Search Results for author:
Found 122 papers, 38 papers with code
Detection of Adversarial Examples in Text Classification: Benchmark and Baseline via Robust Density Estimation
Word-level adversarial attacks have shown success in NLP models, drastically decreasing the performance of transformer-based models in recent years.
FEED: Feature-level Ensemble Effect for knowledge Distillation
This paper proposes a versatile and powerful training algorithm named Feature-level Ensemble Effect for knowledge Distillation(FEED), which is inspired by the work of factor transfer.
Korean Language Modeling via Syntactic Guide
Our experiment results prove that the proposed methods improve the model performance of the investigated Korean language understanding tasks.
SeqHAND: RGB-Sequence-Based 3D Hand Pose and Shape Estimation
In this paper, we attempt to not only consider the appearance of a hand but incorporate the temporal movement information of a hand in motion into the learning framework for better 3D hand pose estimation performance, which leads to the necessity of a large scale dataset with sequential RGB hand images.
A Revisit to the Decoder for Camouflaged Object Detection
Camouflaged object detection (COD) aims to generate a fine-grained segmentation map of camouflaged objects hidden in their background.
Unlocking the Potential of Unlabeled Data in Semi-Supervised Domain Generalization
To the best of our knowledge, we are the first to explore a method for incorporating the unconfident-unlabeled samples that were previously disregarded in SSDG setting.
ROODI: Reconstructing Occluded Objects with Denoising Inpainters
Isolating individual 3D Gaussian primitives for each object and handling occlusions in scenes remain far from being solved.
Decompose the model: Mechanistic interpretability in image models with Generalized Integrated Gradients (GIG)
We utilize Pointwise Feature Vectors (PFVs) and Effective Receptive Fields (ERFs) to decompose model embeddings into interpretable Concept Vectors.
MERLIN: Multimodal Embedding Refinement via LLM-based Iterative Navigation for Text-Video Retrieval-Rerank Pipeline
The rapid expansion of multimedia content has made accurately retrieving relevant videos from large collections increasingly challenging.
Practical Dataset Distillation Based on Deep Support Vectors
Conventional dataset distillation requires significant computational resources and assumes access to the entire dataset, an assumption impractical as it presumes all data resides on a central server.
Do not think about pink elephant!
Large Models (LMs) have heightened expectations for the potential of general AI as they are akin to human intelligence.
Coreset Selection for Object Detection
Coreset selection is a method for selecting a small, representative subset of an entire dataset.
Semi-Supervised Domain Adaptation for Wildfire Detection
In this paper, we propose a novel protocol for wildfire detection, leveraging semi-supervised Domain Adaptation for object detection, accompanied by a corresponding dataset designed for use by both academics and industries.
Unleash the Potential of CLIP for Video Highlight Detection
Multimodal and large language models (LLMs) have revolutionized the utilization of open-world knowledge, unlocking novel potentials across various tasks and applications.
Ranked #5 on
Highlight Detection
on QVHighlights
Deep Support Vectors
To this end, we propose the DeepKKT condition, an adaptation of the traditional Karush-Kuhn-Tucker (KKT) condition for deep learning models, and confirm that generated Deep Support Vectors (DSVs) using this condition exhibit properties similar to traditional support vectors.
Harmonizing Visual and Textual Embeddings for Zero-Shot Text-to-Image Customization
In a surge of text-to-image (T2I) models and their customization methods that generate new images of a user-provided subject, current works focus on alleviating the costs incurred by a lengthy per-subject optimization.
HourglassNeRF: Casting an Hourglass as a Bundle of Rays for Few-shot Neural Rendering
Addressing this, we propose HourglassNeRF, an effective regularization-based approach with a novel hourglass casting strategy.
Mitigating the Bias in the Model for Continual Test-Time Adaptation
Continual Test-Time Adaptation (CTA) is a challenging task that aims to adapt a source pre-trained model to continually changing target domains.
Respect the model: Fine-grained and Robust Explanation with Sharing Ratio Decomposition
The truthfulness of existing explanation methods in authentically elucidating the underlying model's decision-making process has been questioned.
Any-Way Meta Learning
Furthermore, we argue that the inherent label equivalence naturally lacks semantic information.
What How and When Should Object Detectors Update in Continually Changing Test Domains?
It is a well-known fact that the performance of deep learning models deteriorates when they encounter a distribution shift at test time.
What, How, and When Should Object Detectors Update in Continually Changing Test Domains?
It is a well-known fact that the performance of deep learning models deteriorates when they encounter a distribution shift at test time.
Open Domain Generalization with a Single Network by Regularization Exploiting Pre-trained Features
To handle this task, the model has to learn a generalizable representation that can be applied to unseen domains while also identify unknown classes that were not present during training.
SAVE: Protagonist Diversification with Structure Agnostic Video Editing
Finally, we decouple the motion from the appearance of the source video with an additional pseudo word.
Fast Sun-aligned Outdoor Scene Relighting based on TensoRF
In this work, we introduce our method of outdoor scene relighting for Neural Radiance Fields (NeRF) named Sun-aligned Relighting TensoRF (SR-TensoRF).
NICE: CVPR 2023 Challenge on Zero-shot Image Captioning
In this report, we introduce NICE (New frontiers for zero-shot Image Captioning Evaluation) project and share the results and outcomes of 2023 challenge.
ConcatPlexer: Additional Dim1 Batching for Faster ViTs
Transformers have demonstrated tremendous success not only in the natural language processing (NLP) domain but also the field of computer vision, igniting various creative approaches and applications.
Advancing Beyond Identification: Multi-bit Watermark for Large Language Models
By independently embedding sub-units of messages, the proposed method outperforms the existing works in terms of robustness and latency.
FlipNeRF: Flipped Reflection Rays for Few-shot Novel View Synthesis
Neural Radiance Field (NeRF) has been a mainstream in novel view synthesis with its remarkable quality of rendered images and simple architecture.
AADiff: Audio-Aligned Video Synthesis with Text-to-Image Diffusion
Recent advances in diffusion models have showcased promising results in the text-to-video (T2V) synthesis task.
Robust Multi-bit Natural Language Watermarking through Invariant Features
Recent years have witnessed a proliferation of valuable original natural language contents found in subscription-based media outlets, web novel platforms, and outputs of large language models.
Active Semi-Supervised Learning by Exploring Per-Sample Uncertainty and Consistency
Active Learning (AL) and Semi-supervised Learning are two techniques that have been studied to reduce the high cost of deep learning by using a small amount of labeled data and a large amount of unlabeled data.
MDPose: Real-Time Multi-Person Pose Estimation via Mixture Density Model
In this work, we propose a novel framework of single-stage instance-aware pose estimation by modeling the joint distribution of human keypoints with a mixture density model, termed as MDPose.
MixNeRF: Modeling a Ray with Mixture Density for Novel View Synthesis from Sparse Inputs
In this work, we propose MixNeRF, an effective training strategy for novel view synthesis from sparse inputs by modeling a ray with a mixture density model.
Analyzing Multimodal Objectives Through the Lens of Generative Diffusion Guidance
Recent years have witnessed astonishing advances in the field of multimodal representation learning, with contrastive learning being the cornerstone for major breakthroughs.
Unifying Vision-Language Representation Space with Single-tower Transformer
Contrastive learning is a form of distance learning that aims to learn invariant features from two related representations.
Leveraging Off-the-shelf Diffusion Model for Multi-attribute Fashion Image Manipulation
Fashion attribute editing is a task that aims to convert the semantic attributes of a given fashion image while preserving the irrelevant regions.
Towards Efficient Neural Scene Graphs by Learning Consistency Fields
Neural Radiance Fields (NeRF) achieves photo-realistic image rendering from novel views, and the Neural Scene Graphs (NSG) \cite{ost2021neural} extends it to dynamic scenes (video) with multiple objects.
Semantics-Guided Object Removal for Facial Images: with Broad Applicability and Robust Style Preservation
U-net, a conventional approach for conditional GANs, retains fine details of unmasked regions but the style of the reconstructed image is inconsistent with the rest of the original image and only works robustly when the size of the occluding object is small enough.
Conservative Generator, Progressive Discriminator: Coordination of Adversaries in Few-shot Incremental Image Synthesis
The capacity to learn incrementally from an online stream of data is an envied trait of human learners, as deep neural networks typically suffer from catastrophic forgetting and stability-plasticity dilemma.
Unsupervised Domain Adaptation for One-stage Object Detector using Offsets to Bounding Box
Most existing domain adaptive object detection methods exploit adversarial feature alignment to adapt the model to a new domain.
End-to-End Multi-Object Detection with a Regularized Mixture Model
Recent end-to-end multi-object detectors simplify the inference pipeline by removing hand-crafted processes such as non-maximum suppression (NMS).
Backdoor Attacks in Federated Learning by Rare Embeddings and Gradient Ensembling
For a less complex dataset, a mere 0. 1% of adversary clients is enough to poison the global model effectively.
Imposing Consistency for Optical Flow Estimation
Imposing consistency through proxy tasks has been shown to enhance data-driven learning and enable self-supervision in various tasks.
MatteFormer: Transformer-Based Image Matting via Prior-Tokens
In this paper, we propose a transformer-based image matting model called MatteFormer, which takes full advantage of trimap information in the transformer block.
Ranked #5 on
Image Matting
on Composition-1K
Pose-MUM : Reinforcing Key Points Relationship for Semi-Supervised Human Pose Estimation
A well-designed strong-weak augmentation strategy and the stable teacher to generate reliable pseudo labels are essential in the teacher-student framework of semi-supervised learning (SSL).
Detection of Word Adversarial Examples in Text Classification: Benchmark and Baseline via Robust Density Estimation
Word-level adversarial attacks have shown success in NLP models, drastically decreasing the performance of transformer-based models in recent years.
MUM: Mix Image Tiles and UnMix Feature Tiles for Semi-Supervised Object Detection
Data augmentation strategy plays a significant role in the SSL framework since it is hard to create a weak-strong augmented input pair without losing label information.
Self-Distilled Self-Supervised Representation Learning
Through self-distillation, the intermediate layers are better suited for instance discrimination, making the performance of an early-exited sub-network not much degraded from that of the full network.
Few-shot Image Generation with Mixup-based Distance Learning
Producing diverse and realistic images with generative models such as GANs typically requires large scale training with vast amount of images.
MUM : Mix Image Tiles and UnMix Feature Tiles for Semi-Supervised Object Detection
Data augmentation strategy plays a significant role in the SSL framework since it is hard to create a weak-strong augmented input pair without losing label information.
Local-Selective Feature Distillation for Single Image Super-Resolution
In this paper, we propose a novel feature distillation (FD) method which is suitable for SISR.
Dynamic Iterative Refinement for Efficient 3D Hand Pose Estimation
While hand pose estimation is a critical component of most interactive extended reality and gesture recognition systems, contemporary approaches are not optimized for computational and memory efficiency.
Exploiting Inter-pixel Correlations in Unsupervised Domain Adaptation for Semantic Segmentation
It creates a set of pseudo labels for the target domain to give explicit supervision.
Self-Evolutionary Optimization for Pareto Front Learning
Utilizing SEO for PFL, we also introduce self-evolutionary Pareto networks (SEPNet), enabling the unified model to approximate the entire Pareto front set that maximizes the hypervolume.
Few-Shot Object Detection by Attending to Per-Sample-Prototype
Second, each support sample is used as a class code to leverage the information by comparing similarities between each support feature and query features.
Ranked #14 on
Few-Shot Object Detection
on MS-COCO (30-shot)
Dynamic Collective Intelligence Learning: Finding Efficient Sparse Model via Refined Gradients for Pruned Weights
To alleviate this problem, dynamic pruning methods have emerged, which try to find diverse sparsity patterns during training by utilizing Straight-Through-Estimator (STE) to approximate gradients of pruned weights.
Normalization Matters in Weakly Supervised Object Localization
Weakly-supervised object localization (WSOL) enables finding an object using a dataset without any localization information.
PQK: Model Compression via Pruning, Quantization, and Knowledge Distillation
Unlike traditional pruning and KD, PQK makes use of unimportant weights pruned in the pruning process to make a teacher network for training a better student network without pre-training the teacher model.
LFI-CAM: Learning Feature Importance for Better Visual Explanation
LFI-CAM generates an attention map for visual explanation during forward propagation, at the same time, leverages the attention map to improve the classification performance through the attention mechanism.
Prototype-based Personalized Pruning
We verify the usefulness of PPP on a couple of tasks in computer vision and Keyword spotting.
The U-Net based GLOW for Optical-Flow-free Video Interframe Generation
Through experiments, we \sam {confirmed the feasibility of the proposed algorithm and would like to suggest the U-Net based Generative Flow as a new possibility for baseline in video frame interpolation.
Maximizing Cosine Similarity Between Spatial Features for Unsupervised Domain Adaptation in Semantic Segmentation
We propose a novel method that tackles the problem of unsupervised domain adaptation for semantic segmentation by maximizing the cosine similarity between the source and the target domain at the feature level.
Learning Dynamic BERT via Trainable Gate Variables and a Bi-modal Regularizer
The BERT model has shown significant success on various natural language processing tasks.
Learning Dynamic Network Using a Reuse Gate Function in Semi-supervised Video Object Segmentation
Current state-of-the-art approaches for Semi-supervised Video Object Segmentation (Semi-VOS) propagates information from previous frames to generate segmentation mask for the current frame.
StyleUV: Diverse and High-fidelity UV Map Generative Model
Meanwhile, with the advent of Generative Adversarial Networks (GANs), there has been great progress in reconstructing realistic 2D images.
TTVOS: Lightweight Video Object Segmentation with Adaptive Template Attention Module and Temporal Consistency Loss
Our template matching method consists of short-term and long-term matching.
Edge Bias in Federated Learning and its Solution by Buffered Knowledge Distillation
Federated learning (FL), which utilizes communication between the server (core) and local devices (edges) to indirectly learn from more data, is an emerging field in deep learning research.
Self-supervised pre-training and contrastive representation learning for multiple-choice video QA
In this paper, we propose novel training schemes for multiple-choice video question answering with a self-supervised pre-training stage and a supervised contrastive learning in the main stage as an auxiliary learning.
On the Orthogonality of Knowledge Distillation with Other Techniques: From an Ensemble Perspective
In this paper, we focus on knowledge distillation and demonstrate that knowledge distillation methods are orthogonal to other efficiency-enhancing methods both analytically and empirically.
Part-Aware Data Augmentation for 3D Object Detection in Point Cloud
In this paper, we propose part-aware data augmentation (PA-AUG) that can better utilize rich information of 3D label to enhance the performance of 3D object detectors.
Procrustean Regression Networks: Learning 3D Structure of Non-Rigid Objects from 2D Annotations
We propose a novel framework for training neural networks which is capable of learning 3D information of non-rigid objects when only 2D annotations are available as ground truths.
SeqHAND:RGB-Sequence-Based 3D Hand Pose and Shape Estimation
In this paper, we attempt to not only consider the appearance of a hand but incorporate the temporal movement information of a hand in motion into the learning framework for better 3D hand pose estimation performance, which leads to the necessity of a large scale dataset with sequential RGB hand images.
Interpolation-based semi-supervised learning for object detection
Despite the data labeling cost for the object detection tasks being substantially more than that of the classification tasks, semi-supervised learning methods for object detection have not been studied much.
Position-based Scaled Gradient for Model Quantization and Pruning
Second, we empirically show that PSG acting as a regularizer to a weight vector is favorable for model compression domains such as quantization and pruning.
KL-Divergence-Based Region Proposal Network for Object Detection
The learning of the region proposal in object detection using the deep neural networks (DNN) is divided into two tasks: binary classification and bounding box regression task.
LSQ+: Improving low-bit quantization through learnable offsets and better initialization
To solve this problem, we propose LSQ+, a natural extension of LSQ, wherein we introduce a general asymmetric quantization scheme with trainable scale and offset parameters that can learn to accommodate the negative activations.
Ranked #18 on
Quantization
on ImageNet
Class-Imbalanced Semi-Supervised Learning
First, we analyze existing SSL methods in imbalanced environments and examine how the class imbalance affects SSL methods.
Feature-map-level Online Adversarial Knowledge Distillation
By training a network to fool the corresponding discriminator, it can learn the other network's feature map distribution.
Consistency-based Semi-supervised Learning for Object detection
Making a precise annotation in a large dataset is crucial to the performance of object detection.
Ranked #18 on
Semi-Supervised Object Detection
on COCO 2% labeled data
QKD: Quantization-aware Knowledge Distillation
First, Self-studying (SS) phase fine-tunes a quantized low-precision student network without KD to obtain a good initialization.
Training Multi-Object Detector by Estimating Bounding Box Distribution for Input Image
Instead of assigning each ground truth to specific locations of network's output, we train a network by estimating the probability density of bounding boxes in an input image using a mixture model.
SINet: Extreme Lightweight Portrait Segmentation Networks with Spatial Squeeze Modules and Information Blocking Decoder
To solve the first problem, we introduce the new extremely lightweight portrait segmentation model SINet, containing an information blocking decoder and spatial squeeze modules.
Ranked #1 on
Portrait Segmentation
on EG1800
Tell Me What They're Holding: Weakly-supervised Object Detection with Transferable Knowledge from Human-object Interaction
In the source domain, we fully train an object detector and the RRPN with full supervision of HOI.
FEED: Feature-level Ensemble for Knowledge Distillation
We name this method as parallel FEED, andexperimental results on CIFAR-100 and ImageNet show that our method has clear performance enhancements, without introducing any additional parameters or computations at test time.
ExtremeC3Net: Extreme Lightweight Portrait Segmentation Networks using Advanced C3-modules
In our qualitative and quantitative analysis on the EG1800 dataset, we show that our method outperforms various existing lightweight segmentation models.
Image Enhancement by Recurrently-trained Super-resolution Network
The newly generated HR images by the repeatedly trained SR network show better image quality and this strategy of training LR to mimic new HR can lead to a more efficient SR network.
Pose estimator and tracker using temporal flow maps for limbs
In this paper, we propose temporal flow maps for limbs (TML) and a multi-stride method to estimate and track human poses.
Ranked #5 on
Pose Tracking
on PoseTrack2018
Selective Self-Training for semi-supervised Learning
While the conventional methods cannot be applied to the new SSL problems where the separated data do not share the classes, our method does not show any performance degradation even if the classes of unlabeled data are different from those of the labeled data.
Feature Fusion for Online Mutual Knowledge Distillation
We propose a learning framework named Feature Fusion Learning (FFL) that efficiently trains a powerful classifier through a fusion module which combines the feature maps generated from parallel neural networks.
Disentangling Options with Hellinger Distance Regularizer
In reinforcement learning (RL), temporal abstraction still remains as an important and unsolved problem.
A Comprehensive Overhaul of Feature Distillation
We investigate the design aspects of feature distillation methods achieving network compression and propose a novel feature distillation method in which the distillation loss is designed to make a synergy among various aspects: teacher transform, student transform, distillation feature position and distance function.
Ranked #49 on
Knowledge Distillation
on ImageNet
Task-oriented Design through Deep Reinforcement Learning
We propose a new low-cost machine-learning-based methodology which assists designers in reducing the gap between the problem and the solution in the design process.
Unpriortized Autoencoder For Image Generation
In this paper, we treat the image generation task using an autoencoder, a representative latent model.
URNet : User-Resizable Residual Networks with Conditional Gating Module
There are methods to reduce the cost by compressing networks or varying its computational path dynamically according to the input image.
C3: Concentrated-Comprehensive Convolution and its application to semantic segmentation
To resolve this problem, we propose a new block called Concentrated-Comprehensive Convolution (C3) which applies the asymmetric convolutions before the depth-wise separable dilated convolution to compensate for the information loss due to dilated convolution.
Genetic-Gated Networks for Deep Reinforcement Learning
We introduce the Genetic-Gated Networks (G2Ns), simple neural networks that combine a gate vector composed of binary genetic genes in the hidden layer(s) of networks.
Sym-parameterized Dynamic Inference for Mixed-Domain Image Translation
Recent advances in image-to-image translation have led to some ways to generate multiple domain images through a single network.
Genetic-Gated Networks for Deep Reinforcement
We introduce the Genetic-Gated Networks (G2Ns), simple neural networks that combine a gate vector composed of binary genetic genes in the hidden layer(s) of networks.
Towards Governing Agent's Efficacy: Action-Conditional $β$-VAE for Deep Transparent Reinforcement Learning
We tackle the blackbox issue of deep neural networks in the settings of reinforcement learning (RL) where neural agents learn towards maximizing reward gains in an uncontrollable way.
Textbook Question Answering with Multi-modal Context Graph Understanding and Self-supervised Open-set Comprehension
Moreover, ablation studies validate that both methods of incorporating f-GCN for extracting knowledge from multi-modal contexts and our newly proposed self-supervised learning process are effective for TQA problems.
HC-Net: Memory-based Incremental Dual-Network System for Continual learning
The C-Net guarantees no degradation in the performance of the previously learned tasks and the H-Net shows high confidence in finding the origin of an input sample.
StackNet: Stacking Parameters for Continual learning
The StackNet guarantees no degradation in the performance of the previously learned tasks and the index module shows high confidence in finding the origin of an input sample.
Motion Feature Network: Fixed Motion Filter for Action Recognition
However, it has an expensive computational cost and requires two-stream (RGB and optical flow) framework.
Vehicle Image Generation Going Well with The Surroundings
Using a subnetwork based on a precedent work of image completion, our model makes the shape of an object.
Semantic Sentence Matching with Densely-connected Recurrent and Co-attentive Information
Inspired by DenseNet, a densely connected convolutional network, we propose a densely-connected co-attentive recurrent neural network, each layer of which uses concatenated information of attentive features as well as hidden features of all the preceding recurrent layers.
Ranked #13 on
Natural Language Inference
on SNLI
3D Human Pose Estimation with Relational Networks
In this paper, we propose a novel 3D human pose estimation algorithm from a single image based on neural networks.
Image Restoration by Estimating Frequency Distribution of Local Patches
In this paper, we propose a method to solve the image restoration problem, which tries to restore the details of a corrupted image, especially due to the loss caused by JPEG compression.
Music Source Separation Using Stacked Hourglass Networks
In this paper, we propose a simple yet effective method for multiple music source separation using convolutional neural networks.
Sound Audio and Speech Processing
MC-GAN: Multi-conditional Generative Adversarial Network for Image Synthesis
This block enables MC-GAN to generate a realistic object image with the desired background by controlling the amount of the background information from the given base image using the foreground information from the text attributes.
Paraphrasing Complex Network: Network Compression via Factor Transfer
Among the model compression methods, a method called knowledge transfer is to train a student network with a stronger teacher network.
Broadcasting Convolutional Network for Visual Relational Reasoning
In this paper, we propose the Broadcasting Convolutional Network (BCN) that extracts key object features from the global field of an entire input image and recognizes their relationship with local features.
Dynamic Graph Generation Network: Generating Relational Knowledge from Diagrams
In this work, we introduce a new algorithm for analyzing a diagram, which contains visual and textual information in an abstract and integrated way.
Butterfly Effect: Bidirectional Control of Classification Performance by Small Additive Perturbation
In addition to this performance enhancement problem, we show that the proposed PGN can be adopted to solve the classical adversarial problem without utilizing the information on the target classifier.
BOOK: Storing Algorithm-Invariant Episodes for Deep Reinforcement Learning
Our method learns hundreds to thousand times faster than the conventional methods by learning only a handful of core cluster information, which shows that deep RL agents can effectively learn through the shared knowledge from other agents.
Residual Features and Unified Prediction Network for Single Stage Detection
They are much faster than two stage detectors that use region proposal networks (RPN) without much degradation in the detection performances.
Where to Play: Retrieval of Video Segments using Natural-Language Queries
Through quantitative and qualitative evaluation, we show that our method is effective for retrieval of video segments using natural language queries.
Superpixel-based Semantic Segmentation Trained by Statistical Process Control
Semantic segmentation, like other fields of computer vision, has seen a remarkable performance advance by the use of deep convolution neural networks.
Enhancement of SSD by concatenating feature maps for object detection
In this paper, we propose and analyze how to use feature maps effectively to improve the performance of the conventional SSD.
3D Human Pose Estimation Using Convolutional Neural Networks with 2D Pose Information
While there has been a success in 2D human pose estimation with convolutional neural networks (CNNs), 3D human pose estimation has not been thoroughly studied.
Ranked #334 on
3D Human Pose Estimation
on Human3.6M
Membership Representation for Detecting Block-Diagonal Structure in Low-Rank or Sparse Subspace Clustering
The proposed method shares the philosophy of the above subspace clustering methods, in that it is a self-expressive system based on a Hadamard product of a membership matrix.
