Dinh Phung
YOU?
Author Swipe
View article: G-reasoner: Foundation Models for Unified Reasoning over Graph-structured Knowledge
G-reasoner: Foundation Models for Unified Reasoning over Graph-structured Knowledge Open
Large language models (LLMs) excel at complex reasoning but remain limited by static and incomplete parametric knowledge. Retrieval-augmented generation (RAG) mitigates this by incorporating external knowledge, yet existing RAGs struggle w…
View article: Unveiling m-Sharpness Through the Structure of Stochastic Gradient Noise
Unveiling m-Sharpness Through the Structure of Stochastic Gradient Noise Open
Sharpness-aware minimization (SAM) has emerged as a highly effective technique for improving model generalization, but its underlying principles are not fully understood. We investigated the phenomenon known as m-sharpness, where the perfo…
View article: IMU: Influence-guided Machine Unlearning
IMU: Influence-guided Machine Unlearning Open
Recent studies have shown that deep learning models are vulnerable to attacks and tend to memorize training data points, raising significant concerns about privacy leakage. This motivates the development of machine unlearning (MU), i.e., a…
View article: Mitigating Semantic Collapse in Generative Personalization with Test-Time Embedding Adjustment
Mitigating Semantic Collapse in Generative Personalization with Test-Time Embedding Adjustment Open
In this paper, we investigate the semantic collapsing problem in generative personalization, an under-explored topic where the learned visual concept ($V$) gradually shifts from its original textual meaning and comes to dominate other conc…
View article: Preserving Clusters in Prompt Learning for Unsupervised Domain Adaptation
Preserving Clusters in Prompt Learning for Unsupervised Domain Adaptation Open
Recent approaches leveraging multi-modal pre-trained models like CLIP for Unsupervised Domain Adaptation (UDA) have shown significant promise in bridging domain gaps and improving generalization by utilizing rich semantic knowledge and rob…
View article: Optimizing Specific and Shared Parameters for Efficient Parameter Tuning
Optimizing Specific and Shared Parameters for Efficient Parameter Tuning Open
Foundation models, with a vast number of parameters and pretraining on massive datasets, achieve state-of-the-art performance across various applications. However, efficiently adapting them to downstream tasks with minimal computational ov…
View article: Why Domain Generalization Fail? A View of Necessity and Sufficiency
Why Domain Generalization Fail? A View of Necessity and Sufficiency Open
Despite a strong theoretical foundation, empirical experiments reveal that existing domain generalization (DG) algorithms often fail to consistently outperform the ERM baseline. We argue that this issue arises because most DG studies focus…
View article: Unbiased Sliced Wasserstein Kernels for High-Quality Audio Captioning
Unbiased Sliced Wasserstein Kernels for High-Quality Audio Captioning Open
Teacher-forcing training for audio captioning usually leads to exposure bias due to training and inference mismatch. Prior works propose the contrastive method to deal with caption degeneration. However, the contrastive method ignores the …
View article: GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation
GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation Open
Retrieval-augmented generation (RAG) has proven effective in integrating knowledge into large language models (LLMs). However, conventional RAGs struggle to capture complex relationships between pieces of knowledge, limiting their performa…
View article: Fantastic Targets for Concept Erasure in Diffusion Models and Where To Find Them
Fantastic Targets for Concept Erasure in Diffusion Models and Where To Find Them Open
Concept erasure has emerged as a promising technique for mitigating the risk of harmful content generation in diffusion models by selectively unlearning undesirable concepts. The common principle of previous works to remove a specific conc…
View article: Explicit Eigenvalue Regularization Improves Sharpness-Aware Minimization
Explicit Eigenvalue Regularization Improves Sharpness-Aware Minimization Open
Sharpness-Aware Minimization (SAM) has attracted significant attention for its effectiveness in improving generalization across various tasks. However, its underlying principles remain poorly understood. In this work, we analyze SAM's trai…
View article: Planning for Success: Exploring LLM Long-term Planning Capabilities in Table Understanding
Planning for Success: Exploring LLM Long-term Planning Capabilities in Table Understanding Open
View article: LLM Reading Tea Leaves: Automatically Evaluating Topic Models with Large Language Models
LLM Reading Tea Leaves: Automatically Evaluating Topic Models with Large Language Models Open
Topic modeling has been a widely used tool for unsupervised text analysis. However, comprehensive evaluations of a topic model remain challenging. Existing evaluation methods are either less comparable across different models (e.g., perple…
View article: Neural Topic Modeling with Large Language Models in the Loop
Neural Topic Modeling with Large Language Models in the Loop Open
View article: PanSplat: 4K Panorama Synthesis with Feed-Forward Gaussian Splatting
PanSplat: 4K Panorama Synthesis with Feed-Forward Gaussian Splatting Open
With the advent of portable 360° cameras, panorama has gained significant attention in applications like virtual reality (VR), virtual tours, robotics, and autonomous driving. As a result, wide-baseline panorama view synthesis has emerged …
View article: Large-Scale Data-Free Knowledge Distillation for ImageNet via Multi-Resolution Data Generation
Large-Scale Data-Free Knowledge Distillation for ImageNet via Multi-Resolution Data Generation Open
Data-Free Knowledge Distillation (DFKD) is an advanced technique that enables knowledge transfer from a teacher model to a student model without relying on original training data. While DFKD methods have achieved success on smaller dataset…
View article: Neural Topic Modeling with Large Language Models in the Loop
Neural Topic Modeling with Large Language Models in the Loop Open
Topic modeling is a fundamental task in natural language processing, allowing the discovery of latent thematic structures in text corpora. While Large Language Models (LLMs) have demonstrated promising capabilities in topic discovery, thei…
View article: Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation
Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation Open
Diffusion models excel at generating visually striking content from text but can inadvertently produce undesirable or harmful content when trained on unfiltered internet data. A practical solution is to selectively removing target concepts…
View article: Leveraging Hierarchical Taxonomies in Prompt-based Continual Learning
Leveraging Hierarchical Taxonomies in Prompt-based Continual Learning Open
Humans perceive the world as a series of sequential events, which can be hierarchically organized with different levels of abstraction based on conceptual knowledge. Drawing inspiration from human learning behaviors, this work proposes a n…
View article: Improving Generalization with Flat Hilbert Bayesian Inference
Improving Generalization with Flat Hilbert Bayesian Inference Open
We introduce Flat Hilbert Bayesian Inference (FHBI), an algorithm designed to enhance generalization in Bayesian inference. Our approach involves an iterative two-step procedure with an adversarial functional perturbation step and a functi…
View article: Connective Viewpoints of Signal-to-Noise Diffusion Models
Connective Viewpoints of Signal-to-Noise Diffusion Models Open
Diffusion models (DM) have become fundamental components of generative models, excelling across various domains such as image creation, audio generation, and complex data interpolation. Signal-to-Noise diffusion models constitute a diverse…
View article: MetaAug: Meta-Data Augmentation for Post-Training Quantization
MetaAug: Meta-Data Augmentation for Post-Training Quantization Open
Post-Training Quantization (PTQ) has received significant attention because it requires only a small set of calibration data to quantize a full-precision model, which is more practical in real-world applications in which full access to a l…
View article: Model and Feature Diversity for Bayesian Neural Networks in Mutual Learning
Model and Feature Diversity for Bayesian Neural Networks in Mutual Learning Open
Bayesian Neural Networks (BNNs) offer probability distributions for model parameters, enabling uncertainty quantification in predictions. However, they often underperform compared to deterministic neural networks. Utilizing mutual learning…
View article: LLM Reading Tea Leaves: Automatically Evaluating Topic Models with Large Language Models
LLM Reading Tea Leaves: Automatically Evaluating Topic Models with Large Language Models Open
Topic modeling has been a widely used tool for unsupervised text analysis. However, comprehensive evaluations of a topic model remain challenging. Existing evaluation methods are either less comparable across different models (e.g., perple…
View article: Agnostic Sharpness-Aware Minimization
Agnostic Sharpness-Aware Minimization Open
Sharpness-aware minimization (SAM) has been instrumental in improving deep neural network training by minimizing both the training loss and the sharpness of the loss landscape, leading the model into flatter minima that are associated with…
View article: PaRa: Personalizing Text-to-Image Diffusion via Parameter Rank Reduction
PaRa: Personalizing Text-to-Image Diffusion via Parameter Rank Reduction Open
Personalizing a large-scale pretrained Text-to-Image (T2I) diffusion model is challenging as it typically struggles to make an appropriate trade-off between its training data distribution and the target distribution, i.e., learning a novel…
View article: Revisiting Deep Audio-Text Retrieval Through the Lens of Transportation
Revisiting Deep Audio-Text Retrieval Through the Lens of Transportation Open
The Learning-to-match (LTM) framework proves to be an effective inverse optimal transport approach for learning the underlying ground metric between two sources of data, facilitating subsequent matching. However, the conventional LTM frame…
View article: Deep Domain Adaptation With Max-Margin Principle for Cross-Project Imbalanced Software Vulnerability Detection
Deep Domain Adaptation With Max-Margin Principle for Cross-Project Imbalanced Software Vulnerability Detection Open
Software vulnerabilities (SVs) have become a common, serious, and crucial concern due to the ubiquity of computer software. Many AI-based approaches have been proposed to solve the software vulnerability detection (SVD) problem to ensure t…
View article: Taming Stable Diffusion for Text to 360° Panorama Image Generation
Taming Stable Diffusion for Text to 360° Panorama Image Generation Open
Generative models, e.g., Stable Diffusion, have enabled the creation of photorealistic images from text prompts. Yet, the generation of 360-degree panorama images from text remains a challenge, particularly due to the dearth of paired text…
View article: Text-Enhanced Data-free Approach for Federated Class-Incremental Learning
Text-Enhanced Data-free Approach for Federated Class-Incremental Learning Open
Federated Class-Incremental Learning (FCIL) is an underexplored yet pivotal issue, involving the dynamic addition of new classes in the context of federated learning. In this field, Data-Free Knowledge Transfer (DFKT) plays a crucial role …