Xingxing Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xingxing Zhang has a total of 95 co-authors. Here are the top co-authors by number of shared publications:
- Yao Zhao
Institute of Information Science, Beijing Jiaotong University, Beijing, China and Visual Intelligence + X International Joint Laboratory, Beijing, China
13 shared publications - Zhenfeng Zhu
Beijing Jiaotong University
13 shared publications - Furu Wei
Microsoft Corporation
8 shared publications - Ming Zhou
Langboat Technology
5 shared publications - Liyuan Wang
Department of Psychological and Cognitive Sciences, Tsinghua University, Beijing, China
4 shared publications - Jun Zhu
Tsinghua University, Beijing, China
3 shared publications
Latest Publications
- 24 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (3) , 3304-3318
BiKT: Unleashing the Potential of GNNs via Bi-Directional Knowledge Transfer
Based on the message-passing paradigm, there has been an amount of research proposing diverse and impressive feature propagation mechanisms to improve the performance of GNNs. However, less focus has been put on feature transformation, another major operation of the message-passing framework. In this paper, we first empirically investigate the performance of the feature transformation operation in several typical GNNs. Unexpectedly, we notice that GNNs do not completely free up the power of the inherent feature transformation operation. By this observation, we propose the Bi-directional Knowledge Transfer (BiKT), a plug-and-play approach to unleash the potential of the feature transformation operations without modifying the original architecture. Taking the feature transformation operation as a derived representation learning model that shares parameters with the original GNN, the direct prediction by this model provides a topological-agnostic knowledge feedback that can further instruct the learning of GNN and the feature transformations therein. On this basis, BiKT not only allows us to acquire knowledge from both the GNN and its derived model but also promotes each other by injecting the knowledge into the other. In addition, a theoretical analysis is further provided to demonstrate that BiKT improves the generalization bound of the GNNs from the perspective of domain adaptation. An extensive group of experiments on up to 7 datasets with 5 typical GNNs demonstrates that BiKT brings up to 0.5% - 4% performance gain over the original GNN, which means a boosted GNN is obtained. Meanwhile, the derived model also shows a powerful performance to compete with or even surpass the original GNN, enabling us to flexibly apply it independently to some other specific downstream tasks.
- 24 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 48 (3) , 3304-3318
BiKT: Unleashing the Potential of GNNs via Bi-Directional Knowledge Transfer
Based on the message-passing paradigm, there has been an amount of research proposing diverse and impressive feature propagation mechanisms to improve the performance of GNNs. However, less focus has been put on feature transformation, another major operation of the message-passing framework. In this paper, we first empirically investigate the performance of the feature transformation operation in several typical GNNs. Unexpectedly, we notice that GNNs do not completely free up the power of the inherent feature transformation operation. By this observation, we propose the Bi-directional Knowledge Transfer (BiKT), a plug-and-play approach to unleash the potential of the feature transformation operations without modifying the original architecture. Taking the feature transformation operation as a derived representation learning model that shares parameters with the original GNN, the direct prediction by this model provides a topological-agnostic knowledge feedback that can further instruct the learning of GNN and the feature transformations therein. On this basis, BiKT not only allows us to acquire knowledge from both the GNN and its derived model but also promotes each other by injecting the knowledge into the other. In addition, a theoretical analysis is further provided to demonstrate that BiKT improves the generalization bound of the GNNs from the perspective of domain adaptation. An extensive group of experiments on up to 7 datasets with 5 typical GNNs demonstrates that BiKT brings up to 0.5% - 4% performance gain over the original GNN, which means a boosted GNN is obtained. Meanwhile, the derived model also shows a powerful performance to compete with or even surpass the original GNN, enabling us to flexibly apply it independently to some other specific downstream tasks.
- 11 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (4) , 4287-4301
Rethinking Obscured Sub-Optimality in Analytic Learning for Exemplar-Free Class-Incremental Learning
Exemplar-free Class-Incremental Learning (EFCIL) poses a significant challenge in mitigating catastrophic forgetting, due to the absence of exemplars. Recently, analytic learning-based methods propose a recursive alignment procedure to execute EFCIL in a phase-invariant manner and show state-of-the-art performance. However, they heavily rely on a frozen feature extractor trained with the initial dataset to avoid the misalignment between feature and label spaces, ignoring the importance of acquiring generalizable features across incremental tasks for performance improvement. To tackle this, we rethink the obscured sub-optimality of analytic learning-based methods, particularly through empirical reevaluation, and then introduce the Multi-head analytic learning (Muheal) approach. Muheal forms the multi-head model with a delicate feature extractor, thereby introducing a feature optimization procedure and a forgetting compensation module to balance the learning and forgetting. Specifically, within the feature optimization procedure, the feature extractor seeks to learn more generalizable features in a self-supervised manner using the fully-connected classification head. An analytic learning-based classification head follows to align the feature-label space. Additionally, we employ the compensation module to generate and align pseudo-features with a replicated analytic head, thus preventing overfitting and testing. Comprehensive experiments on several benchmark datasets have demonstrated that Muheal significantly outperforms existing state-of-the-art EFCIL methods and is comparable, if not superior, to methods that use replay techniques.
- 11 November 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (4) , 4287-4301
Rethinking Obscured Sub-Optimality in Analytic Learning for Exemplar-Free Class-Incremental Learning
Exemplar-free Class-Incremental Learning (EFCIL) poses a significant challenge in mitigating catastrophic forgetting, due to the absence of exemplars. Recently, analytic learning-based methods propose a recursive alignment procedure to execute EFCIL in a phase-invariant manner and show state-of-the-art performance. However, they heavily rely on a frozen feature extractor trained with the initial dataset to avoid the misalignment between feature and label spaces, ignoring the importance of acquiring generalizable features across incremental tasks for performance improvement. To tackle this, we rethink the obscured sub-optimality of analytic learning-based methods, particularly through empirical reevaluation, and then introduce the Multi-head analytic learning (Muheal) approach. Muheal forms the multi-head model with a delicate feature extractor, thereby introducing a feature optimization procedure and a forgetting compensation module to balance the learning and forgetting. Specifically, within the feature optimization procedure, the feature extractor seeks to learn more generalizable features in a self-supervised manner using the fully-connected classification head. An analytic learning-based classification head follows to align the feature-label space. Additionally, we employ the compensation module to generate and align pseudo-features with a replicated analytic head, thus preventing overfitting and testing. Comprehensive experiments on several benchmark datasets have demonstrated that Muheal significantly outperforms existing state-of-the-art EFCIL methods and is comparable, if not superior, to methods that use replay techniques.
- 16 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (3) , 3231-3245
Dynamic Confidence Variance for Generalized Coreset in Active Learning
Active Learning (AL) aims to reduce data annotation costs by selecting the most informative samples from an unlabeled data pool. Traditional AL methods often rely on a single snapshot to identify uncertain or representative samples, often overlooking the poor generalization of a single model. Recent AL studies have attempted to address this issue by tracking a broader range of training dynamics for data selection, typically using averaging or accumulating manner. However, both our theoretical and experimental analyses reveal that these methods obscure the variability inherent in the training process, potentially prioritizing hard-to-learn samples that result in poor generalization. In this paper, we propose a novel AL method termed as Dynamic Confidence Variance (DCoV), that seamlessly integrates variability with the training dynamic to effectively identify a well-generalized Coreset. DCoV leverages the variance of the model’s prediction confidence throughout the training process for active sampling and model training. Our theoretical analysis demonstrates that DCoV provides a lower bound on the population risk of the model learned from selected labeled subset, spanning the entire training process. Extensive experiments demonstrate that our approach significantly outperforms existing state-of-the-art AL methods on various balanced and imbalanced benchmark datasets across various modalities.
- 16 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (3) , 3231-3245
Dynamic Confidence Variance for Generalized Coreset in Active Learning
Active Learning (AL) aims to reduce data annotation costs by selecting the most informative samples from an unlabeled data pool. Traditional AL methods often rely on a single snapshot to identify uncertain or representative samples, often overlooking the poor generalization of a single model. Recent AL studies have attempted to address this issue by tracking a broader range of training dynamics for data selection, typically using averaging or accumulating manner. However, both our theoretical and experimental analyses reveal that these methods obscure the variability inherent in the training process, potentially prioritizing hard-to-learn samples that result in poor generalization. In this paper, we propose a novel AL method termed as Dynamic Confidence Variance (DCoV), that seamlessly integrates variability with the training dynamic to effectively identify a well-generalized Coreset. DCoV leverages the variance of the model’s prediction confidence throughout the training process for active sampling and model training. Our theoretical analysis demonstrates that DCoV provides a lower bound on the population risk of the model learned from selected labeled subset, spanning the entire training process. Extensive experiments demonstrate that our approach significantly outperforms existing state-of-the-art AL methods on various balanced and imbalanced benchmark datasets across various modalities.
- Open Access
Continual Action Quality Assessment via Adaptive Manifold-Aligned Graph Regularization
Action Quality Assessment (AQA) quantifies human actions in videos, supporting applications in sports scoring, rehabilitation, and skill evaluation. A major challenge lies in the non-stationary nature of quality distributions in real-world scenarios, which limits the generalization ability of conventional methods. We introduce Continual AQA (CAQA), which equips AQA with Continual Learning (CL) capabilities to handle evolving distributions while mitigating catastrophic forgetting. Although parameter-efficient fine-tuning of pretrained models has shown promise in CL for image classification, we find it insufficient for CAQA. Our empirical and theoretical analyses reveal two insights: (i) Full-Parameter Fine-Tuning (FPFT) is necessary for effective representation learning; yet (ii) uncontrolled FPFT induces overfitting and feature manifold shift, thereby aggravating forgetting. To address this, we propose Adaptive Manifold-Aligned Graph Regularization (MAGR++), which couples backbone fine-tuning that stabilizes shallow layers while adapting deeper ones with a two-step feature rectification pipeline: a manifold projector to translate deviated historical features into the current representation space, and a graph regularizer to align local and global distributions. We construct four CAQA benchmarks from three datasets with tailored evaluation protocols and strong baselines, enabling systematic cross-dataset comparison. Extensive experiments show that MAGR++ achieves state-of-the-art performance, with average correlation gains of 3.6% offline and 12.2% online over the strongest baseline, confirming its robustness and effectiveness. Our code is available at https://github.com/ZhouKanglei/MAGRPP.
Open AccessContinual Action Quality Assessment via Adaptive Manifold-Aligned Graph Regularization
Action Quality Assessment (AQA) quantifies human actions in videos, supporting applications in sports scoring, rehabilitation, and skill evaluation. A major challenge lies in the non-stationary nature of quality distributions in real-world scenarios, which limits the generalization ability of conventional methods. We introduce Continual AQA (CAQA), which equips AQA with Continual Learning (CL) capabilities to handle evolving distributions while mitigating catastrophic forgetting. Although parameter-efficient fine-tuning of pretrained models has shown promise in CL for image classification, we find it insufficient for CAQA. Our empirical and theoretical analyses reveal two insights: (i) Full-Parameter Fine-Tuning (FPFT) is necessary for effective representation learning; yet (ii) uncontrolled FPFT induces overfitting and feature manifold shift, thereby aggravating forgetting. To address this, we propose Adaptive Manifold-Aligned Graph Regularization (MAGR++), which couples backbone fine-tuning that stabilizes shallow layers while adapting deeper ones with a two-step feature rectification pipeline: a manifold projector to translate deviated historical features into the current representation space, and a graph regularizer to align local and global distributions. We construct four CAQA benchmarks from three datasets with tailored evaluation protocols and strong baselines, enabling systematic cross-dataset comparison. Extensive experiments show that MAGR++ achieves state-of-the-art performance, with average correlation gains of 3.6% offline and 12.2% online over the strongest baseline, confirming its robustness and effectiveness. Our code is available at https://github.com/ZhouKanglei/MAGRPP.
- 4 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 34 (10577149) , 3718-3732
PHI: Bridging Domain Shift in Long-Term Action Quality Assessment via Progressive Hierarchical Instruction
Long-term Action Quality Assessment (AQA) aims to evaluate the quantitative performance of actions in long videos. However, existing methods face challenges due to domain shifts between the pre-trained large-scale action recognition backbones and the specific AQA task, thereby hindering their performance. This arises since fine-tuning resource-intensive backbones on small AQA datasets is impractical. We address this by identifying two levels of domain shift: task-level, regarding differences in task objectives, and feature-level, regarding differences in important features. For feature-level shifts, which are more detrimental, we propose Progressive Hierarchical Instruction (PHI) with two strategies. First, Gap Minimization Flow (GMF) leverages flow matching to progressively learn a fast flow path that reduces the domain gap between initial and desired features across shallow to deep layers. Additionally, a temporally-enhanced attention module captures long-range dependencies essential for AQA. Second, List-wise Contrastive Regularization (LCR) facilitates coarse-to-fine alignment by comprehensively comparing batch pairs to learn fine-grained cues while mitigating domain shift. Integrating these modules, PHI offers an effective solution. Experiments demonstrate that PHI achieves state-of-the-art performance on three representative long-term AQA datasets, proving its superiority in addressing the domain shift for long-term AQA.
- 4 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 34 (10577149) , 3718-3732
PHI: Bridging Domain Shift in Long-Term Action Quality Assessment via Progressive Hierarchical Instruction
Long-term Action Quality Assessment (AQA) aims to evaluate the quantitative performance of actions in long videos. However, existing methods face challenges due to domain shifts between the pre-trained large-scale action recognition backbones and the specific AQA task, thereby hindering their performance. This arises since fine-tuning resource-intensive backbones on small AQA datasets is impractical. We address this by identifying two levels of domain shift: task-level, regarding differences in task objectives, and feature-level, regarding differences in important features. For feature-level shifts, which are more detrimental, we propose Progressive Hierarchical Instruction (PHI) with two strategies. First, Gap Minimization Flow (GMF) leverages flow matching to progressively learn a fast flow path that reduces the domain gap between initial and desired features across shallow to deep layers. Additionally, a temporally-enhanced attention module captures long-range dependencies essential for AQA. Second, List-wise Contrastive Regularization (LCR) facilitates coarse-to-fine alignment by comprehensively comparing batch pairs to learn fine-grained cues while mitigating domain shift. Integrating these modules, PHI offers an effective solution. Experiments demonstrate that PHI achieves state-of-the-art performance on three representative long-term AQA datasets, proving its superiority in addressing the domain shift for long-term AQA.
- 18 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (8) , 6687-6702
HiDe-PET: Continual Learning via Hierarchical Decomposition of Parameter-Efficient Tuning
The deployment of pre-trained models (PTMs) has greatly advanced the field of continual learning (CL), enabling positive knowledge transfer and resilience to catastrophic forgetting. To sustain these advantages for sequentially arriving tasks, a promising direction involves keeping the pre-trained backbone frozen while employing parameter-efficient tuning (PET) techniques to instruct representation learning. Despite the popularity of Prompt-based PET for CL, its empirical design often leads to sub-optimal performance in our evaluation of different PTMs and target tasks. To this end, we propose a unified framework for CL with PTMs and PET that provides both theoretical and empirical advancements. We first perform an in-depth theoretical analysis of the CL objective in a pre-training context, decomposing it into hierarchical components namely within-task prediction, task-identity inference and task-adaptive prediction. We then present Hierarchical Decomposition PET (HiDe-PET), an innovative approach that explicitly optimizes the decomposed objective through incorporating task-specific and task-shared knowledge via mainstream PET techniques along with efficient recovery of pre-trained representations. Leveraging this framework, we delve into the distinct impacts of implementation strategy, PET technique and PET architecture, as well as adaptive knowledge accumulation amidst pronounced distribution changes. Finally, across various CL scenarios, our approach demonstrates remarkably superior performance over a broad spectrum of recent strong baselines.
- 18 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (8) , 6687-6702
HiDe-PET: Continual Learning via Hierarchical Decomposition of Parameter-Efficient Tuning
The deployment of pre-trained models (PTMs) has greatly advanced the field of continual learning (CL), enabling positive knowledge transfer and resilience to catastrophic forgetting. To sustain these advantages for sequentially arriving tasks, a promising direction involves keeping the pre-trained backbone frozen while employing parameter-efficient tuning (PET) techniques to instruct representation learning. Despite the popularity of Prompt-based PET for CL, its empirical design often leads to sub-optimal performance in our evaluation of different PTMs and target tasks. To this end, we propose a unified framework for CL with PTMs and PET that provides both theoretical and empirical advancements. We first perform an in-depth theoretical analysis of the CL objective in a pre-training context, decomposing it into hierarchical components namely within-task prediction, task-identity inference and task-adaptive prediction. We then present Hierarchical Decomposition PET (HiDe-PET), an innovative approach that explicitly optimizes the decomposed objective through incorporating task-specific and task-shared knowledge via mainstream PET techniques along with efficient recovery of pre-trained representations. Leveraging this framework, we delve into the distinct impacts of implementation strategy, PET technique and PET architecture, as well as adaptive knowledge accumulation amidst pronounced distribution changes. Finally, across various CL scenarios, our approach demonstrates remarkably superior performance over a broad spectrum of recent strong baselines.
- Top Cited PapersOpen Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 24914-24937
Chain-of-Reasoning: Towards Unified Mathematical Reasoning in Large Language Models via a Multi-Paradigm Perspective
Yiyao Yu, Yuxiang Zhang, Dongdong Zhang, Xiao Liang, Hengyuan Zhang, Xingxing Zhang, Mahmoud Khademi, Hany Hassan Awadalla, Junjie Wang, Yujiu Yang, Furu Wei. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 24914-24937
Top Cited PapersOpen AccessChain-of-Reasoning: Towards Unified Mathematical Reasoning in Large Language Models via a Multi-Paradigm Perspective
Yiyao Yu, Yuxiang Zhang, Dongdong Zhang, Xiao Liang, Hengyuan Zhang, Xingxing Zhang, Mahmoud Khademi, Hany Hassan Awadalla, Junjie Wang, Yujiu Yang, Furu Wei. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 10966-10975
One-shot In-context Part Segmentation
In this paper, we present the One-shot In-context Part Segmentation (OIParts) framework, designed to tackle the challenges of part segmentation by leveraging visual foundation models (VFMs). Existing training-based one-shot part segmentation methods that utilize VFMs encounter difficulties when faced with scenarios where the one-shot image and test image exhibit significant variance in appearance and perspective, or when the object in the test image is partially visible. We argue that training on the one-shot example often leads to overfitting, thereby compromising the model's generalization capability. Our framework offers a novel approach to part segmentation that is training-free, flexible, and data-efficient, requiring only a single in-context example for precise segmentation with superior generalization ability. By thoroughly exploring the complementary strengths of VFMs, specifically DINOv2 and Stable Diffusion, we introduce an adaptive channel selection approach by minimizing the intra-class distance for better exploiting these two features, thereby enhancing the discriminatory power of the extracted features for the fine-grained parts. We have achieved remarkable segmentation performance across diverse object categories. The OIParts framework not only eliminates the need for extensive labeled data but also demonstrates superior generalization ability. Through comprehensive experimentation on three benchmark datasets, we have demonstrated the superiority of our proposed method over existing part segmentation approaches in one-shot settings. Code is available at https://github.com/dai647/OIParts.
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 10966-10975
Open AccessOne-shot In-context Part Segmentation
In this paper, we present the One-shot In-context Part Segmentation (OIParts) framework, designed to tackle the challenges of part segmentation by leveraging visual foundation models (VFMs). Existing training-based one-shot part segmentation methods that utilize VFMs encounter difficulties when faced with scenarios where the one-shot image and test image exhibit significant variance in appearance and perspective, or when the object in the test image is partially visible. We argue that training on the one-shot example often leads to overfitting, thereby compromising the model's generalization capability. Our framework offers a novel approach to part segmentation that is training-free, flexible, and data-efficient, requiring only a single in-context example for precise segmentation with superior generalization ability. By thoroughly exploring the complementary strengths of VFMs, specifically DINOv2 and Stable Diffusion, we introduce an adaptive channel selection approach by minimizing the intra-class distance for better exploiting these two features, thereby enhancing the discriminatory power of the extracted features for the fine-grained parts. We have achieved remarkable segmentation performance across diverse object categories. The OIParts framework not only eliminates the need for extensive labeled data but also demonstrates superior generalization ability. Through comprehensive experimentation on three benchmark datasets, we have demonstrated the superiority of our proposed method over existing part segmentation approaches in one-shot settings. Code is available at https://github.com/dai647/OIParts.
- Open Access
- 25 October 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Recommender Systems
- Vol. 3 (1) , 1-15
Determinantal Point Processes Guided Crowd-wise Mixture-of-Experts for Recommendation in Alipay
Facing the challenges of sparsity and long tail in thousands of Mini-apps recommendation scenarios deployed on Alipay platform, there is a great need for a simple, effective, and easy-to-deploy industrial solution. To address this issue, we follow the strategy of “divide and conquer” and propose a crowd-based recommendation model by using D eterminantal P oint P rocesse s on C rowd-wise M ixture- o f- E xperts (DPPs-CMoE). Specifically, under the guidance of DPPs-based prototypical tags, the user profiling space is sequentially divided into multiple crowds, with each of them taking on a unique latent specificity; Meanwhile, by treating the modeling of crowd specificity as one of multiple tasks, a crowd-wise architecture is adopted to seamlessly unify the multiple expert networks from the overall user space and the gating network from each of independent crowd spaces. The effectiveness of the proposed method has been illustrated in the experimental results on a mini-apps recommendation scenario deployed in Alipay APPs.
- 25 October 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Recommender Systems
- Vol. 3 (1) , 1-15
Open AccessDeterminantal Point Processes Guided Crowd-wise Mixture-of-Experts for Recommendation in Alipay
Facing the challenges of sparsity and long tail in thousands of Mini-apps recommendation scenarios deployed on Alipay platform, there is a great need for a simple, effective, and easy-to-deploy industrial solution. To address this issue, we follow the strategy of “divide and conquer” and propose a crowd-based recommendation model by using D eterminantal P oint P rocesse s on C rowd-wise M ixture- o f- E xperts (DPPs-CMoE). Specifically, under the guidance of DPPs-based prototypical tags, the user profiling space is sequentially divided into multiple crowds, with each of them taking on a unique latent specificity; Meanwhile, by treating the modeling of crowd specificity as one of multiple tasks, a crowd-wise architecture is adopted to seamlessly unify the multiple expert networks from the overall user space and the gating network from each of independent crowd spaces. The effectiveness of the proposed method has been illustrated in the experimental results on a mini-apps recommendation scenario deployed in Alipay APPs.
- 18 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 33 (10577149) , 5370-5381
To Boost Zero-Shot Generalization for Embodied Reasoning With Vision-Language Pre-Training
Recently, there exists an increased research interest in embodied artificial intelligence (EAI), which involves an agent learning to perform a specific task when dynamically interacting with the surrounding 3D environment. Thereinto, a new challenge is that many unseen objects may appear due to the increased number of object categories in 3D scenes. It makes developing models with strong zero-shot generalization ability to new objects necessary. Existing work tries to achieve this goal by providing embodied agents with massive high-quality human annotations closely related to the task to be learned, while it is too costly in practice. Inspired by recent advances in pre-trained models in 2D visual tasks, we attempt to boost zero-shot generalization for embodied reasoning with vision-language pre-training that can encode common sense as general prior knowledge. To further improve its performance on a specific task, we rectify the pre-trained representation through masked scene graph modeling (MSGM) in a self-supervised manner, where the task-specific knowledge is learned from iterative message passing. Our method can improve a variety of representative embodied reasoning tasks by a large margin ( e.g ., over 5.0% w.r.t. answer accuracy on MP3D-EQA dataset that consists of many real-world scenes with a large number of new objects during testing), and achieve the new state-of-the-art performance.
- 18 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 33 (10577149) , 5370-5381
To Boost Zero-Shot Generalization for Embodied Reasoning With Vision-Language Pre-Training
Recently, there exists an increased research interest in embodied artificial intelligence (EAI), which involves an agent learning to perform a specific task when dynamically interacting with the surrounding 3D environment. Thereinto, a new challenge is that many unseen objects may appear due to the increased number of object categories in 3D scenes. It makes developing models with strong zero-shot generalization ability to new objects necessary. Existing work tries to achieve this goal by providing embodied agents with massive high-quality human annotations closely related to the task to be learned, while it is too costly in practice. Inspired by recent advances in pre-trained models in 2D visual tasks, we attempt to boost zero-shot generalization for embodied reasoning with vision-language pre-training that can encode common sense as general prior knowledge. To further improve its performance on a specific task, we rectify the pre-trained representation through masked scene graph modeling (MSGM) in a self-supervised manner, where the task-specific knowledge is learned from iterative message passing. Our method can improve a variety of representative embodied reasoning tasks by a large margin ( e.g ., over 5.0% w.r.t. answer accuracy on MP3D-EQA dataset that consists of many real-world scenes with a large number of new objects during testing), and achieve the new state-of-the-art performance.