PyTorchKR Papers

최신 논문

총 9,339편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2206.05239공식 코드 · Papers with Code

StructCoder: Structure-Aware Transformer for Code Generation

There has been a recent surge of interest in automating software engineering tasks using deep learning. This paper addresses the problem of code generation, where the goal is to generate target code given source code in a different language or a natural language description. Most state-of-the-art deep learning models for code generation use training strategies primarily designed for natural language. However, understanding and generating code requires a more rigorous comprehension of the code syntax and semantics. With this motivation, we develop an encoder-decoder Transformer model where both the encoder and decoder are explicitly trained to recognize the syntax and data flow in the source and target codes, respectively. We not only make the encoder structure-aware by leveraging the source code's syntax tree and data flow graph, but we also support the decoder in preserving the syntax and data flow of the target code by introducing two novel auxiliary tasks: AST (Abstract Syntax Tree) paths prediction and data flow prediction. To the best of our knowledge, this is the first work to introduce a structure-aware Transformer decoder that models both syntax and data flow to enhance the quality of generated code. The proposed StructCoder model achieves state-of-the-art performance on code translation and text-to-code generation tasks in the CodeXGLUE benchmark, and improves over baselines of similar size on the APPS code generation benchmark. Our code is publicly available at https://github.com/reddy-lab-code-research/StructCoder/.

arXiv:2206.04337

Precoder and Decoder Co-Designs for Radar and Communication Spectrum Sharing

Radar and modern communication systems are both evaluating towards higher frequency bands and massive antenna arrays, thus increasing their similarities in terms of hardware structure, channel characteristics, and signal processing pipelines. To suppress the cross-system interference caused by communications and radar systems with shared spectral and hardware resources, the co-design philosophy, wherein the communications and radar/sensing systems can operate in parallel with jointly optimized performance, has drawn substantial attention from both academia and industry. In this paper, we propose a nullspace-based joint precoder-decoder design for spectrum sharing between multicarrier radar and multiuser multicarrier communication systems, by employing the maximizing signal interference noise ratio (max-SINR) criterion and interference alignment (IA) constraints. By projecting the cross-system interference to the designed null spaces, a maximum degree of freedom upper bound for the $K+1$-radar-communication-user interference channel can be achieved. Our simulation studies demonstrate that interference can be practically fully canceled in both communication and radar systems. This leads to improved detection performance in radar and a higher rate in communication subsystems. A significant performance gain over a nullspace-based precoder-only design is also obtained.

arXiv:2206.04571공식 코드 · Papers with Code

Revisiting End-to-End Speech-to-Text Translation From Scratch

End-to-end (E2E) speech-to-text translation (ST) often depends on pretraining its encoder and/or decoder using source transcripts via speech recognition or text translation tasks, without which translation performance drops substantially. However, transcripts are not always available, and how significant such pretraining is for E2E ST has rarely been studied in the literature. In this paper, we revisit this question and explore the extent to which the quality of E2E ST trained on speech-translation pairs alone can be improved. We reexamine several techniques proven beneficial to ST previously, and offer a set of best practices that biases a Transformer-based E2E ST system toward training from scratch. Besides, we propose parameterized distance penalty to facilitate the modeling of locality in the self-attention model for speech. On four benchmarks covering 23 languages, our experiments show that, without using any transcripts or pretraining, the proposed system reaches and even outperforms previous studies adopting pretraining, although the gap remains in (extremely) low-resource settings. Finally, we discuss neural acoustic feature modeling, where a neural model is designed to extract acoustic features from raw speech signals directly, with the goal to simplify inductive biases and add freedom to the model in describing speech. For the first time, we demonstrate its feasibility and show encouraging results on ST tasks.

arXiv:2206.04465

Joint Encoder-Decoder Self-Supervised Pre-training for ASR

Self-supervised learning (SSL) has shown tremendous success in various speech-related downstream tasks, including Automatic Speech Recognition (ASR). The output embeddings of the SSL model are treated as powerful short-time representations of the speech signal. However, in the ASR task, the main objective is to get the correct sequence of acoustic units, characters, or byte-pair encodings (BPEs). Usually, encoder-decoder architecture works exceptionally well for a sequence-to-sequence task like ASR. Therefore, in this paper, we propose a new paradigm that exploits the power of a decoder during self-supervised learning. We use Hidden Unit BERT (HuBERT) SSL framework to compute the conventional masked prediction loss for the encoder. In addition, we have introduced a decoder in the SSL framework and proposed a target preparation strategy for the decoder. Finally, we use a multitask SSL setup wherein we jointly optimize both the encoder and decoder losses. We hypothesize that the presence of a decoder in the SSL model helps it learn an acoustic unit-based language model, which might improve the performance of an ASR downstream task. We compare our proposed SSL model with HuBERT and show up to 25% relative improvement in performance on ASR by finetuning on various LibriSpeech subsets.

arXiv:2206.04328

Novel projection schemes for graph-based Light Field coding

In Light Field compression, graph-based coding is powerful to exploit signal redundancy along irregular shapes and obtains good energy compaction. However, apart from high time complexity to process high dimensional graphs, their graph construction method is highly sensitive to the accuracy of disparity information between viewpoints. In real world Light Field or synthetic Light Field generated by computer software, the use of disparity information for super-rays projection might suffer from inaccuracy due to vignetting effect and large disparity between views in the two types of Light Fields respectively. This paper introduces two novel projection schemes resulting in less error in disparity information, in which one projection scheme can also significantly reduce time computation for both encoder and decoder. Experimental results show projection quality of super-pixels across views can be considerably enhanced using the proposals, along with rate-distortion performance when compared against original projection scheme and HEVC-based or JPEG Pleno-based coding approaches.

arXiv:2206.04503

cycle text2face: cycle text-to-face gan via transformers

Text-to-face is a subset of text-to-image that require more complex architecture due to their more detailed production. In this paper, we present an encoder-decoder model called Cycle Text2Face. Cycle Text2Face is a new initiative in the encoder part, it uses a sentence transformer and GAN to generate the image described by the text. The Cycle is completed by reproducing the text of the face in the decoder part of the model. Evaluating the model using the CelebA dataset, leads to better results than previous GAN-based models. In measuring the quality of the generate face, in addition to satisfying the human audience, we obtain an FID score of 3.458. This model, with high-speed processing, provides quality face images in the short time.

arXiv:2206.03551

NOMAD: Nonlinear Manifold Decoders for Operator Learning

Supervised learning in function spaces is an emerging area of machine learning research with applications to the prediction of complex physical systems such as fluid flows, solid mechanics, and climate modeling. By directly learning maps (operators) between infinite dimensional function spaces, these models are able to learn discretization invariant representations of target functions. A common approach is to represent such target functions as linear combinations of basis elements learned from data. However, there are simple scenarios where, even though the target functions form a low dimensional submanifold, a very large number of basis elements is needed for an accurate linear representation. Here we present NOMAD, a novel operator learning framework with a nonlinear decoder map capable of learning finite dimensional representations of nonlinear submanifolds in function spaces. We show this method is able to accurately learn low dimensional representations of solution manifolds to partial differential equations while outperforming linear models of larger size. Additionally, we compare to state-of-the-art operator learning methods on a complex fluid dynamics benchmark and achieve competitive performance with a significantly smaller model size and training cost.

arXiv:2206.04045

STable: Table Generation Framework for Encoder-Decoder Models

The output structure of database-like tables, consisting of values structured in horizontal rows and vertical columns identifiable by name, can cover a wide range of NLP tasks. Following this constatation, we propose a framework for text-to-table neural models applicable to problems such as extraction of line items, joint entity and relation extraction, or knowledge base population. The permutation-based decoder of our proposal is a generalized sequential method that comprehends information from all cells in the table. The training maximizes the expected log-likelihood for a table's content across all random permutations of the factorization order. During the content inference, we exploit the model's ability to generate cells in any order by searching over possible orderings to maximize the model's confidence and avoid substantial error accumulation, which other sequential models are prone to. Experiments demonstrate a high practical value of the framework, which establishes state-of-the-art results on several challenging datasets, outperforming previous solutions by up to 15%.

arXiv:2206.03858

Rotation-Equivariant Conditional Spherical Neural Fields for Learning a Natural Illumination Prior

Inverse rendering is an ill-posed problem. Previous work has sought to resolve this by focussing on priors for object or scene shape or appearance. In this work, we instead focus on a prior for natural illuminations. Current methods rely on spherical harmonic lighting or other generic representations and, at best, a simplistic prior on the parameters. We propose a conditional neural field representation based on a variational auto-decoder with a SIREN network and, extending Vector Neurons, build equivariance directly into the network. Using this, we develop a rotation-equivariant, high dynamic range (HDR) neural illumination model that is compact and able to express complex, high-frequency features of natural environment maps. Training our model on a curated dataset of 1.6K HDR environment maps of natural scenes, we compare it against traditional representations, demonstrate its applicability for an inverse rendering task and show environment map completion from partial observations. A PyTorch implementation, our dataset and trained models can be found at jadgardner.github.io/RENI.

arXiv:2206.03826

Towards Understanding Why Mask-Reconstruction Pretraining Helps in Downstream Tasks

For unsupervised pretraining, mask-reconstruction pretraining (MRP) approaches, e.g. MAE and data2vec, randomly mask input patches and then reconstruct the pixels or semantic features of these masked patches via an auto-encoder. Then for a downstream task, supervised fine-tuning the pretrained encoder remarkably surpasses the conventional ``supervised learning'' (SL) trained from scratch. However, it is still unclear 1) how MRP performs semantic feature learning in the pretraining phase and 2) why it helps in downstream tasks. To solve these problems, we first theoretically show that on an auto-encoder of a two/one-layered convolution encoder/decoder, MRP can capture all discriminative features of each potential semantic class in the pretraining dataset. Then considering the fact that the pretraining dataset is of huge size and high diversity and thus covers most features in downstream dataset, in fine-tuning phase, the pretrained encoder can capture as much features as it can in downstream datasets, and would not lost these features with theoretical guarantees. In contrast, SL only randomly captures some features due to lottery ticket hypothesis. So MRP provably achieves better performance than SL on the classification tasks. Experimental results testify to our data assumptions and also our theoretical implications.

arXiv:2206.03544

A Penny for Your (visual) Thoughts: Self-Supervised Reconstruction of Natural Movies from Brain Activity

Reconstructing natural videos from fMRI brain recordings is very challenging, for two main reasons: (i) As fMRI data acquisition is difficult, we only have a limited amount of supervised samples, which is not enough to cover the huge space of natural videos; and (ii) The temporal resolution of fMRI recordings is much lower than the frame rate of natural videos. In this paper, we propose a self-supervised approach for natural-movie reconstruction. By employing cycle-consistency over Encoding-Decoding natural videos, we can: (i) exploit the full framerate of the training videos, and not be limited only to clips that correspond to fMRI recordings; (ii) exploit massive amounts of external natural videos which the subjects never saw inside the fMRI machine. These enable increasing the applicable training data by several orders of magnitude, introducing natural video priors to the decoding network, as well as temporal coherence. Our approach significantly outperforms competing methods, since those train only on the limited supervised data. We further introduce a new and simple temporal prior of natural videos, which - when folded into our fMRI decoder further - allows us to reconstruct videos at a higher frame-rate (HFR) of up to x8 of the original fMRI sample rate.

arXiv:2206.03318

LegoNN: Building Modular Encoder-Decoder Models

State-of-the-art encoder-decoder models (e.g. for machine translation (MT) or automatic speech recognition (ASR)) are constructed and trained end-to-end as an atomic unit. No component of the model can be (re-)used without the others, making it impossible to share parts, e.g. a high resourced decoder, across tasks. We describe LegoNN, a procedure for building encoder-decoder architectures in a way so that its parts can be applied to other tasks without the need for any fine-tuning. To achieve this reusability, the interface between encoder and decoder modules is grounded to a sequence of marginal distributions over a pre-defined discrete vocabulary. We present two approaches for ingesting these marginals; one is differentiable, allowing the flow of gradients across the entire network, and the other is gradient-isolating. To enable the portability of decoder modules between MT tasks for different source languages and across other tasks like ASR, we introduce a modality agnostic encoder which consists of a length control mechanism to dynamically adapt encoders' output lengths in order to match the expected input length range of pre-trained decoders. We present several experiments to demonstrate the effectiveness of LegoNN models: a trained language generation LegoNN decoder module from German-English (De-En) MT task can be reused without any fine-tuning for the Europarl English ASR and the Romanian-English (Ro-En) MT tasks, matching or beating the performance of baseline. After fine-tuning, LegoNN models improve the Ro-En MT task by 1.5 BLEU points and achieve 12.5% relative WER reduction on the Europarl ASR task. To show how the approach generalizes, we compose a LegoNN ASR model from three modules -- each has been learned within different end-to-end trained models on three different datasets -- achieving an overall WER reduction of 19.5%.

arXiv:2206.03246

Portfolio Transformer for Attention-Based Asset Allocation

Traditional approaches to financial asset allocation start with returns forecasting followed by an optimization stage that decides the optimal asset weights. Any errors made during the forecasting step reduce the accuracy of the asset weightings, and hence the profitability of the overall portfolio. The Portfolio Transformer (PT) network, introduced here, circumvents the need to predict asset returns and instead directly optimizes the Sharpe ratio, a risk-adjusted performance metric widely used in practice. The PT is a novel end-to-end portfolio optimization framework, inspired by the numerous successes of attention mechanisms in natural language processing. With its full encoder-decoder architecture, specialized time encoding layers, and gating components, the PT has a high capacity to learn long-term dependencies among portfolio assets and hence can adapt more quickly to changing market conditions such as the COVID-19 pandemic. To demonstrate its robustness, the PT is compared against other algorithms, including the current LSTM-based state of the art, on three different datasets, with results showing that it offers the best risk-adjusted performance.

arXiv:2206.03173

Speaker-Guided Encoder-Decoder Framework for Emotion Recognition in Conversation

The emotion recognition in conversation (ERC) task aims to predict the emotion label of an utterance in a conversation. Since the dependencies between speakers are complex and dynamic, which consist of intra- and inter-speaker dependencies, the modeling of speaker-specific information is a vital role in ERC. Although existing researchers have proposed various methods of speaker interaction modeling, they cannot explore dynamic intra- and inter-speaker dependencies jointly, leading to the insufficient comprehension of context and further hindering emotion prediction. To this end, we design a novel speaker modeling scheme that explores intra- and inter-speaker dependencies jointly in a dynamic manner. Besides, we propose a Speaker-Guided Encoder-Decoder (SGED) framework for ERC, which fully exploits speaker information for the decoding of emotion. We use different existing methods as the conversational context encoder of our framework, showing the high scalability and flexibility of the proposed framework. Experimental results demonstrate the superiority and effectiveness of SGED.

arXiv:2206.03391

Data Stealing Attack on Medical Images: Is it Safe to Export Networks from Data Lakes?

In privacy-preserving machine learning, it is common that the owner of the learned model does not have any physical access to the data. Instead, only a secured remote access to a data lake is granted to the model owner without any ability to retrieve data from the data lake. Yet, the model owner may want to export the trained model periodically from the remote repository and a question arises whether this may cause is a risk of data leakage. In this paper, we introduce the concept of data stealing attack during the export of neural networks. It consists in hiding some information in the exported network that allows the reconstruction outside the data lake of images initially stored in that data lake. More precisely, we show that it is possible to train a network that can perform lossy image compression and at the same time solve some utility tasks such as image segmentation. The attack then proceeds by exporting the compression decoder network together with some image codes that leads to the image reconstruction outside the data lake. We explore the feasibility of such attacks on databases of CT and MR images, showing that it is possible to obtain perceptually meaningful reconstructions of the target dataset, and that the stolen dataset can be used in turns to solve a broad range of tasks. Comprehensive experiments and analyses show that data stealing attacks should be considered as a threat for sensitive imaging data sources.

arXiv:2206.03113

Wavelet Prior Attention Learning in Axial Inpainting Network

Image inpainting is the task of filling masked or unknown regions of an image with visually realistic contents, which has been remarkably improved by Deep Neural Networks (DNNs) recently. Essentially, as an inverse problem, the inpainting has the underlying challenges of reconstructing semantically coherent results without texture artifacts. Many previous efforts have been made via exploiting attention mechanisms and prior knowledge, such as edges and semantic segmentation. However, these works are still limited in practice by an avalanche of learnable prior parameters and prohibitive computational burden. To this end, we propose a novel model -- Wavelet prior attention learning in Axial Inpainting Network (WAIN), whose generator contains the encoder, decoder, as well as two key components of Wavelet image Prior Attention (WPA) and stacked multi-layer Axial-Transformers (ATs). Particularly, the WPA guides the high-level feature aggregation in the multi-scale frequency domain, alleviating the textual artifacts. Stacked ATs employ unmasked clues to help model reasonable features along with low-level features of horizontal and vertical axes, improving the semantic coherence. Extensive quantitative and qualitative experiments on Celeba-HQ and Places2 datasets are conducted to validate that our WAIN can achieve state-of-the-art performance over the competitors. The codes and models will be released.

arXiv:2206.02588

Coding of volumetric content with MIV using VVC subpictures

Storage and transport of six degrees of freedom (6DoF) dynamic volumetric visual content for immersive applications requires efficient compression. ISO/IEC MPEG has recently been working on a standard that aims to efficiently code and deliver 6DoF immersive visual experiences. This standard is called the MIV. MIV uses regular 2D video codecs to code the visual data. MPEG jointly with ITU-T VCEG, has also specified the VVC standard. VVC introduced recently the concept of subpicture. This tool was specifically designed to provide independent accessibility and decodability of sub-bitstreams for omnidirectional applications. This paper shows the benefit of using subpictures in the MIV use-case. While different ways in which subpictures could be used in MIV are discussed, a particular case study is selected. Namely, subpictures are used for parallel encoding and to reduce the number of decoder instances. Experimental results show that the cost of using subpictures in terms of bitrate overhead is negligible (0.1% to 0.4%), when compared to the overall bitrate. The number of decoder instances on the other hand decreases by a factor of two.

arXiv:2206.02743

A Neural Corpus Indexer for Document Retrieval

Current state-of-the-art document retrieval solutions mainly follow an index-retrieve paradigm, where the index is hard to be directly optimized for the final retrieval target. In this paper, we aim to show that an end-to-end deep neural network unifying training and indexing stages can significantly improve the recall performance of traditional methods. To this end, we propose Neural Corpus Indexer (NCI), a sequence-to-sequence network that generates relevant document identifiers directly for a designated query. To optimize the recall performance of NCI, we invent a prefix-aware weight-adaptive decoder architecture, and leverage tailored techniques including query generation, semantic document identifiers, and consistency-based regularization. Empirical studies demonstrated the superiority of NCI on two commonly used academic benchmarks, achieving +21.4% and +16.8% relative enhancement for Recall@1 on NQ320k dataset and R-Precision on TriviaQA dataset, respectively, compared to the best baseline method.

arXiv:2206.02079

Multilingual Neural Machine Translation with Deep Encoder and Multiple Shallow Decoders

Recent work in multilingual translation advances translation quality surpassing bilingual baselines using deep transformer models with increased capacity. However, the extra latency and memory costs introduced by this approach may make it unacceptable for efficiency-constrained applications. It has recently been shown for bilingual translation that using a deep encoder and shallow decoder (DESD) can reduce inference latency while maintaining translation quality, so we study similar speed-accuracy trade-offs for multilingual translation. We find that for many-to-one translation we can indeed increase decoder speed without sacrificing quality using this approach, but for one-to-many translation, shallow decoders cause a clear quality drop. To ameliorate this drop, we propose a deep encoder with multiple shallow decoders (DEMSD) where each shallow decoder is responsible for a disjoint subset of target languages. Specifically, the DEMSD model with 2-layer decoders is able to obtain a 1.8x speedup on average compared to a standard transformer model with no drop in translation quality.

arXiv:2206.01874공식 코드 · Papers with Code

An Unpooling Layer for Graph Generation

We propose a novel and trainable graph unpooling layer for effective graph generation. Given a graph with features, the unpooling layer enlarges this graph and learns its desired new structure and features. Since this unpooling layer is trainable, it can be applied to graph generation either in the decoder of a variational autoencoder or in the generator of a generative adversarial network (GAN). We prove that the unpooled graph remains connected and any connected graph can be sequentially unpooled from a 3-nodes graph. We apply the unpooling layer within the GAN generator. Since the most studied instance of graph generation is molecular generation, we test our ideas in this context. Using the QM9 and ZINC datasets, we demonstrate the improvement obtained by using the unpooling layer instead of an adjacency-matrix-based approach.