Sun‐Yuan Kung
YOU?
Author Swipe
View article: One Token to Fool LLM-as-a-Judge
One Token to Fool LLM-as-a-Judge Open
Large language models (LLMs) are increasingly trusted as automated judges, assisting evaluation and providing reward signals for training other models, particularly in reference-based settings like Reinforcement Learning with Verifiable Re…
View article: Enhancing Extended Weather Forecasts in the TCWAGFS Model Using Deep Learning Method for SST Bias Correction
Enhancing Extended Weather Forecasts in the TCWAGFS Model Using Deep Learning Method for SST Bias Correction Open
The extended weather (> 7 days) and the seasonal climate predictions are highly dependent on the status of Madden Julian Oscillation (MJO) and El Niño-Southern Oscillation (ENSO). Both the evolutions of MJO and ENSO are found to be correla…
View article: Object Detection Using ESRGAN With a Sequential Transfer Learning on Remote Sensing Embedded Systems
Object Detection Using ESRGAN With a Sequential Transfer Learning on Remote Sensing Embedded Systems Open
The field of remote sensing has experienced rapid advancement owing to the widespread utilization of image sensors, drones, and satellites for data collection. However, object detection in remote sensing poses challenges owing to small obj…
View article: SRODNet: Object Detection Network Based on Super Resolution for Autonomous Vehicles
SRODNet: Object Detection Network Based on Super Resolution for Autonomous Vehicles Open
Object detection methods have been applied in several aerial and traffic surveillance applications. However, object detection accuracy decreases in low-resolution (LR) images owing to feature loss. To address this problem, we propose a sin…
View article: MILAN: Masked Image Pretraining on Language Assisted Representation
MILAN: Masked Image Pretraining on Language Assisted Representation Open
Self-attention based transformer models have been dominating many computer vision tasks in the past few years. Their superb model qualities heavily depend on the excessively large labeled image datasets. In order to reduce the reliance on …
View article: Evolving transferable neural pruning functions
Evolving transferable neural pruning functions Open
Structural design of neural networks is crucial for the success of deep\nlearning. While most prior works in evolutionary learning aim at directly\nsearching the structure of a network, few attempts have been made on another\npromising tra…
View article: XNAS: A Regressive/Progressive NAS for Deep Learning
XNAS: A Regressive/Progressive NAS for Deep Learning Open
Deep learning has achieved great and broad breakthroughs in many real-world applications. In particular, the task of training the network parameters has been masterly handled by back-propagation learning. However, the pursuit on optimal ne…
View article: CHEX: CHannel EXploration for CNN Model Compression
CHEX: CHannel EXploration for CNN Model Compression Open
Channel pruning has been broadly recognized as an effective technique to reduce the computation and memory cost of deep convolutional neural networks. However, conventional pruning methods have limitations in that: they are restricted to p…
View article: Multi-Dimensional Model Compression of Vision Transformer
Multi-Dimensional Model Compression of Vision Transformer Open
Vision transformers (ViT) have recently attracted considerable attentions, but the huge computational cost remains an issue for practical deployment. Previous ViT pruning methods tend to prune the model along one dimension solely, which ma…
View article: Class-Discriminative CNN Compression
Class-Discriminative CNN Compression Open
Compressing convolutional neural networks (CNNs) by pruning and distillation has received ever-increasing focus in the community. In particular, designing a class-discrimination based approach would be desired as it fits seamlessly with th…
View article: Evolving Transferable Pruning Functions
Evolving Transferable Pruning Functions Open
Channel pruning has made major headway in the design of efficient deep learning models. Conventional approaches adopt human-made pruning functions to score channels' importance for channel pruning, which requires domain knowledge and could…
View article: Privacy Enhancing Machine Learning via Removal of Unwanted Dependencies
Privacy Enhancing Machine Learning via Removal of Unwanted Dependencies Open
The rapid rise of IoT and Big Data has facilitated copious data-driven applications to enhance our quality of life. However, the omnipresent and all-encompassing nature of the data collection can generate privacy concerns. Hence, there is …
View article: Few-shot Learning via Dependency Maximization and Instance Discriminant Analysis
Few-shot Learning via Dependency Maximization and Instance Discriminant Analysis Open
We study the few-shot learning (FSL) problem, where a model learns to recognize new objects with extremely few labeled training data per category. Most of previous FSL approaches resort to the meta-learning paradigm, where the model accumu…
View article: A compressive multi-kernel method for privacy-preserving machine learning
A compressive multi-kernel method for privacy-preserving machine learning Open
As the analytic tools become more powerful, and more data are generated on a daily basis, the issue of data privacy arises. This leads to the study of the design of privacy-preserving machine learning algorithms. Given two objectives, name…
View article: CovSegNet: A Multi Encoder–Decoder Architecture for Improved Lesion Segmentation of COVID-19 Chest CT Scans
CovSegNet: A Multi Encoder–Decoder Architecture for Improved Lesion Segmentation of COVID-19 Chest CT Scans Open
Automatic lung lesion segmentation of chest computer tomography (CT) scans is considered a pivotal stage toward accurate diagnosis and severity measurement of COVID-19. Traditional U-shaped encoder-decoder architecture and its variants suf…
View article: Content-Aware GAN Compression
Content-Aware GAN Compression Open
Generative adversarial networks (GANs), e.g., StyleGAN2, play a vital role in various image generation and synthesis tasks, yet their notoriously high computational cost hinders their efficient deployment on edge devices. Directly applying…
View article: CovSegNet: A Multi Encoder-Decoder Architecture for Improved Lesion Segmentation of COVID-19 Chest CT Scans
CovSegNet: A Multi Encoder-Decoder Architecture for Improved Lesion Segmentation of COVID-19 Chest CT Scans Open
Automatic lung lesions segmentation of chest CT scans is considered a pivotal stage towards accurate diagnosis and severity measurement of COVID-19. Traditional U-shaped encoder-decoder architecture and its variants suffer from diminutions…
View article: A Novel Multi-Stage Training Approach for Human Activity Recognition From Multimodal Wearable Sensor Data Using Deep Neural Network
A Novel Multi-Stage Training Approach for Human Activity Recognition From Multimodal Wearable Sensor Data Using Deep Neural Network Open
Deep neural network is an effective choice to automatically recognize human\nactions utilizing data from various wearable sensors. These networks automate\nthe process of feature extraction relying completely on data. However, various\nnoi…
View article: A Feature-map Discriminant Perspective for Pruning Deep Neural Networks
A Feature-map Discriminant Perspective for Pruning Deep Neural Networks Open
Network pruning has become the de facto tool to accelerate deep neural networks for mobile and edge applications. Recently, feature-map discriminant based channel pruning has shown promising results, as it aligns well with the CNN objectiv…
View article: Rethinking Class-Discrimination Based CNN Channel Pruning
Rethinking Class-Discrimination Based CNN Channel Pruning Open
Channel pruning has received ever-increasing focus on network compression. In particular, class-discrimination based channel pruning has made major headway, as it fits seamlessly with the classification objective of CNNs and provides good …
View article: Soft-Root-Sign Activation Function
Soft-Root-Sign Activation Function Open
The choice of activation function in deep networks has a significant effect on the training dynamics and task performance. At present, the most effective and widely-used activation function is ReLU. However, because of the non-zero mean, n…
View article: Exploiting Operation Importance for Differentiable Neural Architecture Search
Exploiting Operation Importance for Differentiable Neural Architecture Search Open
Recently, differentiable neural architecture search methods significantly reduce the search cost by constructing a super network and relax the architecture representation by assigning architecture weights to the candidate operations. All t…
View article: Temporal Action Localization using Long Short-Term Dependency
Temporal Action Localization using Long Short-Term Dependency Open
Temporal action localization in untrimmed videos is an important but difficult task. Difficulties are encountered in the application of existing methods when modeling temporal structures of videos. In the present study, we developed a nove…