PyTorchKR Papers

최신 논문

총 9,642편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2110.08811

Attention W-Net: Improved Skip Connections for better Representations

Segmentation of macro and microvascular structures in fundoscopic retinal images plays a crucial role in the detection of multiple retinal and systemic diseases, yet it is a difficult problem to solve. Most neural network approaches face several issues such as lack of enough parameters, overfitting and/or incompatibility between internal feature-spaces. We propose Attention W-Net, a new U-Net based architecture for retinal vessel segmentation to address these problems. In this architecture, we have two main contributions: Attention Block and regularisation measures. Our Attention Block uses attention between encoder and decoder features, resulting in higher compatibility upon addition. Our regularisation measures include augmentation and modifications to the ResNet Block used, which greatly prevent overfitting. We observe an F1 and AUC of 0.8407 and 0.9833 on the DRIVE and 0.8174 and 0.9865 respectively on the CHASE-DB1 datasets - a sizeable improvement over its backbone as well as competitive performance among contemporary state-of-the-art methods.

arXiv:2110.08636공식 코드 · Papers with Code

DPC: Unsupervised Deep Point Correspondence via Cross and Self Construction

We present a new method for real-time non-rigid dense correspondence between point clouds based on structured shape construction. Our method, termed Deep Point Correspondence (DPC), requires a fraction of the training data compared to previous techniques and presents better generalization capabilities. Until now, two main approaches have been suggested for the dense correspondence problem. The first is a spectral-based approach that obtains great results on synthetic datasets but requires mesh connectivity of the shapes and long inference processing time while being unstable in real-world scenarios. The second is a spatial approach that uses an encoder-decoder framework to regress an ordered point cloud for the matching alignment from an irregular input. Unfortunately, the decoder brings considerable disadvantages, as it requires a large amount of training data and struggles to generalize well in cross-dataset evaluations. DPC's novelty lies in its lack of a decoder component. Instead, we use latent similarity and the input coordinates themselves to construct the point cloud and determine correspondence, replacing the coordinate regression done by the decoder. Extensive experiments show that our construction scheme leads to a performance boost in comparison to recent state-of-the-art correspondence methods. Our code is publicly available at https://github.com/dvirginz/DPC.

arXiv:2110.08568공식 코드 · Papers with Code

ASFormer: Transformer for Action Segmentation

Algorithms for the action segmentation task typically use temporal models to predict what action is occurring at each frame for a minute-long daily activity. Recent studies have shown the potential of Transformer in modeling the relations among elements in sequential data. However, there are several major concerns when directly applying the Transformer to the action segmentation task, such as the lack of inductive biases with small training sets, the deficit in processing long input sequence, and the limitation of the decoder architecture to utilize temporal relations among multiple action segments to refine the initial predictions. To address these concerns, we design an efficient Transformer-based model for action segmentation task, named ASFormer, with three distinctive characteristics: (i) We explicitly bring in the local connectivity inductive priors because of the high locality of features. It constrains the hypothesis space within a reliable scope, and is beneficial for the action segmentation task to learn a proper target function with small training sets. (ii) We apply a pre-defined hierarchical representation pattern that efficiently handles long input sequences. (iii) We carefully design the decoder to refine the initial predictions from the encoder. Extensive experiments on three public datasets demonstrate that effectiveness of our methods. Code is available at \url{https://github.com/ChinaYi/ASFormer}.

arXiv:2110.08578

Visual-aware Attention Dual-stream Decoder for Video Captioning

Video captioning is a challenging task that captures different visual parts and describes them in sentences, for it requires visual and linguistic coherence. The attention mechanism in the current video captioning method learns to assign weight to each frame, promoting the decoder dynamically. This may not explicitly model the correlation and the temporal coherence of the visual features extracted in the sequence frames.To generate semantically coherent sentences, we propose a new Visual-aware Attention (VA) model, which concatenates dynamic changes of temporal sequence frames with the words at the previous moment, as the input of attention mechanism to extract sequence features.In addition, the prevalent approaches widely use the teacher-forcing (TF) learning during training, where the next token is generated conditioned on the previous ground-truth tokens. The semantic information in the previously generated tokens is lost. Therefore, we design a self-forcing (SF) stream that takes the semantic information in the probability distribution of the previous token as input to enhance the current token.The Dual-stream Decoder (DD) architecture unifies the TF and SF streams, generating sentences to promote the annotated captioning for both streams.Meanwhile, with the Dual-stream Decoder utilized, the exposure bias problem is alleviated, caused by the discrepancy between the training and testing in the TF learning.The effectiveness of the proposed Visual-aware Attention Dual-stream Decoder (VADD) is demonstrated through the result of experimental studies on Microsoft video description (MSVD) corpus and MSR-Video to text (MSR-VTT) datasets.

arXiv:2110.07844공식 코드 · Papers with Code

Modeling Endorsement for Multi-Document Abstractive Summarization

A crucial difference between single- and multi-document summarization is how salient content manifests itself in the document(s). While such content may appear at the beginning of a single document, essential information is frequently reiterated in a set of documents related to a particular topic, resulting in an endorsement effect that increases information salience. In this paper, we model the cross-document endorsement effect and its utilization in multiple document summarization. Our method generates a synopsis from each document, which serves as an endorser to identify salient content from other documents. Strongly endorsed text segments are used to enrich a neural encoder-decoder model to consolidate them into an abstractive summary. The method has a great potential to learn from fewer examples to identify salient content, which alleviates the need for costly retraining when the set of documents is dynamically adjusted. Through extensive experiments on benchmark multi-document summarization datasets, we demonstrate the effectiveness of our proposed method over strong published baselines. Finally, we shed light on future research directions and discuss broader challenges of this task using a case study.

arXiv:2110.08207공식 코드 · Papers with Code

Multitask Prompted Training Enables Zero-Shot Task Generalization

Large language models have recently been shown to attain reasonable zero-shot generalization on a diverse set of tasks (Brown et al., 2020). It has been hypothesized that this is a consequence of implicit multitask learning in language models' pretraining (Radford et al., 2019). Can zero-shot generalization instead be directly induced by explicit multitask learning? To test this question at scale, we develop a system for easily mapping any natural language tasks into a human-readable prompted form. We convert a large set of supervised datasets, each with multiple prompts with diverse wording. These prompted datasets allow for benchmarking the ability of a model to perform completely held-out tasks. We fine-tune a pretrained encoder-decoder model (Raffel et al., 2020; Lester et al., 2021) on this multitask mixture covering a wide variety of tasks. The model attains strong zero-shot performance on several standard datasets, often outperforming models up to 16x its size. Further, our approach attains strong performance on a subset of tasks from the BIG-bench benchmark, outperforming models up to 6x its size. All trained models are available at https://github.com/bigscience-workshop/t-zero and all prompts are available at https://github.com/bigscience-workshop/promptsource.

arXiv:2110.08130

Breaking Down Multilingual Machine Translation

While multilingual training is now an essential ingredient in machine translation (MT) systems, recent work has demonstrated that it has different effects in different multilingual settings, such as many-to-one, one-to-many, and many-to-many learning. These training settings expose the encoder and the decoder in a machine translation model with different data distributions. In this paper, we examine how different varieties of multilingual training contribute to learning these two components of the MT model. Specifically, we compare bilingual models with encoders and/or decoders initialized by multilingual training. We show that multilingual training is beneficial to encoders in general, while it only benefits decoders for low-resource languages (LRLs). We further find the important attention heads for each language pair and compare their correlations during inference. Our analysis sheds light on how multilingual translation models work and enables us to propose methods to improve performance by training with highly related languages. Our many-to-one models for high-resource languages and one-to-many models for LRL outperform the best results reported by Aharoni et al. (2019)

arXiv:2110.07993

Pose-guided Generative Adversarial Net for Novel View Action Synthesis

We focus on the problem of novel-view human action synthesis. Given an action video, the goal is to generate the same action from an unseen viewpoint. Naturally, novel view video synthesis is more challenging than image synthesis. It requires the synthesis of a sequence of realistic frames with temporal coherency. Besides, transferring the different actions to a novel target view requires awareness of action category and viewpoint change simultaneously. To address these challenges, we propose a novel framework named Pose-guided Action Separable Generative Adversarial Net (PAS-GAN), which utilizes pose to alleviate the difficulty of this task. First, we propose a recurrent pose-transformation module which transforms actions from the source view to the target view and generates novel view pose sequence in 2D coordinate space. Second, a well-transformed pose sequence enables us to separatethe action and background in the target view. We employ a novel local-global spatial transformation module to effectively generate sequential video features in the target view using these action and background features. Finally, the generated video features are used to synthesize human action with the help of a 3D decoder. Moreover, to focus on dynamic action in the video, we propose a novel multi-scale action-separable loss which further improves the video quality. We conduct extensive experiments on two large-scale multi-view human action datasets, NTU-RGBD and PKU-MMD, demonstrating the effectiveness of PAS-GAN which outperforms existing approaches.

arXiv:2110.07717

Deep Human-guided Conditional Variational Generative Modeling for Automated Urban Planning

Urban planning designs land-use configurations and can benefit building livable, sustainable, safe communities. Inspired by image generation, deep urban planning aims to leverage deep learning to generate land-use configurations. However, urban planning is a complex process. Existing studies usually ignore the need of personalized human guidance in planning, and spatial hierarchical structure in planning generation. Moreover, the lack of large-scale land-use configuration samples poses a data sparsity challenge. This paper studies a novel deep human guided urban planning method to jointly solve the above challenges. Specifically, we formulate the problem into a deep conditional variational autoencoder based framework. In this framework, we exploit the deep encoder-decoder design to generate land-use configurations. To capture the spatial hierarchy structure of land uses, we enforce the decoder to generate both the coarse-grained layer of functional zones, and the fine-grained layer of POI distributions. To integrate human guidance, we allow humans to describe what they need as texts and use these texts as a model condition input. To mitigate training data sparsity and improve model robustness, we introduce a variational Gaussian embedding mechanism. It not just allows us to better approximate the embedding space distribution of training data and sample a larger population to overcome sparsity, but also adds more probabilistic randomness into the urban planning generation to improve embedding diversity so as to improve robustness. Finally, we present extensive experiments to validate the enhanced performances of our method.

arXiv:2110.07358

Memory-Based Semantic Parsing

We present a memory-based model for context-dependent semantic parsing. Previous approaches focus on enabling the decoder to copy or modify the parse from the previous utterance, assuming there is a dependency between the current and previous parses. In this work, we propose to represent contextual information using an external memory. We learn a context memory controller that manages the memory by maintaining the cumulative meaning of sequential user utterances. We evaluate our approach on three semantic parsing benchmarks. Experimental results show that our model can better process context-dependent information and demonstrates improved performance without using task-specific decoders.

arXiv:2110.07182

Adversarial examples by perturbing high-level features in intermediate decoder layers

We propose a novel method for creating adversarial examples. Instead of perturbing pixels, we use an encoder-decoder representation of the input image and perturb intermediate layers in the decoder. This changes the high-level features provided by the generative model. Therefore, our perturbation possesses semantic meaning, such as a longer beak or green tints. We formulate this task as an optimization problem by minimizing the Wasserstein distance between the adversarial and initial images under a misclassification constraint. We employ the projected gradient method with a simple inexact projection. Due to the projection, all iterations are feasible, and our method always generates adversarial images. We perform numerical experiments on the MNIST and ImageNet datasets in both targeted and untargeted settings. We demonstrate that our adversarial images are much less vulnerable to steganographic defence techniques than pixel-based attacks. Moreover, we show that our method modifies key features such as edges and that defence techniques based on adversarial training are vulnerable to our attacks.

arXiv:2110.07205공식 코드 · Papers with Code

SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing

Motivated by the success of T5 (Text-To-Text Transfer Transformer) in pre-trained natural language processing models, we propose a unified-modal SpeechT5 framework that explores the encoder-decoder pre-training for self-supervised speech/text representation learning. The SpeechT5 framework consists of a shared encoder-decoder network and six modal-specific (speech/text) pre/post-nets. After preprocessing the input speech/text through the pre-nets, the shared encoder-decoder network models the sequence-to-sequence transformation, and then the post-nets generate the output in the speech/text modality based on the output of the decoder. Leveraging large-scale unlabeled speech and text data, we pre-train SpeechT5 to learn a unified-modal representation, hoping to improve the modeling capability for both speech and text. To align the textual and speech information into this unified semantic space, we propose a cross-modal vector quantization approach that randomly mixes up speech/text states with latent units as the interface between encoder and decoder. Extensive evaluations show the superiority of the proposed SpeechT5 framework on a wide variety of spoken language processing tasks, including automatic speech recognition, speech synthesis, speech translation, voice conversion, speech enhancement, and speaker identification. We release our code and model at https://github.com/microsoft/SpeechT5.

arXiv:2110.06986공식 코드 · Papers with Code

ADMM-DAD net: a deep unfolding network for analysis compressed sensing

In this paper, we propose a new deep unfolding neural network based on the ADMM algorithm for analysis Compressed Sensing. The proposed network jointly learns a redundant analysis operator for sparsification and reconstructs the signal of interest. We compare our proposed network with a state-of-the-art unfolded ISTA decoder, that also learns an orthogonal sparsifier. Moreover, we consider not only image, but also speech datasets as test examples. Computational experiments demonstrate that our proposed network outperforms the state-of-the-art deep unfolding network, consistently for both real-world image and speech datasets.

arXiv:2110.06920

Semantics-aware Attention Improves Neural Machine Translation

The integration of syntactic structures into Transformer machine translation has shown positive results, but to our knowledge, no work has attempted to do so with semantic structures. In this work we propose two novel parameter-free methods for injecting semantic information into Transformers, both rely on semantics-aware masking of (some of) the attention heads. One such method operates on the encoder, through a Scene-Aware Self-Attention (SASA) head. Another on the decoder, through a Scene-Aware Cross-Attention (SACrA) head. We show a consistent improvement over the vanilla Transformer and syntax-aware models for four language pairs. We further show an additional gain when using both semantic and syntactic structures in some language pairs.

arXiv:2110.06672공식 코드 · Papers with Code

The Deep Generative Decoder: MAP estimation of representations improves modeling of single-cell RNA data

Learning low-dimensional representations of single-cell transcriptomics has become instrumental to its downstream analysis. The state of the art is currently represented by neural network models such as variational autoencoders (VAEs) which use a variational approximation of the likelihood for inference. We here present the Deep Generative Decoder (DGD), a simple generative model that computes model parameters and representations directly via maximum a posteriori (MAP) estimation. The DGD handles complex parameterized latent distributions naturally unlike VAEs which typically use a fixed Gaussian distribution, because of the complexity of adding other types. We first show its general functionality on a commonly used benchmark set, Fashion-MNIST. Secondly, we apply the model to multiple single-cell data sets. Here the DGD learns low-dimensional, meaningful and well-structured latent representations with sub-clustering beyond the provided labels. The advantages of this approach are its simplicity and its capability to provide representations of much smaller dimensionality than a comparable VAE.

arXiv:2110.06634

End-to-end translation of human neural activity to speech with a dual-dual generative adversarial network

In a recent study of auditory evoked potential (AEP) based brain-computer interface (BCI), it was shown that, with an encoder-decoder framework, it is possible to translate human neural activity to speech (T-CAS). However, current encoder-decoder-based methods achieve T-CAS often with a two-step method where the information is passed between the encoder and decoder with a shared dimension reduction vector, which may result in a loss of information. A potential approach to this problem is to design an end-to-end method by using a dual generative adversarial network (DualGAN) without dimension reduction of passing information, but it cannot realize one-to-one signal-to-signal translation (see Fig.1 (a) and (b)). In this paper, we propose an end-to-end model to translate human neural activity to speech directly, create a new electroencephalogram (EEG) datasets for participants with good attention by design a device to detect participants' attention, and introduce a dual-dual generative adversarial network (Dual-DualGAN) (see Fig. 1 (c) and (d)) to address an end-to-end translation of human neural activity to speech (ET-CAS) problem by group labelling EEG signals and speech signals, inserting a transition domain to realize cross-domain mapping. In the transition domain, the transition signals are cascaded by the corresponding EEG and speech signals in a certain proportion, which can build bridges for EEG and speech signals without corresponding features, and realize one-to-one cross-domain EEG-to-speech translation. The proposed method can translate word-length and sentence-length sequences of neural activity to speech. Experimental evaluation has been conducted to show that the proposed method significantly outperforms state-of-the-art methods on both words and sentences of auditory stimulus.

arXiv:2110.06393

Attention-guided Generative Models for Extractive Question Answering

We propose a novel method for applying Transformer models to extractive question answering (QA) tasks. Recently, pretrained generative sequence-to-sequence (seq2seq) models have achieved great success in question answering. Contributing to the success of these models are internal attention mechanisms such as cross-attention. We propose a simple strategy to obtain an extractive answer span from the generative model by leveraging the decoder cross-attention patterns. Viewing cross-attention as an architectural prior, we apply joint training to further improve QA performance. Empirical results show that on open-domain question answering datasets like NaturalQuestions and TriviaQA, our method approaches state-of-the-art performance on both generative and extractive inference, all while using much fewer parameters. Furthermore, this strategy allows us to perform hallucination-free inference while conferring significant improvements to the model's ability to rerank relevant passages.

arXiv:2110.06615

CLIP4Caption: CLIP for Video Caption

Video captioning is a challenging task since it requires generating sentences describing various diverse and complex videos. Existing video captioning models lack adequate visual representation due to the neglect of the existence of gaps between videos and texts. To bridge this gap, in this paper, we propose a CLIP4Caption framework that improves video captioning based on a CLIP-enhanced video-text matching network (VTM). This framework is taking full advantage of the information from both vision and language and enforcing the model to learn strongly text-correlated video features for text generation. Besides, unlike most existing models using LSTM or GRU as the sentence decoder, we adopt a Transformer structured decoder network to effectively learn the long-range visual and language dependency. Additionally, we introduce a novel ensemble strategy for captioning tasks. Experimental results demonstrate the effectiveness of our method on two datasets: 1) on MSR-VTT dataset, our method achieved a new state-of-the-art result with a significant gain of up to 10% in CIDEr; 2) on the private test data, our method ranking 2nd place in the ACM MM multimedia grand challenge 2021: Pre-training for Video Understanding Challenge. It is noted that our model is only trained on the MSR-VTT dataset.

arXiv:2110.06550

Saliency Detection via Global Context Enhanced Feature Fusion and Edge Weighted Loss

UNet-based methods have shown outstanding performance in salient object detection (SOD), but are problematic in two aspects. 1) Indiscriminately integrating the encoder feature, which contains spatial information for multiple objects, and the decoder feature, which contains global information of the salient object, is likely to convey unnecessary details of non-salient objects to the decoder, hindering saliency detection. 2) To deal with ambiguous object boundaries and generate accurate saliency maps, the model needs additional branches, such as edge reconstructions, which leads to increasing computational cost. To address the problems, we propose a context fusion decoder network (CFDN) and near edge weighted loss (NEWLoss) function. The CFDN creates an accurate saliency map by integrating global context information and thus suppressing the influence of the unnecessary spatial information. NEWLoss accelerates learning of obscure boundaries without additional modules by generating weight maps on object boundaries. Our method is evaluated on four benchmarks and achieves state-of-the-art performance. We prove the effectiveness of the proposed method through comparative experiments.

arXiv:2110.06823

A Speaker-aware Parallel Hierarchical Attentive Encoder-Decoder Model for Multi-turn Dialogue Generation

This paper presents a novel open-domain dialogue generation model emphasizing the differentiation of speakers in multi-turn conversations. Differing from prior work that solely relies on the content of conversation history to generate a response, we argue that capturing relative social relations among utterances (i.e., generated by either the same speaker or different persons) benefits the machine capturing fine-grained context information from a conversation history to improve context coherence in the generated response. Given that, we propose a speaker-aware Parallel Hierarchical Attentive Encoder-Decoder (PHAED) model that aims to model each utterance with the awareness of its speaker and contextual associations with the same speaker's previous messages. Specifically, in a conversation involving two speakers, we regard the utterances from one speaker as responses and those from the other as queries. After understanding queries via our encoder with inner-query and inter-query encodings, our decoder reuses the hidden states of previously generated responses, instead of reconstructing these by the encoder, to generate a new response. Our empirical results show that PHAED outperforms the state-of-the-art in both automatic and human evaluations. Furthermore, our ablation study shows that dialogue models with speaker tokens can generally decrease the possibility of generating non-coherent responses regarding the conversation context.