no code implementations • 19 Apr 2024 • Yang Ye, Shihao Ji
This paper proposes GDPNet, the first hybrid Generative and Discriminative PointNet that extends JEM for point cloud classification and generation.
1 code implementation • 25 Aug 2023 • Hui Ye, Rajshekhar Sunderraman, Shihao Ji
We then extract the dense text representations from the fine-tuned Transformer.
Multi Label Text Classification Multi-Label Text Classification +3
1 code implementation • 8 Mar 2023 • Xiulong Yang, Shihao Ji
Despite its simplicity, M-EBM significantly improves unconditional EBMs in training stability and speed on a host of benchmark datasets, such as CIFAR10, CIFAR100, CelebA-HQ, and ImageNet 32x32.
1 code implementation • 11 Nov 2022 • Yang Li, Xin Ma, Raj Sunderraman, Shihao Ji, Suprateek Kundu
We compare the prediction performance for different intelligence measures based on static FC, dynamic FC, and region level time series acquired from the Adolescent Brain Cognitive Development (ABCD) study involving close to 7000 individuals.
1 code implementation • 11 Oct 2022 • Yang Ye, Xiulong Yang, Shihao Ji
Traditional task-agnostic sampling methods, such as farthest point sampling (FPS), do not consider downstream tasks when sampling point clouds, and thus non-informative points to the tasks are often sampled.
1 code implementation • CVPR 2023 • Xiulong Yang, Qing Su, Shihao Ji
This question has recently been answered in the affirmative, introducing the field of Joint Energy-based Model (JEM), which achieves high classification accuracy and image generation quality simultaneously.
1 code implementation • 16 Aug 2022 • Xiulong Yang, Sheng-Min Shih, Yinlin Fu, Xiaoting Zhao, Shihao Ji
Diffusion Denoising Probability Models (DDPM) and Vision Transformer (ViT) have demonstrated significant progress in generative tasks and discriminative tasks, respectively, and thus far these models have largely been developed in their own domains.
1 code implementation • COLING 2022 • Mingchen Li, Shihao Ji
Therefore, this paper focuses on query graph generation from natural language questions.
1 code implementation • 9 Mar 2022 • Qing Su, Shihao Ji
Current stereo matching techniques are challenged by restricted searching space, occluded regions and sheer size.
1 code implementation • CVPR 2022 • Qing Su, Shihao Ji
Current stereo matching techniques are challenged by restricted searching space, occluded regions and sheer size.
Ranked #2 on Stereo Depth Estimation on KITTI2015 (D1-all All metric)
1 code implementation • 8 Nov 2021 • Xiang Li, Shihao Ji
Extensive experiments on VGGFace, Traffic Sign and ImageNet show that GDPA achieves higher attack success rates than state-of-the-art patch attacks, while adversarially trained model with GDPA demonstrates superior robustness to adversarial patch attacks than competing methods.
1 code implementation • ICCV 2021 • Xiulong Yang, Shihao Ji
1) We propose a proximal SGLD to generate samples in the proximity of samples from the previous step, which improves the stability.
1 code implementation • 6 Jul 2021 • Hui Ye, Xiulong Yang, Martin Takac, Rajshekhar Sunderraman, Shihao Ji
To address this issue, we propose a contrastive learning approach to improve the quality and enhance the semantic consistency of synthetic images.
Ranked #9 on Text-to-Image Generation on CUB
1 code implementation • 30 Jun 2021 • Yang Li, Shihao Ji
We term our algorithm Dep-$L_0$ as it prunes networks via a dependency-enabled $L_0$ regularization.
no code implementations • 28 Apr 2021 • Krishanu Sarker, Sharbani Pandit, Anupam Sarker, Saeid Belkasim, Shihao Ji
In this work, we introduce uncertainty estimation to detect confusing cases for expert referral to address the unreliability of state-of-the-art (SOTA) DNNs on COVID-19 detection.
1 code implementation • 1 Jan 2021 • Xiulong Yang, Hui Ye, Yang Ye, Xiang Li, Shihao Ji
We show that our Generative MMC (GMMC) can be trained discriminatively, generatively, or jointly for image classification and generation.
no code implementations • 25 Sep 2020 • Krishanu Sarker, Xiulong Yang, Yang Li, Saeid Belkasim, Shihao Ji
The success of Deep Neural Networks (DNNs) highly depends on data quality.
1 code implementation • 30 Aug 2020 • Kaiyang Li, Guangchun Luo, Yang Ye, Wei Li, Shihao Ji, Zhipeng Cai
In this paper, we propose Adversarial Privacy Graph Embedding (APGE), a graph adversarial training framework that integrates the disentangling and purging mechanisms to remove users' private information from learned node representations.
1 code implementation • 4 Dec 2019 • Xiulong Yang, Shihao Ji
In this paper, we propose xAT and xVAT, new adversarial training algorithms, that generate \textbf{multiplicative} perturbations to input examples for robust training of DNNs.
1 code implementation • 2 Dec 2019 • Yang Ye, Shihao Ji
Among the variants of GNNs, Graph Attention Networks (GATs) learn to assign dense attention coefficients over all neighbors of a node for feature aggregation, and improve the performance of many graph learning tasks.
no code implementations • 20 Nov 2019 • Azim Ahmadzadeh, Sushant S. Mahajan, Dustin J. Kempton, Rafal A. Angryk, Shihao Ji
Despite the known challenges in the identification and characterization of filaments by the existing module, which in turn are inherited into any other module that intends to learn from such outputs, Mask R-CNN shows promising results.
1 code implementation • 13 Aug 2019 • Yang Li, Shihao Ji
To the best of our knowledge, this is the first learning framework that unifies network sparsification and network expansion in an end-to-end training pipeline.
1 code implementation • 9 Aug 2019 • Xiang Li, Shihao Ji
Explaining the prediction of deep neural networks (DNNs) and semantic image compression are two active research areas of deep learning with a numerous of applications in decision-critical systems, such as surveillance cameras, drones and self-driving cars, where interpretable decision is critical and storage/network bandwidth is limited.
1 code implementation • 9 Apr 2019 • Yang Li, Shihao Ji
Thanks to the flexibility of ARM, many smooth or non-smooth parametric functions, such as scaled sigmoid or hard sigmoid, can be used to parameterize this binary optimization problem and the unbiasness of the ARM estimator is retained, while the hard concrete estimator has to rely on the hard sigmoid function to achieve conditional computation and thus accelerated training.
1 code implementation • 17 Dec 2018 • Xiang Li, Shihao Ji
The proposed method is generic and can defend white-box and black-box attacks without the need of retraining the original CNN classifiers, and can further strengthen the defense by retraining CNN or end-to-end finetuning the whole pipeline.
no code implementations • 16 Dec 2018 • Krishanu Sarker, Mohamed Masoud, Saeid Belkasim, Shihao Ji
Due to lack of depth information, RGB video based activity recognition performs poorly compared to RGB-D video based solutions.
1 code implementation • 18 Nov 2016 • Shihao Ji, Nadathur Satish, Sheng Li, Pradeep Dubey
Word2vec is a widely used algorithm for extracting low-dimensional vector representations of words.
no code implementations • 31 May 2016 • Jiong Zhang, Parameswaran Raman, Shihao Ji, Hsiang-Fu Yu, S. V. N. Vishwanathan, Inderjit S. Dhillon
Moreover, it requires the parameters to fit in the memory of a single processor; this is problematic when the number of parameters is in billions.
no code implementations • 15 Apr 2016 • Shihao Ji, Nadathur Satish, Sheng Li, Pradeep Dubey
In combination, these techniques allow us to scale up the computation near linearly across cores and nodes, and process hundreds of millions of words per second, which is the fastest word2vec implementation to the best of our knowledge.
1 code implementation • 21 Nov 2015 • Shihao Ji, S. V. N. Vishwanathan, Nadathur Satish, Michael J. Anderson, Pradeep Dubey
One way to understand BlackOut is to view it as an extension of the DropOut strategy to the output layer, wherein we use a discriminative training loss and a weighted sampling scheme.
2 code implementations • EMNLP 2016 • Shihao Ji, Hyokun Yun, Pinar Yanardag, Shin Matsushima, S. V. N. Vishwanathan
Then, based on this insight, we propose a novel framework WordRank that efficiently estimates word representations via robust ranking, in which the attention mechanism and robustness to noise are readily achieved via the DCG-like ranking losses.