no code implementations • ICML 2020 • Quan Hoang, Trung Le, Dinh Phung
We propose a novel gradient-based tractable approach for the Blahut-Arimoto (BA) algorithm to compute the rate-distortion function where the BA algorithm is fully parameterized.
1 code implementation • 21 Mar 2024 • Minh-Tuan Tran, Trung Le, Xuan-May Le, Mehrtash Harandi, Dinh Phung
In this field, Data-Free Knowledge Transfer (DFKT) plays a crucial role in addressing catastrophic forgetting and data privacy problems.
no code implementations • 19 Mar 2024 • Anh Bui, Vy Vo, Tung Pham, Dinh Phung, Trung Le
There has long been plenty of theoretical and empirical evidence supporting the success of ensemble learning.
no code implementations • 18 Mar 2024 • Anh Bui, Khanh Doan, Trung Le, Paul Montague, Tamas Abraham, Dinh Phung
Generative models have demonstrated remarkable potential in generating visually impressive content from textual descriptions.
1 code implementation • 9 Mar 2024 • Cuong Pham, Van-Anh Nguyen, Trung Le, Dinh Phung, Gustavo Carneiro, Thanh-Toan Do
Inspired by the benefits of the frequency domain, we propose a novel module that functions as an attention mechanism in the frequency domain.
1 code implementation • 23 Feb 2024 • Vy Vo, He Zhao, Trung Le, Edwin V. Bonilla, Dinh Phung
Merely filling in missing values with existing imputation methods and subsequently applying structure learning on the complete data is empirical shown to be sub-optimal.
no code implementations • 29 Jan 2024 • Tuan Nguyen, Van Nguyen, Trung Le, He Zhao, Quan Hung Tran, Dinh Phung
Additionally, we propose minimizing class-aware Higher-order Moment Matching (HMM) to align the corresponding class regions on the source and target domains.
no code implementations • 11 Jan 2024 • Jing Wu, Trung Le, Munawar Hayat, Mehrtash Harandi
In this work, we introduce an unlearning algorithm for diffusion models.
no code implementations • 1 Jan 2024 • Parul Gupta, Tuan Nguyen, Abhinav Dhall, Munawar Hayat, Trung Le, Thanh-Toan Do
The task of Visual Relationship Recognition (VRR) aims to identify relationships between two interacting objects in an image and is particularly challenging due to the widely-spread and highly imbalanced distribution of <subject, relation, object> triplets.
no code implementations • 10 Dec 2023 • Khanh Doan, Quyen Tran, Tung Lam Tran, Tuan Nguyen, Dinh Phung, Trung Le
To address this, we propose the Gradient Projection Class-Prototype Conditional Diffusion Model (GPPDM), a GR-based approach for continual learning that enhances image quality in generators and thus reduces the CF in classifiers.
no code implementations • 26 Nov 2023 • Quyen Tran, Lam Tran, Khoat Than, Toan Tran, Dinh Phung, Trung Le
Drawing inspiration from prompt tuning techniques applied to Large Language Models, recent methods based on pre-trained ViT networks have achieved remarkable results in the field of Continual Learning.
no code implementations • 16 Nov 2023 • Ngoc N. Tran, Lam Tran, Hoang Phan, Anh Bui, Tung Pham, Toan Tran, Dinh Phung, Trung Le
Contrastive learning (CL) is a self-supervised training paradigm that allows us to extract meaningful features without any label information.
1 code implementation • NeurIPS 2023 • Lu Mi, Trung Le, Tianxing He, Eli Shlizerman, Uygar Sümbül
This suggests that neuronal activity is a combination of its time-invariant identity and the inputs the neuron receives from the rest of the circuit.
no code implementations • 2 Oct 2023 • Thanh Nguyen-Duc, Trung Le, Roland Bammer, He Zhao, Jianfei Cai, Dinh Phung
Medical semi-supervised segmentation is a technique where a model is trained to segment objects of interest in medical images with limited annotated data.
1 code implementation • 30 Sep 2023 • Minh-Tuan Tran, Trung Le, Xuan-May Le, Mehrtash Harandi, Quan Hung Tran, Dinh Phung
In this paper, we propose a novel Noisy Layer Generation method (NAYER) which relocates the random source from the input to a noisy layer and utilizes the meaningful constant label-text embedding (LTE) as the input.
no code implementations • 29 Sep 2023 • Tuan Truong, Hoang-Phi Nguyen, Tung Pham, Minh-Tuan Tran, Mehrtash Harandi, Dinh Phung, Trung Le
Motivated by this analysis, we introduce our algorithm, Riemannian Sharpness-Aware Minimization (RSAM).
1 code implementation • NeurIPS 2023 • Van-Anh Nguyen, Trung Le, Anh Tuan Bui, Thanh-Toan Do, Dinh Phung
Interestingly, our developed theories allow us to flexibly incorporate the concept of sharpness awareness into training, whether it's a single model, ensemble models, or Bayesian Neural Networks, by considering specific forms of the center model distribution.
1 code implementation • 26 May 2023 • Michael Fu, Trung Le, Van Nguyen, Chakkrit Tantithamthavorn, Dinh Phung
Prior studies found that vulnerabilities across different vulnerable programs may exhibit similar vulnerable scopes, implicitly forming discernible vulnerability patterns that can be learned by DL models through supervised training.
1 code implementation • 25 May 2023 • Vy Vo, Trung Le, Long-Tung Vuong, He Zhao, Edwin Bonilla, Dinh Phung
Estimating the parameters of a probabilistic directed graphical model from incomplete data remains a long-standing challenge.
no code implementations • 17 May 2023 • Ngoc N. Tran, Son Duong, Hoang Phan, Tung Pham, Dinh Phung, Trung Le
Self-supervised learning aims to extract meaningful features from unlabeled data for further downstream tasks.
1 code implementation • 26 Apr 2023 • Anh Bui, Trung Le, He Zhao, Quan Tran, Paul Montague, Dinh Phung
The key factor for the success of adversarial training is the capability to generate qualified and divergent adversarial examples which satisfy some objectives/goals (e. g., finding adversarial examples that maximize the model losses for simultaneously attacking multiple models).
no code implementations • 21 Apr 2023 • Pengfei Fang, Mehrtash Harandi, Trung Le, Dinh Phung
Hyperbolic geometry, a Riemannian manifold endowed with constant sectional negative curvature, has been considered an alternative embedding space in many learning scenarios, \eg, natural language processing, graph learning, \etc, as a result of its intriguing property of encoding the data's hierarchical structure (like irregular graph or tree-likeness data).
no code implementations • 12 Feb 2023 • Tung-Long Vuong, Trung Le, He Zhao, Chuanxia Zheng, Mehrtash Harandi, Jianfei Cai, Dinh Phung
Learning deep discrete latent presentations offers a promise of better symbolic and summarized abstractions that are more useful to subsequent downstream tasks.
no code implementations • 5 Dec 2022 • Ngoc N. Tran, Anh Tuan Bui, Dinh Phung, Trung Le
On the other hand, in order to achieve that, we need to devise even stronger adversarial attacks to challenge these defense models.
no code implementations • 30 Nov 2022 • Quyen Tran, Hoang Phan, Khoat Than, Dinh Phung, Trung Le
To address this issue, in this work, we first propose an online mixture model learning approach based on nice properties of the mature optimal transport theory (OT-MM).
no code implementations • 24 Nov 2022 • Hoang Phan, Lam Tran, Ngoc N. Tran, Nhat Ho, Dinh Phung, Trung Le
Multi-Task Learning (MTL) is a widely-used and powerful learning paradigm for training deep neural networks that allows learning more than one objective by a single backbone.
1 code implementation • 14 Oct 2022 • Van-Anh Nguyen, Khanh Pham Dinh, Long Tung Vuong, Thanh-Toan Do, Quan Hung Tran, Dinh Phung, Trung Le
Our approach departs from the computational process of ViTs with a focus on visualizing the local and global information in input images and the latent feature embeddings at multiple levels.
1 code implementation • 27 Sep 2022 • Vy Vo, Trung Le, Van Nguyen, He Zhao, Edwin Bonilla, Gholamreza Haffari, Dinh Phung
Interpretable machine learning seeks to understand the reasoning process of complex black-box systems that are long notorious for lack of explainability.
1 code implementation • 20 Sep 2022 • Van Nguyen, Trung Le, Chakkrit Tantithamthavorn, John Grundy, Hung Nguyen, Seyit Camtepe, Paul Quirk, Dinh Phung
In this paper we propose a novel end-to-end deep learning-based approach to identify the vulnerability-relevant code statements of a specific function.
1 code implementation • 19 Sep 2022 • Van Nguyen, Trung Le, Chakkrit Tantithamthavorn, John Grundy, Hung Nguyen, Dinh Phung
However, there are still two open and significant issues for SVD in terms of i) learning automatic representations to improve the predictive performance of SVD, and ii) tackling the scarcity of labeled vulnerabilities datasets that conventionally need laborious labeling effort by experts.
1 code implementation • 7 Jul 2022 • Vy Vo, Van Nguyen, Trung Le, Quan Hung Tran, Gholamreza Haffari, Seyit Camtepe, Dinh Phung
A popular attribution-based approach is to exploit local neighborhoods for learning instance-specific explainers in an additive manner.
no code implementations • 9 Jun 2022 • Trung Le, Eli Shlizerman
Modeling neural population dynamics underlying noisy single-trial spiking activities is essential for relating neural observation and behavior.
1 code implementation • 4 Jun 2022 • Hoang Phan, Ngoc Tran, Trung Le, Toan Tran, Nhat Ho, Dinh Phung
Furthermore, when analysing its asymptotic properties, SVGD reduces exactly to a single-objective optimization problem and can be viewed as a probabilistic version of this single-objective optimization problem.
1 code implementation • 1 Mar 2022 • Hoang Phan, Trung Le, Trung Phung, Tuan Anh Bui, Nhat Ho, Dinh Phung
First, they purely focus on local regularization to strengthen model robustness, missing a global regularization effect which is useful in many real-world applications (e. g., domain adaptation, domain generalization, and adversarial machine learning).
1 code implementation • ICLR 2022 • Tuan Anh Bui, Trung Le, Quan Tran, He Zhao, Dinh Phung
We introduce a new Wasserstein cost function and a new series of risk functions, with which we show that standard AT methods are special cases of their counterparts in our framework.
no code implementations • NeurIPS 2021 • Trung Phung, Trung Le, Long Vuong, Toan Tran, Anh Tran, Hung Bui, Dinh Phung
Domain adaptation (DA) benefits from the rigorous theoretical works that study its insightful characteristics and various aspects, e. g., learning domain-invariant representations and its trade-off.
no code implementations • 29 Oct 2021 • Trung Le, Dat Do, Tuan Nguyen, Huy Nguyen, Hung Bui, Nhat Ho, Dinh Phung
We study the label shift problem between the source and target domains in general domain adaptation (DA) settings.
no code implementations • 21 Oct 2021 • Ehsan K. Ardestani, Changkyu Kim, Seung Jae Lee, Luoshang Pan, Valmiki Rampersad, Jens Axboe, Banit Agrawal, Fuxun Yu, Ansha Yu, Trung Le, Hector Yuen, Shishir Juluri, Akshat Nanda, Manoj Wodekar, Dheevatsa Mudigere, Krishnakumar Nair, Maxim Naumov, Chris Peterson, Mikhail Smelyanskiy, Vijay Rao
Deep Learning Recommendation Models (DLRM) are widespread, account for a considerable data center footprint, and grow by more than 1. 5x per year.
1 code implementation • 14 Oct 2021 • Van-Anh Nguyen, Dai Quoc Nguyen, Van Nguyen, Trung Le, Quan Hung Tran, Dinh Phung
Identifying vulnerabilities in the source code is essential to protect the software systems from cyber security attacks.
1 code implementation • 1 Oct 2021 • Van-Anh Nguyen, Tuan Nguyen, Trung Le, Quan Hung Tran, Dinh Phung
To address the second challenge, we propose to bridge the gap between the target domain and the mixture of source domains in the latent space via a generator or feature extractor.
no code implementations • 29 Sep 2021 • Siqi Xia, Shijie Liu, Trung Le, Dinh Phung, Sarah Erfani, Benjamin I. P. Rubinstein, Christopher Leckie, Paul Montague
More specifically, by minimizing the WS distance of interest, an adversarial example is pushed toward the cluster of benign examples sharing the same label on the latent space, which helps to strengthen the generalization ability of the classifier on the adversarial examples.
no code implementations • 29 Sep 2021 • Van Nguyen, Trung Le, John C. Grundy, Dinh Phung
Software vulnerabilities existing in a program or function of computer systems have been becoming a serious and crucial concern.
no code implementations • 29 Sep 2021 • Long Tung Vuong, Trung Quoc Phung, Toan Tran, Anh Tuan Tran, Dinh Phung, Trung Le
To achieve a satisfactory generalization performance on prediction tasks in an unseen domain, existing domain generalization (DG) approaches often rely on the strict assumption of fixed domain-invariant features and common hypotheses learned from a set of training domains.
1 code implementation • 13 May 2021 • Trung Le, Ryan Poplin, Fred Bertsch, Andeep Singh Toor, Margaret L. Oh
We introduce a new dataset called SyntheticFur built specifically for machine learning training.
1 code implementation • UAI 2021 • Tuan Nguyen, Trung Le, He Zhao, Quan Hung Tran, Truyen Nguyen, Dinh Phung
To this end, we propose in this paper a novel model for multi-source DA using the theory of optimal transport and imitation learning.
Imitation Learning Multi-Source Unsupervised Domain Adaptation +1
no code implementations • 27 Apr 2021 • Mahmoud Hossam, Trung Le, Michael Papasimeon, Viet Huynh, Dinh Phung
Generating realistic sequences is a central task in many machine learning applications.
no code implementations • 27 Apr 2021 • Mahmoud Hossam, Trung Le, He Zhao, Viet Huynh, Dinh Phung
There has been recently a growing interest in studying adversarial examples on natural language models in the black-box setting.
2 code implementations • 11 Feb 2021 • Khai Nguyen, Dang Nguyen, Quoc Nguyen, Tung Pham, Hung Bui, Dinh Phung, Trung Le, Nhat Ho
To address these problems, we propose a novel mini-batch scheme for optimal transport, named Batch of Mini-batches Optimal Transport (BoMb-OT), that finds the optimal coupling between mini-batches and it can be seen as an approximation to a well-defined distance on the space of probability measures.
1 code implementation • 25 Jan 2021 • Anh Bui, Trung Le, He Zhao, Paul Montague, Seyit Camtepe, Dinh Phung
Central to this approach is the selection of positive (similar) and negative (dissimilar) sets to provide the model the opportunity to `contrast' between data and class representation in the latent space.
1 code implementation • ICCV 2021 • Van-Anh Nguyen, Tuan Nguyen, Trung Le, Quan Hung Tran, Dinh Phung
To address the second challenge, we propose to bridge the gap between the target domain and the mixture of source domains in the latent space via a generator or feature extractor.
Multi-Source Unsupervised Domain Adaptation Unsupervised Domain Adaptation
no code implementations • 4 Dec 2020 • Ngan Le, Trung Le, Kashu Yamazaki, Toan Duc Bui, Khoa Luu, Marios Savides
Our proposed Offset Curves (OsC) loss consists of three main fitting terms.
no code implementations • COLING 2020 • Quan Tran, Nhan Dam, Tuan Lai, Franck Dernoncourt, Trung Le, Nham Le, Dinh Phung
Interpretability and explainability of deep neural networks are challenging due to their scale, complexity, and the agreeable notions on which the explaining process rests.
1 code implementation • 14 Oct 2020 • Mahmoud Hossam, Trung Le, He Zhao, Dinh Phung
Training robust deep learning models for down-stream tasks is a critical challenge.
no code implementations • 13 Oct 2020 • He Zhao, Thanh Nguyen, Trung Le, Paul Montague, Olivier De Vel, Tamas Abraham, Dinh Phung
Deep neural network image classifiers are reported to be susceptible to adversarial evasion attacks, which use carefully crafted images created to mislead a classifier.
1 code implementation • 21 Sep 2020 • Anh Bui, Trung Le, He Zhao, Paul Montague, Olivier deVel, Tamas Abraham, Dinh Phung
An important technique of this approach is to control the transferability of adversarial examples among ensemble members.
1 code implementation • ICLR 2021 • He Zhao, Dinh Phung, Viet Huynh, Trung Le, Wray Buntine
Recently, Neural Topic Models (NTMs) inspired by variational autoencoders have obtained increasingly research interest due to their promising results on text analysis.
Ranked #5 on Topic Models on 20NewsGroups
1 code implementation • ECCV 2020 • Anh Bui, Trung Le, He Zhao, Paul Montague, Olivier deVel, Tamas Abraham, Dinh Phung
The fact that deep neural networks are susceptible to crafted perturbations severely impacts the use of deep learning in certain domains of application.
1 code implementation • 16 Apr 2020 • Mahmoud Hossam, Trung Le, Viet Huynh, Michael Papasimeon, Dinh Phung
One of the challenging problems in sequence generation tasks is the optimized generation of sequences with specific desired goals.
no code implementations • 3 Oct 2019 • He Zhao, Trung Le, Paul Montague, Olivier De Vel, Tamas Abraham, Dinh Phung
Deep neural network image classifiers are reported to be susceptible to adversarial evasion attacks, which use carefully crafted images created to mislead a classifier.
no code implementations • ICLR 2019 • Tue Le, Tuan Nguyen, Trung Le, Dinh Phung, Paul Montague, Olivier De Vel, Lizhen Qu
Due to the sharp increase in the severity of the threat imposed by software vulnerabilities, the detection of vulnerabilities in binary code has become an important concern in the software industry, such as the embedded systems industry, and in the field of computer security.
no code implementations • 25 Jan 2019 • Trung Le, Dinh Phung
Previous work has questioned the conditions under which the decision regions of a neural network are connected and further showed the implications of the corresponding theory to the problem of adversarial manipulation of classifiers.
no code implementations • 15 Nov 2018 • Trung Le, Khanh Nguyen, Nhat Ho, Hung Bui, Dinh Phung
The underlying idea of deep domain adaptation is to bridge the gap between source and target domains in a joint space so that a supervised classifier trained on labeled source data can be nicely transferred to the target domain.
1 code implementation • ICLR 2018 • Quan Hoang, Tu Dinh Nguyen, Trung Le, Dinh Phung
We propose in this paper a new approach to train the Generative Adversarial Nets (GANs) with a mixture of generators to overcome the mode collapsing problem.
no code implementations • 6 Nov 2017 • Trung Le, Tu Dinh Nguyen, Dinh Phung
In this paper, we propose a new viewpoint for GANs, which is termed as the minimizing general loss viewpoint.
no code implementations • 19 Sep 2017 • Tung Pham, Trung Le, Hang Dang
In this paper, we propose applying Stochastic Gradient Descent (SGD) framework to the first phase of support-based clustering for finding the domain of novelty and a new strategy to perform the clustering assignment.
no code implementations • 19 Sep 2017 • Trung Le, Khanh Nguyen, Tu Dinh Nguyen, Dinh Phung
With this spirit, in this paper, we propose Analogical-based Bayesian Optimization that can maximize black-box function over a domain where only a similarity score can be defined.
2 code implementations • NeurIPS 2017 • Tu Dinh Nguyen, Trung Le, Hung Vu, Dinh Phung
We develop theoretical analysis to show that, given the maximal discriminators, optimizing the generator of D2GAN reduces to minimizing both KL and reverse KL divergences between data distribution and the distribution induced from the data generated by the generator, hence effectively avoiding the mode collapsing problem.
Ranked #18 on Image Generation on STL-10 (Inception score metric)
no code implementations • 16 Aug 2017 • Trung Le, Hung Vu, Tu Dinh Nguyen, Dinh Phung
Training model to generate data has increasingly attracted research attention and become important in modern world applications.
no code implementations • 8 Aug 2017 • Quan Hoang, Tu Dinh Nguyen, Trung Le, Dinh Phung
A minimax formulation is able to establish among a classifier, a discriminator, and a set of generators in a similar spirit with GAN.
no code implementations • NeurIPS 2016 • Trung Le, Tu Nguyen, Vu Nguyen, Dinh Phung
However, this approach still suffers from a serious shortcoming as it needs to use a high dimensional random feature space to achieve a sufficiently accurate kernel approximation.
no code implementations • 22 Jun 2016 • Trung Le, Khanh Nguyen, Van Nguyen, Vu Nguyen, Dinh Phung
Acquiring labels are often costly, whereas unlabeled data are usually easy to obtain in modern machine learning applications.
1 code implementation • 22 Apr 2016 • Trung Le, Tu Dinh Nguyen, Vu Nguyen, Dinh Phung
One of the most challenging problems in kernel online learning is to bound the model size and to promote the model sparsity.