PyTorchKR Papers

최신 논문

총 9,217편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2411.19726

Towards Santali Linguistic Inclusion: Building the First Santali-to-English Translation Model using mT5 Transformer and Data Augmentation

Around seven million individuals in India, Bangladesh, Bhutan, and Nepal speak Santali, positioning it as nearly the third most commonly used Austroasiatic language. Despite its prominence among the Austroasiatic language family's Munda subfamily, Santali lacks global recognition. Currently, no translation models exist for the Santali language. Our paper aims to include Santali to the NPL spectrum. We aim to examine the feasibility of building Santali translation models based on available Santali corpora. The paper successfully addressed the low-resource problem and, with promising results, examined the possibility of creating a functional Santali machine translation model in a low-resource setup. Our study shows that Santali-English parallel corpus performs better when in transformers like mt5 as opposed to untrained transformers, proving that transfer learning can be a viable technique that works with Santali language. Besides the mT5 transformer, Santali-English performs better than Santali-Bangla parallel corpus as the mT5 has been trained in way more English data than Bangla data. Lastly, our study shows that with data augmentation, our model performs better.

arXiv:2411.19371공식 코드 · Papers with Code

Parameter-Efficient Transfer Learning for Music Foundation Models

More music foundation models are recently being released, promising a general, mostly task independent encoding of musical information. Common ways of adapting music foundation models to downstream tasks are probing and fine-tuning. These common transfer learning approaches, however, face challenges. Probing might lead to suboptimal performance because the pre-trained weights are frozen, while fine-tuning is computationally expensive and is prone to overfitting. Our work investigates the use of parameter-efficient transfer learning (PETL) for music foundation models which integrates the advantage of probing and fine-tuning. We introduce three types of PETL methods: adapter-based methods, prompt-based methods, and reparameterization-based methods. These methods train only a small number of parameters, and therefore do not require significant computational resources. Results show that PETL methods outperform both probing and fine-tuning on music auto-tagging. On key detection and tempo estimation, they achieve similar results as fine-tuning with significantly less training cost. However, the usefulness of the current generation of foundation model on key and tempo tasks is questioned by the similar results achieved by training a small model from scratch. Code available at https://github.com/suncerock/peft-music/

arXiv:2411.18926

Data Augmentation with Diffusion Models for Colon Polyp Localization on the Low Data Regime: How much real data is enough?

The scarcity of data in medical domains hinders the performance of Deep Learning models. Data augmentation techniques can alleviate that problem, but they usually rely on functional transformations of the data that do not guarantee to preserve the original tasks. To approximate the distribution of the data using generative models is a way of reducing that problem and also to obtain new samples that resemble the original data. Denoising Diffusion models is a promising Deep Learning technique that can learn good approximations of different kinds of data like images, time series or tabular data. Automatic colonoscopy analysis and specifically Polyp localization in colonoscopy videos is a task that can assist clinical diagnosis and treatment. The annotation of video frames for training a deep learning model is a time consuming task and usually only small datasets can be obtained. The fine tuning of application models using a large dataset of generated data could be an alternative to improve their performance. We conduct a set of experiments training different diffusion models that can generate jointly colonoscopy images with localization annotations using a combination of existing open datasets. The generated data is used on various transfer learning experiments in the task of polyp localization with a model based on YOLO v9 on the low data regime.

arXiv:2411.19041공식 코드 · Papers with Code

TAMT: Temporal-Aware Model Tuning for Cross-Domain Few-Shot Action Recognition

Going beyond few-shot action recognition (FSAR), cross-domain FSAR (CDFSAR) has attracted recent research interests by solving the domain gap lying in source-to-target transfer learning. Existing CDFSAR methods mainly focus on joint training of source and target data to mitigate the side effect of domain gap. However, such kind of methods suffer from two limitations: First, pair-wise joint training requires retraining deep models in case of one source data and multiple target ones, which incurs heavy computation cost, especially for large source and small target data. Second, pre-trained models after joint training are adopted to target domain in a straightforward manner, hardly taking full potential of pre-trained models and then limiting recognition performance. To overcome above limitations, this paper proposes a simple yet effective baseline, namely Temporal-Aware Model Tuning (TAMT) for CDFSAR. Specifically, our TAMT involves a decoupled paradigm by performing pre-training on source data and fine-tuning target data, which avoids retraining for multiple target data with single source. To effectively and efficiently explore the potential of pre-trained models in transferring to target domain, our TAMT proposes a Hierarchical Temporal Tuning Network (HTTN), whose core involves local temporal-aware adapters (TAA) and a global temporal-aware moment tuning (GTMT). Particularly, TAA learns few parameters to recalibrate the intermediate features of frozen pre-trained models, enabling efficient adaptation to target domains. Furthermore, GTMT helps to generate powerful video representations, improving match performance on the target domain. Experiments on several widely used video benchmarks show our TAMT outperforms the recently proposed counterparts by 13%$\sim$31%, achieving new state-of-the-art CDFSAR results.

arXiv:2411.19539

Knowledge Management for Automobile Failure Analysis Using Graph RAG

This paper presents a knowledge management system for automobile failure analysis using retrieval-augmented generation (RAG) with large language models (LLMs) and knowledge graphs (KGs). In the automotive industry, there is a growing demand for knowledge transfer of failure analysis from experienced engineers to young engineers. However, failure events are phenomena that occur in a chain reaction, making them difficult for beginners to analyze them. While knowledge graphs, which can describe semantic relationships and structure information is effective in representing failure events, due to their capability of representing the relationships between components, there is much information in KGs, so it is challenging for young engineers to extract and understand sub-graphs from the KG. On the other hand, there is increasing interest in the use of Graph RAG, a type of RAG that combines LLMs and KGs for knowledge management. However, when using the current Graph RAG framework with an existing knowledge graph for automobile failures, several issues arise because it is difficult to generate executable queries for a knowledge graph database which is not constructed by LLMs. To address this, we focused on optimizing the Graph RAG pipeline for existing knowledge graphs. Using an original Q&A dataset, the ROUGE F1 score of the sentences generated by the proposed method showed an average improvement of 157.6% compared to the current method. This highlights the effectiveness of the proposed method for automobile failure analysis.

arXiv:2411.19230

Pre-Training Graph Contrastive Masked Autoencoders are Strong Distillers for EEG

Effectively utilizing extensive unlabeled high-density EEG data to improve performance in scenarios with limited labeled low-density EEG data presents a significant challenge. In this paper, we address this by framing it as a graph transfer learning and knowledge distillation problem. We propose a Unified Pre-trained Graph Contrastive Masked Autoencoder Distiller, named EEG-DisGCMAE, to bridge the gap between unlabeled/labeled and high/low-density EEG data. To fully leverage the abundant unlabeled EEG data, we introduce a novel unified graph self-supervised pre-training paradigm, which seamlessly integrates Graph Contrastive Pre-training and Graph Masked Autoencoder Pre-training. This approach synergistically combines contrastive and generative pre-training techniques by reconstructing contrastive samples and contrasting the reconstructions. For knowledge distillation from high-density to low-density EEG data, we propose a Graph Topology Distillation loss function, allowing a lightweight student model trained on low-density data to learn from a teacher model trained on high-density data, effectively handling missing electrodes through contrastive distillation. To integrate transfer learning and distillation, we jointly pre-train the teacher and student models by contrasting their queries and keys during pre-training, enabling robust distillers for downstream tasks. We demonstrate the effectiveness of our method on four classification tasks across two clinical EEG datasets with abundant unlabeled data and limited labeled data. The experimental results show that our approach significantly outperforms contemporary methods in both efficiency and accuracy.

arXiv:2411.18919공식 코드 · Papers with Code

Federated Continual Graph Learning

Managing evolving graph data presents substantial challenges in storage and privacy, and training graph neural networks (GNNs) on such data often leads to catastrophic forgetting, impairing performance on earlier tasks. Despite existing continual graph learning (CGL) methods mitigating this to some extent, they rely on centralized architectures and ignore the potential of distributed graph databases to leverage collective intelligence. To this end, we propose Federated Continual Graph Learning (FCGL) to adapt GNNs across multiple evolving graphs under storage and privacy constraints. Our empirical study highlights two core challenges: local graph forgetting (LGF), where clients lose prior knowledge when adapting to new tasks, and global expertise conflict (GEC), where the global GNN exhibits sub-optimal performance in both adapting to new tasks and retaining old ones, arising from inconsistent client expertise during server-side parameter aggregation. To address these, we introduce POWER, a framework that preserves experience nodes with maximum local-global coverage locally to mitigate LGF, and leverages pseudo-prototype reconstruction with trajectory-aware knowledge transfer to resolve GEC. Experiments on various graph datasets demonstrate POWER's superiority over federated adaptations of CGL baselines and vision-centric federated continual learning approaches.

arXiv:2411.19525

LokiTalk: Learning Fine-Grained and Generalizable Correspondences to Enhance NeRF-based Talking Head Synthesis

Despite significant progress in talking head synthesis since the introduction of Neural Radiance Fields (NeRF), visual artifacts and high training costs persist as major obstacles to large-scale commercial adoption. We propose that identifying and establishing fine-grained and generalizable correspondences between driving signals and generated results can simultaneously resolve both problems. Here we present LokiTalk, a novel framework designed to enhance NeRF-based talking heads with lifelike facial dynamics and improved training efficiency. To achieve fine-grained correspondences, we introduce Region-Specific Deformation Fields, which decompose the overall portrait motion into lip movements, eye blinking, head pose, and torso movements. By hierarchically modeling the driving signals and their associated regions through two cascaded deformation fields, we significantly improve dynamic accuracy and minimize synthetic artifacts. Furthermore, we propose ID-Aware Knowledge Transfer, a plug-and-play module that learns generalizable dynamic and static correspondences from multi-identity videos, while simultaneously extracting ID-specific dynamic and static features to refine the depiction of individual characters. Comprehensive evaluations demonstrate that LokiTalk delivers superior high-fidelity results and training efficiency compared to previous methods. The code will be released upon acceptance.

논문 ID: document/9702532

Multi-Stage Deep Transfer Learning for EmIoT-enabled Human-Computer Interaction

Emotional Internet of Things (EmIoT), which provides Internet of Things (IoT) devices cognitive and socialization capabilities, has been regarded as a future direction to improve users’ experiences. With the development of intelligent techniques, the requirement of EmIoT is not only sensing the users’ emotional states but also providing emotional feedbacks. Human–computer interaction has been studied to achieve speech interaction with IoT devices. The recent advances in neural text-to-speech (TTS) have made “human parity” synthesized speech possible for IoT-enabled human–computer interaction. Furthermore, emotion control can be achieved by using the emotional codes in a unified model, referred to as emotional TTS (or ETTS for short). Such ETTS models have achieved promising emotional expressiveness using large-scale emotion-annotated English data set; however, they are not practical in IoT environments with other mainstream languages, especially for Chinese. In fact, the limited available large-scale emotion-annotated data set is challenging the development of Chinese ETTS. To address that we propose a multistage deep transfer learning scheme to design a high-quality Chinese ETTS system under a small-scale training corpus to achieve EmIoT in Mandarin environments. In this scheme, the pretrained knowledge from the former stages corresponding to a large-scale neutral English and a medium-scale emotional English corpora is transferred to a Mandarin ETTS model. Thereby, the trained model can achieve high-quality emotional speech with limited available emotional corpus, which is able to serve various EmIoT-oriented applications. The experiments have been conducted to demonstrate the effectiveness and superiority of the proposed model as compared to other counterparts in terms of naturalness and emotional expressiveness. We refer readers to visit our demo Webpage 1 enjoy the synthesized speech samples.

arXiv:2411.18441공식 코드 · Papers with Code

Deep learning-based spatio-temporal fusion for high-fidelity ultra-high-speed x-ray radiography

Full-field ultra-high-speed (UHS) x-ray imaging experiments have been well established to characterize various processes and phenomena. However, the potential of UHS experiments through the joint acquisition of x-ray videos with distinct configurations has not been fully exploited. In this paper, we investigate the use of a deep learning-based spatio-temporal fusion (STF) framework to fuse two complementary sequences of x-ray images and reconstruct the target image sequence with high spatial resolution, high frame rate, and high fidelity. We applied a transfer learning strategy to train the model and compared the peak signal-to-noise ratio (PSNR), average absolute difference (AAD), and structural similarity (SSIM) of the proposed framework on two independent x-ray datasets with those obtained from a baseline deep learning model, a Bayesian fusion framework, and the bicubic interpolation method. The proposed framework outperformed the other methods with various configurations of the input frame separations and image noise levels. With 3 subsequent images from the low resolution (LR) sequence of a 4-time lower spatial resolution and another 2 images from the high resolution (HR) sequence of a 20-time lower frame rate, the proposed approach achieved an average PSNR of 37.57 dB and 35.15 dB, respectively. When coupled with the appropriate combination of high-speed cameras, the proposed approach will enhance the performance and therefore scientific value of the UHS x-ray imaging experiments.

arXiv:2411.17707

A Composite Fault Diagnosis Model for NPPs Based on Bayesian-EfficientNet Module

This article focuses on the faults of important mechanical components such as pumps, valves, and pipelines in the reactor coolant system, main steam system, condensate system, and main feedwater system of nuclear power plants (NPPs). It proposes a composite multi-fault diagnosis model based on Bayesian algorithm and EfficientNet large model using data-driven deep learning fault diagnosis technology. The aim is to evaluate the effectiveness of automatic deep learning-based large model technology through transfer learning in nuclear power plant scenarios.

arXiv:2306.16830

Sampling weights of deep neural networks

We introduce a probability distribution, combined with an efficient sampling algorithm, for weights and biases of fully-connected neural networks. In a supervised learning context, no iterative optimization or gradient computations of internal network parameters are needed to obtain a trained network. The sampling is based on the idea of random feature models. However, instead of a data-agnostic distribution, e.g., a normal distribution, we use both the input and the output training data to sample shallow and deep networks. We prove that sampled networks are universal approximators. For Barron functions, we show that the $L^2$-approximation error of sampled shallow networks decreases with the square root of the number of neurons. Our sampling scheme is invariant to rigid body transformations and scaling of the input data, which implies many popular pre-processing techniques are not required. In numerical experiments, we demonstrate that sampled networks achieve accuracy comparable to iteratively trained ones, but can be constructed orders of magnitude faster. Our test cases involve a classification benchmark from OpenML, sampling of neural operators to represent maps in function spaces, and transfer learning using well-known architectures.

arXiv:2411.18459

What do physics-informed DeepONets learn? Understanding and improving training for scientific computing applications

Physics-informed deep operator networks (DeepONets) have emerged as a promising approach toward numerically approximating the solution of partial differential equations (PDEs). In this work, we aim to develop further understanding of what is being learned by physics-informed DeepONets by assessing the universality of the extracted basis functions and demonstrating their potential toward model reduction with spectral methods. Results provide clarity about measuring the performance of a physics-informed DeepONet through the decays of singular values and expansion coefficients. In addition, we propose a transfer learning approach for improving training for physics-informed DeepONets between parameters of the same PDE as well as across different, but related, PDEs where these models struggle to train well. This approach results in significant error reduction and learned basis functions that are more effective in representing the solution of a PDE.

arXiv:2411.18406

When does a bridge become an aeroplane?

Despite recent advances in population-based structural health monitoring (PBSHM), knowledge transfer between highly-disparate structures (i.e., heterogeneous populations) remains a challenge. It has been proposed that heterogeneous transfer may be accomplished via intermediate structures that bridge the gap in information between the structures of interest. A key aspect of the technique is the idea that by varying parameters such as material properties and geometry, one structure can be continuously morphed into another. The current work demonstrates the development of these interpolating structures, via case studies involving the parameterisation of (and transfer between) a simple, simulated 'bridge' and 'aeroplane'. The facetious question 'When is a bridge not an aeroplane?' has been previously asked in the context of predicting positive transfer based on structural similarity. While the obvious answer to this question is 'Always,' the current work demonstrates that in some cases positive transfer can be achieved between highly-disparate systems.

arXiv:2411.18259공식 코드 · Papers with Code

Transfer Learning for Deep Learning-based Prediction of Lattice Thermal Conductivity

Machine learning promises to accelerate the material discovery by enabling high-throughput prediction of desirable macro-properties from atomic-level descriptors or structures. However, the limited data available about precise values of these properties have been a barrier, leading to predictive models with limited precision or the ability to generalize. This is particularly true of lattice thermal conductivity (LTC): existing datasets of precise (ab initio, DFT-based) computed values are limited to a few dozen materials with little variability. Based on such datasets, we study the impact of transfer learning on both the precision and generalizability of a deep learning model (ParAIsite). We start from an existing model (MEGNet~\cite{Chen2019}) and show that improvements are obtained by fine-tuning a pre-trained version on different tasks. Interestingly, we also show that a much greater improvement is obtained when first fine-tuning it on a large datasets of low-quality approximations of LTC (based on the AGL model) and then applying a second phase of fine-tuning with our high-quality, smaller-scale datasets. The promising results obtained pave the way not only towards a greater ability to explore large databases in search of low thermal conductivity materials but also to methods enabling increasingly precise predictions in areas where quality data are rare.

arXiv:2411.18054

Using different sources of ground truths and transfer learning to improve the generalization of photometric redshift estimation

In this work, we explore methods to improve galaxy redshift predictions by combining different ground truths. Traditional machine learning models rely on training sets with known spectroscopic redshifts, which are precise but only represent a limited sample of galaxies. To make redshift models more generalizable to the broader galaxy population, we investigate transfer learning and directly combining ground truth redshifts derived from photometry and spectroscopy. We use the COSMOS2020 survey to create a dataset, TransferZ, which includes photometric redshift estimates derived from up to 35 imaging filters using template fitting. This dataset spans a wider range of galaxy types and colors compared to spectroscopic samples, though its redshift estimates are less accurate. We first train a base neural network on TransferZ and then refine it using transfer learning on a dataset of galaxies with more precise spectroscopic redshifts (GalaxiesML). In addition, we train a neural network on a combined dataset of TransferZ and GalaxiesML. Both methods reduce bias by $\sim$ 5x, RMS error by $\sim$ 1.5x, and catastrophic outlier rates by 1.3x on GalaxiesML, compared to a baseline trained only on TransferZ. However, we also find a reduction in performance for RMS and bias when evaluated on TransferZ data. Overall, our results demonstrate these approaches can meet cosmological requirements.

arXiv:2411.18021

Can bidirectional encoder become the ultimate winner for downstream applications of foundation models?

Over the past few decades, Artificial Intelligence(AI) has progressed from the initial machine learning stage to the deep learning stage, and now to the stage of foundational models. Foundational models have the characteristics of pre-training, transfer learning, and self-supervised learning, and pre-trained models can be fine-tuned and applied to various downstream tasks. Under the framework of foundational models, models such as Bidirectional Encoder Representations from Transformers(BERT) and Generative Pre-trained Transformer(GPT) have greatly advanced the development of natural language processing(NLP), especially the emergence of many models based on BERT. BERT broke through the limitation of only using one-way methods for language modeling in pre-training by using a masked language model. It can capture bidirectional context information to predict the masked words in the sequence, this can improve the feature extraction ability of the model. This makes the model very useful for downstream tasks, especially for specialized applications. The model using the bidirectional encoder can better understand the domain knowledge and be better applied to these downstream tasks. So we hope to help understand how this technology has evolved and improved model performance in various natural language processing tasks under the background of foundational models and reveal its importance in capturing context information and improving the model's performance on downstream tasks. This article analyzes one-way and bidirectional models based on GPT and BERT and compares their differences based on the purpose of the model. It also briefly analyzes BERT and the improvements of some models based on BERT. The model's performance on the Stanford Question Answering Dataset(SQuAD) and General Language Understanding Evaluation(GLUE) was compared.

arXiv:2411.18115공식 코드 · Papers with Code

Spectral-Spatial Transformer with Active Transfer Learning for Hyperspectral Image Classification

The classification of hyperspectral images (HSI) is a challenging task due to the high spectral dimensionality and limited labeled data typically available for training. In this study, we propose a novel multi-stage active transfer learning (ATL) framework that integrates a Spatial-Spectral Transformer (SST) with an active learning process for efficient HSI classification. Our approach leverages a pre-trained (initially trained) SST model, fine-tuned iteratively on newly acquired labeled samples using an uncertainty-diversity (Spatial-Spectral Neighborhood Diversity) querying mechanism. This mechanism identifies the most informative and diverse samples, thereby optimizing the transfer learning process to reduce both labeling costs and model uncertainty. We further introduce a dynamic freezing strategy, selectively freezing layers of the SST model to minimize computational overhead while maintaining adaptability to spectral variations in new data. One of the key innovations in our work is the self-calibration of spectral and spatial attention weights, achieved through uncertainty-guided active learning. This not only enhances the model's robustness in handling dynamic and disjoint spectral profiles but also improves generalization across multiple HSI datasets. Additionally, we present a diversity-promoting sampling strategy that ensures the selected samples span distinct spectral regions, preventing overfitting to particular spectral classes. Experiments on benchmark HSI datasets demonstrate that the SST-ATL framework significantly outperforms existing CNN and SST-based methods, offering superior accuracy, efficiency, and computational performance. The source code can be accessed at \url{https://github.com/mahmad000/ATL-SST}.

arXiv:2411.17870공식 코드 · Papers with Code

Breast Tumor Classification Using EfficientNet Deep Learning Model

Precise breast cancer classification on histopathological images has the potential to greatly improve the diagnosis and patient outcome in oncology. The data imbalance problem largely stems from the inherent imbalance within medical image datasets, where certain tumor subtypes may appear much less frequently. This constitutes a considerable limitation in biased model predictions that can overlook critical but rare classes. In this work, we adopted EfficientNet, a state-of-the-art convolutional neural network (CNN) model that balances high accuracy with computational cost efficiency. To address data imbalance, we introduce an intensive data augmentation pipeline and cost-sensitive learning, improving representation and ensuring that the model does not overly favor majority classes. This approach provides the ability to learn effectively from rare tumor types, improving its robustness. Additionally, we fine-tuned the model using transfer learning, where weights in the beginning trained on a binary classification task were adopted to multi-class classification, improving the capability to detect complex patterns within the BreakHis dataset. Our results underscore significant improvements in the binary classification performance, achieving an exceptional recall increase for benign cases from 0.92 to 0.95, alongside an accuracy enhancement from 97.35 % to 98.23%. Our approach improved the performance of multi-class tasks from 91.27% with regular augmentation to 94.54% with intensive augmentation, reaching 95.04% with transfer learning. This framework demonstrated substantial gains in precision in the minority classes, such as Mucinous carcinoma and Papillary carcinoma, while maintaining high recall consistently across these critical subtypes, as further confirmed by confusion matrix analysis.

arXiv:2411.18456

Synthetic ECG Generation for Data Augmentation and Transfer Learning in Arrhythmia Classification

Deep learning models need a sufficient amount of data in order to be able to find the hidden patterns in it. It is the purpose of generative modeling to learn the data distribution, thus allowing us to sample more data and augment the original dataset. In the context of physiological data, and more specifically electrocardiogram (ECG) data, given its sensitive nature and expensive data collection, we can exploit the benefits of generative models in order to enlarge existing datasets and improve downstream tasks, in our case, classification of heart rhythm. In this work, we explore the usefulness of synthetic data generated with different generative models from Deep Learning namely Diffweave, Time-Diffusion and Time-VQVAE in order to obtain better classification results for two open source multivariate ECG datasets. Moreover, we also investigate the effects of transfer learning, by fine-tuning a synthetically pre-trained model and then progressively adding increasing proportions of real data. We conclude that although the synthetic samples resemble the real ones, the classification improvement when simply augmenting the real dataset is barely noticeable on individual datasets, but when both datasets are merged the results show an increase across all metrics for the classifiers when using synthetic samples as augmented data. From the fine-tuning results the Time-VQVAE generative model has shown to be superior to the others but not powerful enough to achieve results close to a classifier trained with real data only. In addition, methods and metrics for measuring closeness between synthetic data and the real one have been explored as a side effect of the main research questions of this study.