledge from unlabeled data by aligning instance features with both augmented features and the calibrated prototypes, which can boost model performance on both known and novel categories with less noise. Experiments on three benchmark datasets show that our model outperforms SOTA methods, especially on novel categories. Theoretical analysis is provided for an in-depth understanding of our model in general. Our code and data are available at https://github.com/Lackel/TAN.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Generalized Category Discovery with Decoupled Prototypical Network

Wenbin An
Feng Tian
Qinghua Zheng
Wei Ding
Qianying Wang
Ping Chen

Generalized Category Discovery (GCD) aims to recognize both known and novel categories from a set of unlabeled data, based on another dataset labeled with only known categories. Without considering differences between known and novel categories, current methods learn about them in a coupled manner, which can hurt model's generalization and discriminative ability. Furthermore, the coupled training approach prevents these models transferring category-specific knowledge explicitly from labeled data to unlabeled data, which can lose high-level semantic information and impair model performance. To mitigate above limitations, we present a novel model called Decoupled Prototypical Network (DPN). By formulating a bipartite matching problem for category prototypes, DPN can not only decouple known and novel categories to achieve different training targets effectively, but also align known categories in labeled and unlabeled data to transfer category-specific knowledge explicitly and capture high-level semantics. Furthermore, DPN can learn more discriminative features for both known and novel categories through our proposed Semantic-aware Prototypical Learning (SPL). Besides capturing meaningful semantic information, SPL can also alleviate the noise of hard pseudo labels through semantic-weighted soft assignment. Extensive experiments show that DPN outperforms state-of-the-art models by a large margin on all evaluation metrics across multiple benchmark datasets. Code and data are available at https://github.com/Lackel/DPN.

PDF Details DOI

IS Journal 2023 Journal Article

New User Intent Discovery With Robust Pseudo Label Training and Source Domain Joint Training

Wenbin An
Feng Tian
Ping Chen
Qinghua Zheng
Wei Ding

Discovering new user intents based on existing intents from constantly incoming unlabeled data is an important task in many intelligent systems deployed in the real world (e. g. , dialogue systems). Since data with new intents are completely unlabeled, most current approaches employ clustering methods to generate pseudo labels to train their models. However, due to intent gaps between existing and new intents, pseudo labels generated by these models are noisy, and prior knowledge from existing intents is not fully utilized. To mitigate these issues, we propose a robust pseudo label training and source domain joint-training network to refine the noisy pseudo labels and make full use of prior knowledge. Experimental results on three intent detection datasets show that our model is more effective and robust than state-of-the-art methods. The code and data are released at https://github.com/Lackel/PTJN.

Details DOI

EAAI Journal 2020 Journal Article

Affinity matrix with large eigenvalue gap for graph-based subspace clustering and semi-supervised classification

Xiaofang Liu
Jun Wang
Dansong Cheng
Feng Tian
Yongqiang Zhang

In the graph-based learning method, the data graph or similarity matrix reveals the relationship between data, and reflects similar attributes within a class and differences between classes. Inspired by Davis–Kahan Theorem that the stability of matrix eigenvector space depends on its spectral distance (i. e. its eigenvalue gap), in this paper, we propose a global local affinity matrix model with low rank subspace sparse representation (GLAM-LRSR) based on global information of eigenvalue gap and local distance between samples. This method approximate the similarity matrix with ideally diagonal block structure from the perspective of maximizing the eigenvalue gap, and the local distance between data is utilized as a regular term to prevent the eigenvalue gap from being too large to ensure the efficacy of similarity matrix. We have shown that the combination of subspace (LRSR) partitioning method such as Sparse Subspace Clustering(SSC) and the similarity matrix constructed by GLAM can improve the accuracy of subspace clustering, and that the similarity matrix constructed by GLAM-LRSR can be successfully applied to graph-based semi-supervised classification task. Our experiments on synthetic data as well as the real-world datasets for face clustering, face recovery and motion segmentation have clearly demonstrate the significant advantages of GLAM-LRSR and its effectiveness.

Details DOI

TCS Journal 2020 Journal Article

Bounded privacy-utility monotonicity indicating bounded tradeoff of differential privacy mechanisms

Hai Liu
Zhenqiang Wu
Changgen Peng
Feng Tian
Laifeng Lu

Differential privacy can achieve the tradeoff between privacy and utility by using privacy metric and utility metric. However, since privacy metric and utility metric may not be bounded, differential privacy can not provide the bounded tradeoff. Moreover, there is no unified method to indicate the bounded tradeoff of differential privacy in the current work. To this end, we proposed the bounded privacy-utility monotonicity indicating the bounded tradeoff of differential privacy. First, we gave the definition of the bounded tradeoff of differential privacy, and we presented the bounded privacy-utility monotonicity of differential privacy based on computational indistinguishability. Second, we theoretically proved the bounded privacy-utility monotonicity of several differential privacy mechanisms based on the bounded metrics of modulus of characteristic function and normalized entropy, including the Laplace mechanism, discrete Laplace mechanism, Gaussian mechanism, exponential mechanism, optimal mechanism, and quaternary mechanism. We also showed that these mechanisms had the bounded privacy-utility monotonicity in the multivariate case. Third, our numerical results further demonstrated that these several differential privacy mechanisms obtained the bounded privacy-utility monotonicity. Finally, we gave an instance of achieving the bounded tradeoff of differential privacy mechanisms based on the bounded privacy-utility monotonicity under semi-honest model, and we discussed the goal of optimization of the bounded tradeoff of differential privacy based on the bounded privacy-utility monotonicity under semi-honest model. Therefore, the bounded privacy-utility monotonicity can be used to indicate the bounded tradeoff of differential privacy under semi-honest model. Furthermore, the bounded privacy-utility monotonicity plays an important role of optimizing the bounded tradeoff of differential privacy under semi-honest model.

Details DOI

IJCAI Conference 2019 Conference Paper

Attributed Subspace Clustering

Jing Wang
Linchuan Xu
Feng Tian
Atsushi Suzuki
Changqing Zhang
Kenji Yamanishi

Existing methods on representation-based subspace clustering mainly treat all features of data as a whole to learn a single self-representation and get one clustering solution. Real data however are often complex and consist of multiple attributes or sub-features, such as a face image has expressions or genders. Each attribute is distinct and complementary on depicting the data. Failing to explore attributes and capture the complementary information among them may lead to an inaccurate representation. Moreover, a single clustering solution is rather limited to depict data, which can often be interpreted from different aspects and grouped into multiple clusters according to attributes. Therefore, we propose an innovative model called attributed subspace clustering (ASC). It simultaneously learns multiple self-representations on latent representations derived from original data. By utilizing Hilbert Schmidt Independence Criterion as a co-regularizing term, ASC enforces that each self-representation is independent and corresponds to a specific attribute. A more comprehensive self-representation is then established by adding these self-representations. Experiments on several benchmark image datasets have demonstrated the effectiveness of ASC not only in terms of clustering accuracy achieved by the integrated representation, but also the diverse interpretation of data, which is beyond what current approaches can offer.

PDF Details

AAAI Conference 2019 Conference Paper

Orderly Subspace Clustering

Jing Wang
Atsushi Suzuki
Linchuan Xu
Feng Tian
Liang Yang
Kenji Yamanishi

Semi-supervised representation-based subspace clustering is to partition data into their underlying subspaces by finding effective data representations with partial supervisions. Essentially, an effective and accurate representation should be able to uncover and preserve the true data structure. Meanwhile, a reliable and easy-to-obtain supervision is desirable for practical learning. To meet these two objectives, in this paper we make the first attempt towards utilizing the orderly relationship, such as the data a is closer to b than to c, as a novel supervision. We propose an orderly subspace clustering approach with a novel regularization term. OSC enforces the learned representations to simultaneously capture the intrinsic subspace structure and reveal orderly structure that is faithful to true data relationship. Experimental results with several benchmarks have demonstrated that aside from more accurate clustering against state-of-the-arts, OSC interprets orderly data structure which is beyond what current approaches can offer.

PDF Details

IJCAI Conference 2018 Conference Paper

Outer Product-based Neural Collaborative Filtering

Xiangnan He
Xiaoyu Du
Xiang Wang
Feng Tian
Jinhui Tang
Tat-Seng Chua

In this work, we contribute a new multi-layer neural network architecture named ONCF to perform collaborative filtering. The idea is to use an outer product to explicitly model the pairwise correlations between the dimensions of the embedding space. In contrast to existing neural recommender models that combine user embedding and item embedding via a simple concatenation or element-wise product, our proposal of using outer product above the embedding layer results in a two-dimensional interaction map that is more expressive and semantically plausible. Above the interaction map obtained by outer product, we propose to employ a convolutional neural network to learn high-order correlations among embedding dimensions. Extensive experiments on two public implicit feedback data demonstrate the effectiveness of our proposed ONCF framework, in particular, the positive effect of using outer product to model the correlations between embedding dimensions in the low level of multi-layer neural recommender model.

PDF Details

IJCAI Conference 2018 Conference Paper

Ranking Preserving Nonnegative Matrix Factorization

Jing Wang
Feng Tian
Weiwei Liu
Xiao Wang
Wenjie Zhang
Kenji Yamanishi

Nonnegative matrix factorization (NMF), a well-known technique to find parts-based representations of nonnegative data, has been widely studied. In reality, ordinal relations often exist among data, such as data i is more related to j than to q. Such relative order is naturally available, and more importantly, it truly reflects the latent data structure. Preserving the ordinal relations enables us to find structured representations of data that are faithful to the relative order, so that the learned representations become more discriminative. However, current NMFs pay no attention to this. In this paper, we make the first attempt towards incorporating the ordinal relations and propose a novel ranking preserving nonnegative matrix factorization (RPNMF) approach, which enforces the learned representations to be ranked according to the relations. We derive iterative updating rules to solve RPNMF's objective function with convergence guaranteed. Experimental results with several datasets for clustering and classification have demonstrated that RPNMF achieves greater performance against the state-of-the-arts, not only in terms of accuracy, but also interpretation of orderly data structure.

PDF Details

IJCAI Conference 2017 Conference Paper

Multi-Component Nonnegative Matrix Factorization

Jing Wang
Feng Tian
Xiao Wang
Hongchuan Yu
Chang Hong Liu
Liang Yang

Real data are usually complex and contain various components. For example, face images have expressions and genders. Each component mainly reflects one aspect of data and provides information others do not have. Therefore, exploring the semantic information of multiple components as well as the diversity among them is of great benefit to understand data comprehensively and in-depth. However, this cannot be achieved by current nonnegative matrix factorization (NMF)-based methods, despite that NMF has shown remarkable competitiveness in learning parts-based representation of data. To overcome this limitation, we propose a novel multi-component nonnegative matrix factorization (MCNMF). Instead of seeking for only one representation of data, MCNMF learns multiple representations simultaneously, with the help of the Hilbert Schmidt Independence Criterion (HSIC) as a diversity term. HSIC explores the diverse information among the representations, where each representation corresponds to a component. By integrating the multiple representations, a more comprehensive representation is then established. A new iterative updating optimization scheme is derived to solve the objective function of MCNMF, along with its correctness and convergence guarantees. Extensive experimental results on real-world datasets have shown that MCNMF not only achieves more accurate performance over the state-of-the-arts using the aggregated representation, but also interprets data from different aspects with the multiple representations, which is beyond what current NMFs can offer.

PDF Details

JMLR Journal 2015 Journal Article

Discrete Reproducing Kernel Hilbert Spaces: Sampling and Distribution of Dirac-masses

Palle Jorgensen
Feng Tian

We study reproducing kernels, and associated reproducing kernel Hilbert spaces (RKHSs) $\mathscr{H}$ over infinite, discrete and countable sets $V$. In this setting we analyze in detail the distributions of the corresponding Dirac point-masses of $V$. Illustrations include certain models from neural networks: An Extreme Learning Machine (ELM) is a neural network-configuration in which a hidden layer of weights are randomly sampled, and where the object is then to compute resulting output. For RKHSs $\mathscr{H}$ of functions defined on a prescribed countable infinite discrete set $V$, we characterize those which contain the Dirac masses $\delta_{x}$ for all points $x$ in $V$. Further examples and applications where this question plays an important role are: (i) discrete Brownian motion-Hilbert spaces, i.e., discrete versions of the Cameron-Martin Hilbert space; (ii) energy-Hilbert spaces corresponding to graph-Laplacians where the set $V$ of vertices is then equipped with a resistance metric; and finally (iii) the study of Gaussian free fields. [abs] [ pdf ][ bib ] &copy JMLR 2015. ( edit, beta )

PDF Details

IJCAI Conference 2015 Conference Paper

Stroke-Based Stylization Learning and Rendering with Inverse Reinforcement Learning

Ning Xie
Tingting Zhao
Feng Tian
Xiao Hua Zhang
Masashi Sugiyama

Among various traditional art forms, brush stroke drawing is one of the widely used styles in modern computer graphic tools such as GIMP, Photoshop and Painter. In this paper, we develop an AI-aided art authoring (A4) system of nonphotorealistic rendering that allows users to automatically generate brush stroke paintings in a specific artist’s style. Within the reinforcement learning framework of brush stroke generation proposed by Xie et al. [Xie et al. , 2012], our contribution in this paper is to learn artists’ drawing styles from video-captured stroke data by inverse reinforcement learning. Through experiments, we demonstrate that our system can successfully learn artists’ styles and render pictures with consistent and smooth brush strokes.

PDF Details

TIST Journal 2011 Journal Article

Understanding, Manipulating and Searching Hand-Drawn Concept Maps

Yingying Jiang
Feng Tian
Xiaolong (Luke) Zhang
Guozhong Dai
Hongan Wang

Concept maps are an important tool to organize, represent, and share knowledge. Building a concept map involves creating text-based concepts and specifying their relationships with line-based links. Current concept map tools usually impose specific task structures for text and link construction, and may increase cognitive burden to generate and interact with concept maps. While pen-based devices (e.g., tablet PCs) offer users more freedom in drawing concept maps with a pen or stylus more naturally, the support for hand-drawn concept map creation and manipulation is still limited, largely due to the lack of methods to recognize the components and structures of hand-drawn concept maps. This article proposes a method to understand hand-drawn concept maps. Our algorithm can extract node blocks, or concept blocks, and link blocks of a hand-drawn concept map by combining dynamic programming and graph partitioning, recognize the text content of each concept node, and build a concept-map structure by relating concepts and links. We also design an algorithm for concept map retrieval based on hand-drawn queries. With our algorithms, we introduce structure-based intelligent manipulation techniques and ink-based retrieval techniques to support the management and modification of hand-drawn concept maps. Results from our evaluation study show high structure recognition accuracy in real time of our method, and good usability of intelligent manipulation and retrieval techniques.

Details DOI