no code implementations • EMNLP 2021 • Arjun Akula, Spandana Gella, Keze Wang, Song-Chun Zhu, Siva Reddy
Our model outperforms the state-of-the-art NMN model on CLEVR-Ref+ dataset with +8. 1% improvement in accuracy on the single-referent test set and +4. 3% on the full test set.
no code implementations • 24 Feb 2025 • Ziyi Tang, Zechuan Chen, Jiarui Yang, Jiayao Mai, Yongsen Zheng, Keze Wang, Jinrui Chen, Liang Lin
Alpha mining, a critical component in quantitative investment, focuses on discovering predictive signals for future asset returns in increasingly complex financial markets.
no code implementations • 11 Feb 2025 • Jusheng Zhang, Zimeng Huang, Yijia Fan, Ningyuan Liu, Mingyan Li, Zhuojie Yang, Jiawei Yao, Jian Wang, Keze Wang
As scaling large language models faces prohibitive costs, multi-agent systems emerge as a promising alternative, though challenged by static knowledge assumptions and coordination inefficiencies.
no code implementations • 9 Feb 2025 • Jusheng Zhang, Yijia Fan, Kaitong Cai, Keze Wang
Although Kolmogorov-Arnold based interpretable networks (KAN) have strong theoretical expressiveness, they face significant parameter explosion and high-frequency feature capture challenges in high-dimensional tasks.
1 code implementation • 20 Jan 2025 • Wentao Wan, Zhuojie Yang, Yongcan Chen, Chenglin Luo, Ruilin Wang, Kehao Cai, Nan Kang, Liang Lin, Keze Wang
Finally, it guides LLMs to use the previously generated major and minor premises to perform syllogistic deductive reasoning to derive the answer to the original question.
no code implementations • 9 Dec 2024 • Haijing Liu, Tao Pu, Hefeng Wu, Keze Wang, Liang Lin
The proposed framework consists of two complementary modules, i. e., intra-category semantic refinement (ISR) module and inter-category semantic transfer (IST) module.
no code implementations • 21 Nov 2024 • Zeqing Wang, Qingyang Ma, Wentao Wan, Haojie Li, Keze Wang, Yonghong Tian
Intuitively, Visual Language Models (VLMs) that have obtained remarkable performance on various visual tasks are quite suitable for this task.
no code implementations • 8 Aug 2024 • Hefeng Wu, Hao Jiang, Keze Wang, Ziyi Tang, Xianghuan He, Liang Lin
The pursuit of greater interpretability in neural networks often results in a degradation of their original performance.
1 code implementation • 2 Mar 2024 • Linsheng Chen, Guangrun Wang, Liuchun Yuan, Keze Wang, Ken Deng, Philip H. S. Torr
Furthermore, the cascading learning of NeRF-VPT introduces adaptability to scenarios with sparse inputs, resulting in a significant enhancement of accuracy for sparse-view novel view synthesis.
1 code implementation • CVPR 2024 • Xingyu Zhou, Leheng Zhang, Xiaorui Zhao, Keze Wang, Leida Li, Shuhang Gu
The core of MIA-VSR is leveraging feature-level temporal continuity between adjacent frames to reduce redundant computations and make more rational use of previously enhanced SR features.
Ranked #4 on
Video Super-Resolution
on Vid4 - 4x upscaling
no code implementations • 18 Dec 2023 • Hui Fu, Zeqing Wang, Ke Gong, Keze Wang, Tianshui Chen, Haojie Li, Haifeng Zeng, Wenxiong Kang
Moreover, to facilitate disentangled representation learning, we introduce four well-designed constraints: an auxiliary style classifier, an auxiliary inverse classifier, a content contrastive loss, and a pair of latent cycle losses, which can effectively contribute to the construction of the identity-related style space and semantic-related content space.
no code implementations • 29 Nov 2023 • Zeqing Wang, Wentao Wan, Qiqing Lao, Runmeng Chen, Minjie Lang, Keze Wang, Liang Lin
Attempt to overcome this limitation and inspired by the human top-down reasoning process, i. e., systematically exploring relevant issues to derive a comprehensive answer, this work introduces a novel, explainable multi-agent collaboration framework by leveraging the expansive knowledge of Large Language Models (LLMs) to enhance the capabilities of VLMs themselves.
1 code implementation • 16 Nov 2023 • Hefeng Wu, Yandong Chen, Lingbo Liu, Tianshui Chen, Keze Wang, Liang Lin
In the localization stage, the Scale-aware Multi-head Localization (SAML) module utilizes the query tensor to predict the confidence, location, and size of each potential object.
no code implementations • 18 Sep 2023 • Wentao Wan, Nan Kang, Zeqing Wang, Zhuojie Yang, Liang Lin, Keze Wang
Specifically, our CLVP distills the capabilities of well-trained task-specific models into the visual sub-modules in a stepwise and anti-forgetting manner.
2 code implementations • 23 Aug 2023 • Ziyi Tang, Ruilin Wang, Weixing Chen, Yongsen Zheng, Zechuan Chen, Yang Liu, Keze Wang, Tianshui Chen, Liang Lin
Drawing inspiration from the orchestration of diverse specialized agents collaborating to tackle intricate tasks, we propose a framework named Causal-Consistency Chain-of-Thought (CaCo-CoT) that harnesses multi-agent collaboration to bolster the faithfulness and causality of foundation models, involving a set of reasoners and evaluators.
2 code implementations • 7 Dec 2021 • Yang Liu, Keze Wang, Lingbo Liu, Haoyuan Lan, Liang Lin
To overcome these limitations, we take advantage of the multi-scale temporal dependencies within videos and proposes a novel video self-supervised learning framework named Temporal Contrastive Graph Learning (TCGL), which jointly models the inter-snippet and intra-snippet temporal dependencies for temporal representation learning with a hybrid graph contrastive learning strategy.
no code implementations • 8 Nov 2021 • Junying Huang, Fan Chen, Keze Wang, Liang Lin, Dongyu Zhang
Aiming at recognizing the samples from novel categories with few reference samples, few-shot learning (FSL) is a challenging problem.
1 code implementation • 3 Sep 2021 • Arjun R. Akula, Keze Wang, Changsong Liu, Sari Saba-Sadiya, Hongjing Lu, Sinisa Todorovic, Joyce Chai, Song-Chun Zhu
More concretely, our CX-ToM framework generates sequence of explanations in a dialog by mediating the differences between the minds of machine and human user.
1 code implementation • ICCV 2021 • Guangrun Wang, Keze Wang, Guangcong Wang, Philip H. S. Torr, Liang Lin
In this paper, we reveal two contradictory phenomena in contrastive learning that we call under-clustering and over-clustering problems, which are major obstacles to learning efficiency.
Ranked #1 on
Self-Supervised Person Re-Identification
on SYSU-30k
no code implementations • 4 Jan 2021 • Yang Liu, Keze Wang, Haoyuan Lan, Liang Lin
To model multi-scale temporal dependencies, our TCGL integrates the prior knowledge about the frame and snippet orders into graph structures, i. e., the intra-/inter- snippet temporal contrastive graphs.
no code implementations • ICCV 2021 • Qingxing Cao, Wentao Wan, Keze Wang, Xiaodan Liang, Liang Lin
The experimental results show that our proposed method can improve current VQA models on OOD split without losing performance on the in-domain test data.
1 code implementation • 14 Dec 2020 • Qingxing Cao, Bailin Li, Xiaodan Liang, Keze Wang, Liang Lin
Specifically, we generate the question-answer pair based on both the Visual Genome scene graph and an external knowledge base with controlled programs to disentangle the knowledge from other biases.
1 code implementation • 30 Nov 2020 • Junfan Lin, Zhongzhan Huang, Keze Wang, Xiaodan Liang, Weiwei Chen, Liang Lin
Although deep reinforcement learning (RL) has been successfully applied to a variety of robotic control tasks, it's still challenging to apply it to real-world tasks, due to the poor sample efficiency.
1 code implementation • 1 Sep 2020 • Yang Liu, Keze Wang, Guanbin Li, Liang Lin
In this paper, we propose a novel framework, named Semantics-aware Adaptive Knowledge Distillation Networks (SAKDN), to enhance action recognition in vision-sensor modality (videos) by adaptively transferring and distilling the knowledge from multiple wearable sensors.
no code implementations • 23 Mar 2020 • Qingxing Cao, Xiaodan Liang, Keze Wang, Liang Lin
Inspired by the property of a capsule network that can carve a tree structure inside a regular convolutional neural network (CNN), we propose a hierarchical compositional reasoning model called the "Linguistically driven Graph Capsule Network", where the compositional process is guided by the linguistic parse tree.
1 code implementation • 14 Mar 2020 • Junfan Lin, Keze Wang, Ziliang Chen, Xiaodan Liang, Liang Lin
To eliminate this bias and inspired by the propensity score matching technique with causal diagram, we propose a propensity-based patient simulator to effectively answer unrecorded inquiry by drawing knowledge from the other records; Bias (ii) inherently comes along with the passively collected data, and is one of the key obstacles for training the agent towards "learning how" rather than "remembering what".
no code implementations • 29 May 2019 • Hefeng Wu, Yafei Hu, Keze Wang, Hanhui Li, Lin Nie, Hui Cheng
Multi-Person Tracking (MPT) is often addressed within the detection-to-association paradigm.
no code implementations • 4 May 2019 • Yukai Shi, Guanbin Li, Qingxing Cao, Keze Wang, Liang Lin
Face hallucination is a domain-specific super-resolution problem that aims to generate a high-resolution (HR) face image from a low-resolution~(LR) input.
1 code implementation • CVPR 2019 • Guangrun Wang, Keze Wang, Liang Lin
This paper presents a novel adaptively connected neural network (ACNet) to improve the traditional convolutional neural networks (CNNs) {in} two aspects.
Ranked #1 on
Document Classification
on Cora
2 code implementations • arXiv.org 2019 • Keze Wang, Liang Lin, Chenhan Jiang, Chen Qian, Pengxu Wei
Driven by recent computer vision and robotic applications, recovering 3D human poses has become increasingly important and attracted growing interests.
Ranked #286 on
3D Human Pose Estimation
on Human3.6M
no code implementations • 27 Sep 2018 • Ziliang Chen, Keze Wang, Liang Lin
We evaluate T2T across different learners, teachers, and tasks, which significantly demonstrates that structured knowledge can be inherited by the teachers to further benefit learners' training.
1 code implementation • 30 Jun 2018 • Keze Wang, Liang Lin, Xiaopeng Yan, Ziliang Chen, Dongyu Zhang, Lei Zhang
The proposed process can be compatible with mini-batch based training (i. e., using a batch of unlabeled or partially labeled data as a one-time input) for object detection.
no code implementations • CVPR 2018 • Guanbin Li, Yuan Xie, Tianhao Wei, Keze Wang, Liang Lin
Image saliency detection has recently witnessed significant progress due to deep convolutional neural networks.
Ranked #2 on
Video Salient Object Detection
on DAVSOD-Difficult20
(using extra training data)
no code implementations • CVPR 2018 • Keze Wang, Xiaopeng Yan, Dongyu Zhang, Lei Zhang, Liang Lin
Though quite challenging, leveraging large-scale unlabeled or partially labeled images in a cost-effective way has increasingly attracted interests for its great importance to computer vision.
no code implementations • CVPR 2017 • Mude Lin, Liang Lin, Xiaodan Liang, Keze Wang, Hui Cheng
3D human articulated pose recovery from monocular image sequences is very challenging due to the diverse appearances, viewpoints, occlusions, and also the human 3D pose is inherently ambiguous from the monocular imagery.
Ranked #22 on
3D Human Pose Estimation
on HumanEva-I
no code implementations • 28 Jul 2017 • Ziliang Chen, Keze Wang, Xiao Wang, Pai Peng, Ebroul Izquierdo, Liang Lin
Aiming at improving performance of visual classification in a cost-effective manner, this paper proposes an incremental semi-supervised learning paradigm called Deep Co-Space (DCS).
no code implementations • 26 Jul 2017 • Yukai Shi, Keze Wang, Chongyu Chen, Li Xu, Liang Lin
Single image super resolution (SR), which refers to reconstruct a higher-resolution (HR) image from the observed low-resolution (LR) image, has received substantial attention due to its tremendous application potentials.
no code implementations • 13 Jan 2017 • Liang Lin, Keze Wang, Deyu Meng, WangMeng Zuo, Lei Zhang
By naturally combining two recently rising techniques: active learning (AL) and self-paced learning (SPL), our framework is capable of automatically annotating new instances and incorporating them into training under weak expert re-certification.
4 code implementations • 13 Jan 2017 • Keze Wang, Dongyu Zhang, Ya Li, Ruimao Zhang, Liang Lin
In this paper, we propose a novel active learning framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner.
no code implementations • 13 Aug 2016 • Keze Wang, Shengfu Zhai, Hui Cheng, Xiaodan Liang, Liang Lin
In this paper, we propose a novel inference-embedded multi-task learning framework for predicting human pose from still depth images, which is implemented with a deep architecture of neural networks.
no code implementations • 25 Jul 2016 • Yukai Shi, Keze Wang, Li Xu, Liang Lin
Recently, machine learning based single image super resolution (SR) approaches focus on jointly learning representations for high-resolution (HR) and low-resolution (LR) image patch pairs to improve the quality of the super-resolved images.
no code implementations • CVPR 2016 • Keze Wang, Liang Lin, WangMeng Zuo, Shuhang Gu, Lei Zhang
Feature representation and object category classification are two key components of most object detection methods.
no code implementations • 5 Dec 2015 • Liang Lin, Keze Wang, WangMeng Zuo, Meng Wang, Jiebo Luo, Lei Zhang
Understanding human activity is very challenging even with the recently developed 3D/depth sensors.
no code implementations • CVPR 2013 • Keze Wang, Liang Lin, Jiangbo Lu, Chenglong Li, Keyang Shi
In this paper, we propose a unified framework called PISA, which stands for Pixelwise Image Saliency Aggregating various bottom-up cues and priors.
no code implementations • 26 Jan 2015 • Keze Wang, Xiaolong Wang, Liang Lin, Meng Wang, WangMeng Zuo
Our model thus advances existing approaches in two aspects: (i) it acts directly on the raw inputs (grayscale-depth data) to conduct recognition instead of relying on hand-crafted features, and (ii) the model structure can be dynamically adjusted accounting for the temporal variations of human activities, i. e. the network configuration is allowed to be partially activated during inference.
no code implementations • CVPR 2013 • Keyang Shi, Keze Wang, Jiangbo Lu, Liang Lin
By fusing complementary contrast measures in such a pixelwise adaptive manner, the detection effectiveness is significantly boosted.