no code implementations • 16 May 2019 • Yaoyi Li, Jianfu Zhang, Weijie Zhao, Hongtao Lu
A high efficient image matting method based on a weakly annotated mask is in demand for mobile applications.
no code implementations • 24 Nov 2019 • Yi Tu, Li Niu, Weijie Zhao, Dawei Cheng, Liqing Zhang
Aesthetic image cropping is a practical but challenging task which aims at finding the best crops with the highest aesthetic quality in an image.
2 code implementations • 12 Mar 2020 • Weijie Zhao, Deping Xie, Ronglai Jia, Yulei Qian, Ruiquan Ding, Mingming Sun, Ping Li
For example, a sponsored online advertising system can contain more than $10^{11}$ sparse features, making the neural network a massive model with around 10 TB parameters.
1 code implementation • 5 Aug 2020 • Yan Hong, Li Niu, Jianfu Zhang, Weijie Zhao, Chen Fu, Liqing Zhang
In this paper, we propose a Fusing-and-Filling Generative Adversarial Network (F2GAN) to generate realistic and diverse images for a new category with only a few images.
1 code implementation • NeurIPS 2020 • Shaogang Ren, Weijie Zhao, Ping Li
L1 regularization has been broadly employed to pursue model sparsity.
2 code implementations • ICCV 2021 • Khoa Doan, Yingjie Lao, Weijie Zhao, Ping Li
Under this optimization framework, the trigger generator function will learn to manipulate the input with imperceptible noise to preserve the model performance on the clean data and maximize the attack success rate on the poisoned data.
no code implementations • 1 Jan 2021 • Xiangyi Chen, Belhal Karimi, Weijie Zhao, Ping Li
Specifically, we propose a general algorithmic framework that can convert existing adaptive gradient methods to their decentralized counterparts.
no code implementations • 7 Sep 2021 • Xiangyi Chen, Belhal Karimi, Weijie Zhao, Ping Li
Adaptive gradient methods including Adam, AdaGrad, and their variants have been very successful for training deep learning models, such as neural networks.
no code implementations • 29 Sep 2021 • Zhiqi Bu, Ping Li, Weijie Zhao
In this work, we propose the practical adversarial training with differential privacy (DP-Adv), to combine the backbones from both communities and deliver robust and private models with high accuracy.
no code implementations • 5 Jan 2022 • Weijie Zhao, Xuewu Jiao, Mingqing Hu, Xiaoyun Li, Xiangyu Zhang, Ping Li
In this paper, we propose a hardware-aware training workflow that couples the hardware topology into the algorithm design.
no code implementations • 7 Jan 2022 • Ping Li, Weijie Zhao
For example, one can apply GCWS on the outputs of the last layer to boost the accuracy of trained deep neural networks.
1 code implementation • 22 May 2022 • Ping Li, Weijie Zhao
Our framework has parameters $(s, g, w)$.
no code implementations • 22 Jun 2022 • Zhaozhuo Xu, Weijie Zhao, Shulong Tan, Zhixin Zhou, Ping Li
Given a vertex deletion request, we thoroughly investigate solutions to update the connections of the vertex.
no code implementations • Proceedings of the AAAI Conference on Artificial Intelligence 2022 • Yingjie Lao, Weijie Zhao, Peng Yang, Ping Li
After embedding, each model will respond distinctively to these key samples, which creates a model-unique signature as a strong tool for authentication and user identity.
1 code implementation • 18 Jul 2022 • Ping Li, Weijie Zhao
Although the gain formula in Li (2010) was derived for logistic regression loss, it is a generic formula for loss functions with second-derivatives.
1 code implementation • 18 Jul 2022 • Ping Li, Weijie Zhao
In recent prior studies, the pGMM kernel has been extensively evaluated for classification tasks, for logistic regression, support vector machines, as well as deep neural networks.
no code implementations • 26 Sep 2022 • Weijie Zhao, Xuewu Jiao, Xinsheng Luo, Jingxue Li, Belhal Karimi, Ping Li
In this paper, we propose FeatureBox, a novel end-to-end training framework that pipelines the feature extraction and the training on GPU servers to save the intermediate I/O of the feature extraction.
no code implementations • 26 Oct 2022 • Weijie Zhao, Shulong Tan, Ping Li
Typically a three-stage mechanism is employed in those systems: (i) a small collection of items are first retrieved by (e. g.,) approximate near neighbor search algorithms; (ii) then a collection of constraints are applied on the retrieved items; (iii) a fine-grained ranking neural network is employed to determine the final recommendation.
no code implementations • 1 Nov 2022 • Khoa Doan, Shulong Tan, Weijie Zhao, Ping Li
Previous learning-to-hash approaches are also not suitable to solve the fast item ranking problem since they can take a significant amount of time and computation to train the hash functions.
no code implementations • 26 Apr 2023 • Xinyi Zheng, Weijie Zhao, Xiaoyun Li, Ping Li
To retrieve personalized campaigns and creatives while protecting user privacy, digital advertising is shifting from member-based identity to cohort-based identity.
no code implementations • 13 Jun 2023 • Ping Li, Weijie Zhao, Chao Wang, Qi Xia, Alice Wu, Lijun Peng
In this paper, we report our exploration of efficient search in sparse data with graph-based ANN algorithms (e. g., HNSW, or SONG which is the GPU version of HNSW), which are popular in industrial practice, e. g., search and ads (advertising).
no code implementations • 28 Jun 2023 • Ping Li, Weijie Zhao
In this study, we propose to re-use the hashes by partitioning the $B$ bits into $m$ chunks, e. g., $b\times m =B$.
no code implementations • 28 Jun 2023 • Weijie Zhao, Ping Li
Feature interactions can play a crucial role in recommendation systems as they capture complex relationships between user preferences and item characteristics.
1 code implementation • KDD 2023 • Huawei Lin, Jun Woo Chung, Yingjie Lao, Weijie Zhao
To the best of our knowledge, this is the first work that considers machine unlearning on GBDT.
no code implementations • 28 Dec 2023 • Weijie Zhao, Shulong Tan, Ping Li
With the continuous popularity of deep learning and representation learning, fast vector search becomes a vital task in various ranking/retrieval based applications, say recommendation, ads ranking and question answering.