PyTorchKR Papers

최신 논문

총 9,339편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2209.11112공식 코드 · Papers with Code

CMGAN: Conformer-Based Metric-GAN for Monaural Speech Enhancement

In this work, we further develop the conformer-based metric generative adversarial network (CMGAN) model for speech enhancement (SE) in the time-frequency (TF) domain. This paper builds on our previous work but takes a more in-depth look by conducting extensive ablation studies on model inputs and architectural design choices. We rigorously tested the generalization ability of the model to unseen noise types and distortions. We have fortified our claims through DNS-MOS measurements and listening tests. Rather than focusing exclusively on the speech denoising task, we extend this work to address the dereverberation and super-resolution tasks. This necessitated exploring various architectural changes, specifically metric discriminator scores and masking techniques. It is essential to highlight that this is among the earliest works that attempted complex TF-domain super-resolution. Our findings show that CMGAN outperforms existing state-of-the-art methods in the three major speech enhancement tasks: denoising, dereverberation, and super-resolution. For example, in the denoising task using the Voice Bank+DEMAND dataset, CMGAN notably exceeded the performance of prior models, attaining a PESQ score of 3.41 and an SSNR of 11.10 dB. Audio samples and CMGAN implementations are available online.

arXiv:2209.10283

A Comparative Analysis of the Time and Energy Demand of Versatile Video Coding and High Efficiency Video Coding Reference Decoders

This paper investigates the decoding energy and decoding time demand of VTM-7.0 in relation to HM-16.20. We present the first detailed comparison of two video codecs in terms of software decoder energy consumption. The evaluation shows that the energy demand of the VTM decoder is increased significantly compared to HM and that the increase depends on the coding configuration. For the coding configuration randomaccess, we find that the decoding energy is increased by over 80% at a decoding time increase of over 70%. Furthermore, results indicate that the energy demand increases by up to 207% when Single Instruction Multiple Data (SIMD) instructions are disabled, which corresponds to the HM implementation style. By measurements, it is revealed that the coding tools MIP, AMVR, TPM, LFNST, and MTS increase the energy efficiency of the decoder. Furthermore, we propose a new coding configuration based on our analysis, which reduces the energy demand of the VTM decoder by over 17% on average.

arXiv:2209.10268

Extending Video Decoding Energy Models for 360° and HDR Video Formats in HEVC

Research has shown that decoder energy models are helpful tools for improving the energy efficiency in video playback applications. For example, an accurate feature-based bit stream model can reduce the energy consumption of the decoding process. However, until now only sequences of the SDR video format were investigated. Therefore, this paper shows that the decoding energy of HEVC-coded bit streams can be estimated precisely for different video formats and coding bit depths. Therefore, we compare a state-of-the-art model from the literature with a proposed model. We show that bit streams of the 360{\deg}, HDR, and fisheye video format can be estimated with a mean estimation error lower than 3.88% if the setups have the same coding bit depth. Furthermore, it is shown that on average, the energy demand for the decoding of bit streams with a bit depth of 10-bit is 55% higher than with 8-bit.

arXiv:2209.10266

Decoding Energy Modeling For Versatile Video Coding

In previous research, it was shown that the software decoding energy demand of High Efficiency Video Coding (HEVC) can be reduced by 15$\%$ by using a decoding-energy-rate-distortion optimization algorithm. To achieve this, the energy demand of the decoder has to be modeled by a bit stream feature-based model with sufficiently high accuracy. Therefore, we propose two bit stream feature-based models for the upcoming Versatile Video Coding (VVC) standard. The newly introduced models are compared with models from literature, which are used for HEVC. An evaluation of the proposed models reveals that the mean estimation error is similar to the results of the literature and yields an estimation error of 1.85% with 10-fold cross-validation.

arXiv:2209.10158

Position-Aware Relation Learning for RGB-Thermal Salient Object Detection

RGB-Thermal salient object detection (SOD) combines two spectra to segment visually conspicuous regions in images. Most existing methods use boundary maps to learn the sharp boundary. These methods ignore the interactions between isolated boundary pixels and other confident pixels, leading to sub-optimal performance. To address this problem,we propose a position-aware relation learning network (PRLNet) for RGB-T SOD based on swin transformer. PRLNet explores the distance and direction relationships between pixels to strengthen intra-class compactness and inter-class separation, generating salient object masks with clear boundaries and homogeneous regions. Specifically, we develop a novel signed distance map auxiliary module (SDMAM) to improve encoder feature representation, which takes into account the distance relation of different pixels in boundary neighborhoods. Then, we design a feature refinement approach with directional field (FRDF), which rectifies features of boundary neighborhood by exploiting the features inside salient objects. FRDF utilizes the directional information between object pixels to effectively enhance the intra-class compactness of salient regions. In addition, we constitute a pure transformer encoder-decoder network to enhance multispectral feature representation for RGB-T SOD. Finally, we conduct quantitative and qualitative experiments on three public benchmark datasets.The results demonstrate that our proposed method outperforms the state-of-the-art methods.

arXiv:2209.09735

Relaxed Attention for Transformer Models

The powerful modeling capabilities of all-attention-based transformer architectures often cause overfitting and - for natural language processing tasks - lead to an implicitly learned internal language model in the autoregressive transformer decoder complicating the integration of external language models. In this paper, we explore relaxed attention, a simple and easy-to-implement smoothing of the attention weights, yielding a two-fold improvement to the general transformer architecture: First, relaxed attention provides regularization when applied to the self-attention layers in the encoder. Second, we show that it naturally supports the integration of an external language model as it suppresses the implicitly learned internal language model by relaxing the cross attention in the decoder. We demonstrate the benefit of relaxed attention across several tasks with clear improvement in combination with recent benchmark approaches. Specifically, we exceed the former state-of-the-art performance of 26.90% word error rate on the largest public lip-reading LRS3 benchmark with a word error rate of 26.31%, as well as we achieve a top-performing BLEU score of 37.67 on the IWSLT14 (DE$\rightarrow$EN) machine translation task without external language models and virtually no additional model parameters. Code and models will be made publicly available.

arXiv:2209.09244

Flexible Neural Image Compression via Code Editing

Neural image compression (NIC) has outperformed traditional image codecs in rate-distortion (R-D) performance. However, it usually requires a dedicated encoder-decoder pair for each point on R-D curve, which greatly hinders its practical deployment. While some recent works have enabled bitrate control via conditional coding, they impose strong prior during training and provide limited flexibility. In this paper we propose Code Editing, a highly flexible coding method for NIC based on semi-amortized inference and adaptive quantization. Our work is a new paradigm for variable bitrate NIC. Furthermore, experimental results show that our method surpasses existing variable-rate methods, and achieves ROI coding and multi-distortion trade-off with a single decoder.

arXiv:2209.08738

Learning Decoupled Retrieval Representation for Nearest Neighbour Neural Machine Translation

K-Nearest Neighbor Neural Machine Translation (kNN-MT) successfully incorporates external corpus by retrieving word-level representations at test time. Generally, kNN-MT borrows the off-the-shelf context representation in the translation task, e.g., the output of the last decoder layer, as the query vector of the retrieval task. In this work, we highlight that coupling the representations of these two tasks is sub-optimal for fine-grained retrieval. To alleviate it, we leverage supervised contrastive learning to learn the distinctive retrieval representation derived from the original context representation. We also propose a fast and effective approach to constructing hard negative samples. Experimental results on five domains show that our approach improves the retrieval accuracy and BLEU score compared to vanilla kNN-MT.

arXiv:2209.09078공식 코드 · Papers with Code

NIERT: Accurate Numerical Interpolation through Unifying Scattered Data Representations using Transformer Encoder

Interpolation for scattered data is a classical problem in numerical analysis, with a long history of theoretical and practical contributions. Recent advances have utilized deep neural networks to construct interpolators, exhibiting excellent and generalizable performance. However, they still fall short in two aspects: \textbf{1) inadequate representation learning}, resulting from separate embeddings of observed and target points in popular encoder-decoder frameworks and \textbf{2) limited generalization power}, caused by overlooking prior interpolation knowledge shared across different domains. To overcome these limitations, we present a \textbf{N}umerical \textbf{I}nterpolation approach using \textbf{E}ncoder \textbf{R}epresentation of \textbf{T}ransformers (called \textbf{NIERT}). On one hand, NIERT utilizes an encoder-only framework rather than the encoder-decoder structure. This way, NIERT can embed observed and target points into a unified encoder representation space, thus effectively exploiting the correlations among them and obtaining more precise representations. On the other hand, we propose to pre-train NIERT on large-scale synthetic mathematical functions to acquire prior interpolation knowledge, and transfer it to multiple interpolation domains with consistent performance gain. On both synthetic and real-world datasets, NIERT outperforms the existing approaches by a large margin, i.e., 4.3$\sim$14.3$\times$ lower MAE on TFRD subsets, and 1.7/1.8/8.7$\times$ lower MSE on Mathit/PhysioNet/PTV datasets. The source code of NIERT is available at https://github.com/DingShizhe/NIERT.

arXiv:2209.08217공식 코드 · Papers with Code

Delving Globally into Texture and Structure for Image Inpainting

Image inpainting has achieved remarkable progress and inspired abundant methods, where the critical bottleneck is identified as how to fulfill the high-frequency structure and low-frequency texture information on the masked regions with semantics. To this end, deep models exhibit powerful superiority to capture them, yet constrained on the local spatial regions. In this paper, we delve globally into texture and structure information to well capture the semantics for image inpainting. As opposed to the existing arts trapped on the independent local patches, the texture information of each patch is reconstructed from all other patches across the whole image, to match the coarsely filled information, specially the structure information over the masked regions. Unlike the current decoder-only transformer within the pixel level for image inpainting, our model adopts the transformer pipeline paired with both encoder and decoder. On one hand, the encoder captures the texture semantic correlations of all patches across image via self-attention module. On the other hand, an adaptive patch vocabulary is dynamically established in the decoder for the filled patches over the masked regions. Building on this, a structure-texture matching attention module anchored on the known regions comes up to marry the best of these two worlds for progressive inpainting via a probabilistic diffusion process. Our model is orthogonal to the fashionable arts, such as Convolutional Neural Networks (CNNs), Attention and Transformer model, from the perspective of texture and structure information for image inpainting. The extensive experiments over the benchmarks validate its superiority. Our code is available at https://github.com/htyjers/DGTS-Inpainting.

arXiv:2209.09002

MoVQ: Modulating Quantized Vectors for High-Fidelity Image Generation

Although two-stage Vector Quantized (VQ) generative models allow for synthesizing high-fidelity and high-resolution images, their quantization operator encodes similar patches within an image into the same index, resulting in a repeated artifact for similar adjacent regions using existing decoder architectures. To address this issue, we propose to incorporate the spatially conditional normalization to modulate the quantized vectors so as to insert spatially variant information to the embedded index maps, encouraging the decoder to generate more photorealistic images. Moreover, we use multichannel quantization to increase the recombination capability of the discrete codes without increasing the cost of model and codebook. Additionally, to generate discrete tokens at the second stage, we adopt a Masked Generative Image Transformer (MaskGIT) to learn an underlying prior distribution in the compressed latent space, which is much faster than the conventional autoregressive model. Experiments on two benchmark datasets demonstrate that our proposed modulated VQGAN is able to greatly improve the reconstructed image quality as well as provide high-fidelity image generation.

arXiv:2209.07861

Machine Learning Decoder for 5G NR PUCCH Format 0

5G cellular systems depend on the timely exchange of feedback control information between the user equipment and the base station. Proper decoding of this control information is necessary to set up and sustain high throughput radio links. This paper makes the first attempt at using Machine Learning techniques to improve the decoding performance of the Physical Uplink Control Channel Format 0. We use fully connected neural networks to classify the received samples based on the uplink control information content embedded within them. The trained neural network, tested on real-time wireless captures, shows significant improvement in accuracy over conventional DFT-based decoders, even at low SNR. The obtained accuracy results also demonstrate conformance with 3GPP requirements.

arXiv:2209.07634공식 코드 · Papers with Code

Stateful Memory-Augmented Transformers for Efficient Dialogue Modeling

Transformer encoder-decoder models have achieved great performance in dialogue generation tasks, however, their inability to process long dialogue history often leads to truncation of the context To address this problem, we propose a novel memory-augmented transformer that is compatible with existing pre-trained encoder-decoder models and enables efficient preservation of the dialogue history information. By incorporating a separate memory module alongside the pre-trained transformer, the model can effectively interchange information between the memory states and the current input context. We evaluate our model on three dialogue datasets and two language modeling datasets. Experimental results show that our method has achieved superior efficiency and performance compared to other pre-trained Transformer baselines.

arXiv:2209.07600공식 코드 · Papers with Code

STPOTR: Simultaneous Human Trajectory and Pose Prediction Using a Non-Autoregressive Transformer for Robot Following Ahead

In this paper, we develop a neural network model to predict future human motion from an observed human motion history. We propose a non-autoregressive transformer architecture to leverage its parallel nature for easier training and fast, accurate predictions at test time. The proposed architecture divides human motion prediction into two parts: 1) the human trajectory, which is the hip joint 3D position over time and 2) the human pose which is the all other joints 3D positions over time with respect to a fixed hip joint. We propose to make the two predictions simultaneously, as the shared representation can improve the model performance. Therefore, the model consists of two sets of encoders and decoders. First, a multi-head attention module applied to encoder outputs improves human trajectory. Second, another multi-head self-attention module applied to encoder outputs concatenated with decoder outputs facilitates learning of temporal dependencies. Our model is well-suited for robotic applications in terms of test accuracy and speed, and compares favorably with respect to state-of-the-art methods. We demonstrate the real-world applicability of our work via the Robot Follow-Ahead task, a challenging yet practical case study for our proposed model.

arXiv:2209.07989

CurveFormer: 3D Lane Detection by Curve Propagation with Curve Queries and Attention

3D lane detection is an integral part of autonomous driving systems. Previous CNN and Transformer-based methods usually first generate a bird's-eye-view (BEV) feature map from the front view image, and then use a sub-network with BEV feature map as input to predict 3D lanes. Such approaches require an explicit view transformation between BEV and front view, which itself is still a challenging problem. In this paper, we propose CurveFormer, a single-stage Transformer-based method that directly calculates 3D lane parameters and can circumvent the difficult view transformation step. Specifically, we formulate 3D lane detection as a curve propagation problem by using curve queries. A 3D lane query is represented by a dynamic and ordered anchor point set. In this way, queries with curve representation in Transformer decoder iteratively refine the 3D lane detection results. Moreover, a curve cross-attention module is introduced to compute the similarities between curve queries and image features. Additionally, a context sampling module that can capture more relative image features of a curve query is provided to further boost the 3D lane detection performance. We evaluate our method for 3D lane detection on both synthetic and real-world datasets, and the experimental results show that our method achieves promising performance compared with the state-of-the-art approaches. The effectiveness of each component is validated via ablation studies as well.

arXiv:2209.07953

StyleGAN Encoder-Based Attack for Block Scrambled Face Images

In this paper, we propose an attack method to block scrambled face images, particularly Encryption-then-Compression (EtC) applied images by utilizing the existing powerful StyleGAN encoder and decoder for the first time. Instead of reconstructing identical images as plain ones from encrypted images, we focus on recovering styles that can reveal identifiable information from the encrypted images. The proposed method trains an encoder by using plain and encrypted image pairs with a particular training strategy. While state-of-the-art attack methods cannot recover any perceptual information from EtC images, the proposed method discloses personally identifiable information such as hair color, skin color, eyeglasses, gender, etc. Experiments were carried out on the CelebA dataset, and results show that reconstructed images have some perceptual similarities compared to plain images.

arXiv:2209.07857공식 코드 · Papers with Code

GATraj: A Graph- and Attention-based Multi-Agent Trajectory Prediction Model

Trajectory prediction has been a long-standing problem in intelligent systems like autonomous driving and robot navigation. Models trained on large-scale benchmarks have made significant progress in improving prediction accuracy. However, the importance on efficiency for real-time applications has been less emphasized. This paper proposes an attention-based graph model, named GATraj, which achieves a good balance of prediction accuracy and inference speed. We use attention mechanisms to model the spatial-temporal dynamics of agents, such as pedestrians or vehicles, and a graph convolutional network to model their interactions. Additionally, a Laplacian mixture decoder is implemented to mitigate mode collapse and generate diverse multimodal predictions for each agent. GATraj achieves state-of-the-art prediction performance at a much higher speed when tested on the ETH/UCY datasets for pedestrian trajectories, and good performance at about 100 Hz inference speed when tested on the nuScenes dataset for autonomous driving. We conduct extensive experiments to analyze the probability estimation of the Laplacian mixture decoder and compare it with a Gaussian mixture decoder for predicting different multimodalities. Furthermore, comprehensive ablation studies demonstrate the effectiveness of each proposed module in GATraj. The code is released at https://github.com/mengmengliu1998/GATraj.

arXiv:2209.07795

KaliCalib: A Framework for Basketball Court Registration

Tracking the players and the ball in team sports is key to analyse the performance or to enhance the game watching experience with augmented reality. When the only sources for this data are broadcast videos, sports-field registration systems are required to estimate the homography and re-project the ball or the players from the image space to the field space. This paper describes a new basketball court registration framework in the context of the MMSports 2022 camera calibration challenge. The method is based on the estimation by an encoder-decoder network of the positions of keypoints sampled with perspective-aware constraints. The regression of the basket positions and heavy data augmentation techniques make the model robust to different arenas. Ablation studies show the positive effects of our contributions on the challenge test set. Our method divides the mean squared error by 4.7 compared to the challenge baseline.

arXiv:2209.07692공식 코드 · Papers with Code

Answering Numerical Reasoning Questions in Table-Text Hybrid Contents with Graph-based Encoder and Tree-based Decoder

In the real-world question answering scenarios, hybrid form combining both tabular and textual contents has attracted more and more attention, among which numerical reasoning problem is one of the most typical and challenging problems. Existing methods usually adopt encoder-decoder framework to represent hybrid contents and generate answers. However, it can not capture the rich relationship among numerical value, table schema, and text information on the encoder side. The decoder uses a simple predefined operator classifier which is not flexible enough to handle numerical reasoning processes with diverse expressions. To address these problems, this paper proposes a \textbf{Re}lational \textbf{G}raph enhanced \textbf{H}ybrid table-text \textbf{N}umerical reasoning model with \textbf{T}ree decoder (\textbf{RegHNT}). It models the numerical question answering over table-text hybrid contents as an expression tree generation task. Moreover, we propose a novel relational graph modeling method, which models alignment between questions, tables, and paragraphs. We validated our model on the publicly available table-text hybrid QA benchmark (TAT-QA). The proposed RegHNT significantly outperform the baseline model and achieve state-of-the-art results. We openly released the source code and data at https://github.com/lfy79001/RegHNT (2022-05-05).

arXiv:2209.07496

Unsupervised Opinion Summarization Using Approximate Geodesics

Opinion summarization is the task of creating summaries capturing popular opinions from user reviews. In this paper, we introduce Geodesic Summarizer (GeoSumm), a novel system to perform unsupervised extractive opinion summarization. GeoSumm involves an encoder-decoder based representation learning model, that generates representations of text as a distribution over latent semantic units. GeoSumm generates these representations by performing dictionary learning over pre-trained text representations at multiple decoder layers. We then use these representations to quantify the relevance of review sentences using a novel approximate geodesic distance based scoring mechanism. We use the relevance scores to identify popular opinions in order to compose general and aspect-specific summaries. Our proposed model, GeoSumm, achieves state-of-the-art performance on three opinion summarization datasets. We perform additional experiments to analyze the functioning of our model and showcase the generalization ability of {\X} across different domains.