PyTorchKR Papers

최신 논문

총 9,217편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2411.09539

A Practical Guide to Fine-tuning Language Models with Limited Data

Employing pre-trained Large Language Models (LLMs) has become the de facto standard in Natural Language Processing (NLP) despite their extensive data requirements. Motivated by the recent surge in research focused on training LLMs with limited data, particularly in low-resource domains and languages, this paper surveys recent transfer learning approaches to optimize model performance in downstream tasks where data is scarce. We first address initial and continued pre-training strategies to better leverage prior knowledge in unseen domains and languages. We then examine how to maximize the utility of limited data during fine-tuning and few-shot learning. The final section takes a task-specific perspective, reviewing models and methods suited for different levels of data scarcity. Our goal is to provide practitioners with practical guidelines for overcoming the challenges posed by constrained data while also highlighting promising directions for future research.

arXiv:2411.09286

A Centralized-Distributed Transfer Model for Cross-Domain Recommendation Based on Multi-Source Heterogeneous Transfer Learning

Cross-domain recommendation (CDR) methods are proposed to tackle the sparsity problem in click through rate (CTR) estimation. Existing CDR methods directly transfer knowledge from the source domains to the target domain and ignore the heterogeneities among domains, including feature dimensional heterogeneity and latent space heterogeneity, which may lead to negative transfer. Besides, most of the existing methods are based on single-source transfer, which cannot simultaneously utilize knowledge from multiple source domains to further improve the model performance in the target domain. In this paper, we propose a centralized-distributed transfer model (CDTM) for CDR based on multi-source heterogeneous transfer learning. To address the issue of feature dimension heterogeneity, we build a dual embedding structure: domain specific embedding (DSE) and global shared embedding (GSE) to model the feature representation in the single domain and the commonalities in the global space,separately. To solve the latent space heterogeneity, the transfer matrix and attention mechanism are used to map and combine DSE and GSE adaptively. Extensive offline and online experiments demonstrate the effectiveness of our model.

arXiv:2411.09151

Mono2Stereo: Monocular Knowledge Transfer for Enhanced Stereo Matching

The generalization and performance of stereo matching networks are limited due to the domain gap of the existing synthetic datasets and the sparseness of GT labels in the real datasets. In contrast, monocular depth estimation has achieved significant advancements, benefiting from large-scale depth datasets and self-supervised strategies. To bridge the performance gap between monocular depth estimation and stereo matching, we propose leveraging monocular knowledge transfer to enhance stereo matching, namely Mono2Stereo. We introduce knowledge transfer with a two-stage training process, comprising synthetic data pre-training and real-world data fine-tuning. In the pre-training stage, we design a data generation pipeline that synthesizes stereo training data from monocular images. This pipeline utilizes monocular depth for warping and novel view synthesis and employs our proposed Edge-Aware (EA) inpainting module to fill in missing contents in the generated images. In the fine-tuning stage, we introduce a Sparse-to-Dense Knowledge Distillation (S2DKD) strategy encouraging the distributions of predictions to align with dense monocular depths. This strategy mitigates issues with edge blurring in sparse real-world labels and enhances overall consistency. Experimental results demonstrate that our pre-trained model exhibits strong zero-shot generalization capabilities. Furthermore, domain-specific fine-tuning using our pre-trained model and S2DKD strategy significantly increments in-domain performance. The code will be made available soon.

arXiv:2411.09116

P-MMEval: A Parallel Multilingual Multitask Benchmark for Consistent Evaluation of LLMs

Recent advancements in large language models (LLMs) showcase varied multilingual capabilities across tasks like translation, code generation, and reasoning. Previous assessments often limited their scope to fundamental natural language processing (NLP) or isolated capability-specific tasks. To alleviate this drawback, we aim to present a comprehensive multilingual multitask benchmark. First, we introduce P-MMEval, a large-scale benchmark covering effective fundamental and capability-specialized datasets. Furthermore, P-MMEval delivers consistent language coverage across various datasets and provides parallel samples. Finally, we conduct extensive experiments on representative multilingual model series to compare performances across models and tasks, explore the relationship between multilingual performances and factors such as tasks, model sizes, languages, and prompts, and examine the effectiveness of knowledge transfer from English to other languages. The resulting insights are intended to offer valuable guidance for future research. The dataset is available at https://huggingface.co/datasets/Qwen/P-MMEval.

arXiv:2411.09598

Assessing the Performance of the DINOv2 Self-supervised Learning Vision Transformer Model for the Segmentation of the Left Atrium from MRI Images

Accurate left atrium (LA) segmentation from pre-operative scans is crucial for diagnosing atrial fibrillation, treatment planning, and supporting surgical interventions. While deep learning models are key in medical image segmentation, they often require extensive manually annotated data. Foundation models trained on larger datasets have reduced this dependency, enhancing generalizability and robustness through transfer learning. We explore DINOv2, a self-supervised learning vision transformer trained on natural images, for LA segmentation using MRI. The challenges for LA's complex anatomy, thin boundaries, and limited annotated data make accurate segmentation difficult before & during the image-guided intervention. We demonstrate DINOv2's ability to provide accurate & consistent segmentation, achieving a mean Dice score of .871 & a Jaccard Index of .792 for end-to-end fine-tuning. Through few-shot learning across various data sizes & patient counts, DINOv2 consistently outperforms baseline models. These results suggest that DINOv2 effectively adapts to MRI with limited data, highlighting its potential as a competitive tool for segmentation & encouraging broader use in medical imaging.

arXiv:2411.09269공식 코드 · Papers with Code

Harnessing multiple LLMs for Information Retrieval: A case study on Deep Learning methodologies in Biodiversity publications

Deep Learning (DL) techniques are increasingly applied in scientific studies across various domains to address complex research questions. However, the methodological details of these DL models are often hidden in the unstructured text. As a result, critical information about how these models are designed, trained, and evaluated is challenging to access and comprehend. To address this issue, in this work, we use five different open-source Large Language Models (LLMs): Llama-3 70B, Llama-3.1 70B, Mixtral-8x22B-Instruct-v0.1, Mixtral 8x7B, and Gemma 2 9B in combination with Retrieval-Augmented Generation (RAG) approach to extract and process DL methodological details from scientific publications automatically. We built a voting classifier from the outputs of five LLMs to accurately report DL methodological information. We tested our approach using biodiversity publications, building upon our previous research. To validate our pipeline, we employed two datasets of DL-related biodiversity publications: a curated set of 100 publications from our prior work and a set of 364 publications from the Ecological Informatics journal. Our results demonstrate that the multi-LLM, RAG-assisted pipeline enhances the retrieval of DL methodological information, achieving an accuracy of 69.5% (417 out of 600 comparisons) based solely on textual content from publications. This performance was assessed against human annotators who had access to code, figures, tables, and other supplementary information. Although demonstrated in biodiversity, our methodology is not limited to this field; it can be applied across other scientific domains where detailed methodological reporting is essential for advancing knowledge and ensuring reproducibility. This study presents a scalable and reliable approach for automating information extraction, facilitating better reproducibility and knowledge transfer across studies.

arXiv:2411.09101공식 코드 · Papers with Code

Heuristical Comparison of Vision Transformers Against Convolutional Neural Networks for Semantic Segmentation on Remote Sensing Imagery

Vision Transformers (ViT) have recently brought a new wave of research in the field of computer vision. These models have done particularly well in the field of image classification and segmentation. Research on semantic and instance segmentation has emerged to accelerate with the inception of the new architecture, with over 80\% of the top 20 benchmarks for the iSAID dataset being either based on the ViT architecture or the attention mechanism behind its success. This paper focuses on the heuristic comparison of three key factors of using (or not using) ViT for semantic segmentation of remote sensing aerial images on the iSAID. The experimental results observed during the course of the research were under the scrutinization of the following objectives: 1. Use of weighted fused loss function for the maximum mean Intersection over Union (mIoU) score, Dice score, and minimization or conservation of entropy or class representation, 2. Comparison of transfer learning on Meta's MaskFormer, a ViT-based semantic segmentation model, against generic UNet Convolutional Neural Networks (CNNs) judged over mIoU, Dice scores, training efficiency, and inference time, and 3. What do we lose for what we gain? i.e., the comparison of the two models against current state-of-art segmentation models. We show the use of the novel combined weighted loss function significantly boosts the CNN model's performance capacities as compared to transfer learning the ViT. The code for this implementation can be found on \url{https://github.com/ashimdahal/ViT-vs-CNN-ImageSegmentation}.

arXiv:2411.08374

Federated Graph Learning with Graphless Clients

Federated Graph Learning (FGL) is tasked with training machine learning models, such as Graph Neural Networks (GNNs), for multiple clients, each with its own graph data. Existing methods usually assume that each client has both node features and graph structure of its graph data. In real-world scenarios, however, there exist federated systems where only a part of the clients have such data while other clients (i.e. graphless clients) may only have node features. This naturally leads to a novel problem in FGL: how to jointly train a model over distributed graph data with graphless clients? In this paper, we propose a novel framework FedGLS to tackle the problem in FGL with graphless clients. In FedGLS, we devise a local graph learner on each graphless client which learns the local graph structure with the structure knowledge transferred from other clients. To enable structure knowledge transfer, we design a GNN model and a feature encoder on each client. During local training, the feature encoder retains the local graph structure knowledge together with the GNN model via knowledge distillation, and the structure knowledge is transferred among clients in global update. Our extensive experiments demonstrate the superiority of the proposed FedGLS over five baselines.

arXiv:2411.08335

DEEGITS: Deep Learning based Framework for Measuring Heterogenous Traffic State in Challenging Traffic Scenarios

This paper presents DEEGITS (Deep Learning Based Heterogeneous Traffic State Measurement), a comprehensive framework that leverages state-of-the-art convolutional neural network (CNN) techniques to accurately and rapidly detect vehicles and pedestrians, as well as to measure traffic states in challenging scenarios (i.e., congestion, occlusion). In this study, we enhance the training dataset through data fusion, enabling simultaneous detection of vehicles and pedestrians. Image preprocessing and augmentation are subsequently performed to improve the quality and quantity of the dataset. Transfer learning is applied on the YOLOv8 pretrained model to increase the model's capability to identify a diverse array of vehicles. Optimal hyperparameters are obtained using the Grid Search algorithm, with the Stochastic Gradient Descent (SGD) optimizer outperforming other optimizers under these settings. Extensive experimentation and evaluation demonstrate substantial accuracy within the detection framework, with the model achieving 0.794 [email omitted] on the validation set and 0.786 [email omitted] on the test set, surpassing previous benchmarks on similar datasets. The DeepSORT multi-object tracking algorithm is incorporated to track detected vehicles and pedestrians in this study. Finally, the framework is tested to measure heterogeneous traffic states in mixed traffic conditions. Two locations with differing traffic compositions and congestion levels are selected: one motorized-dominant location with moderate density and one non-motorized-dominant location with higher density. Errors are statistically insignificant for both cases, showing correlations from 0.99 to 0.88 and 0.91 to 0.97 for heterogeneous traffic flow and speed measurements, respectively.

arXiv:2411.08171

Comprehensive and Comparative Analysis between Transfer Learning and Custom Built VGG and CNN-SVM Models for Wildfire Detection

Contemporary Artificial Intelligence (AI) and Machine Learning (ML) research places a significant emphasis on transfer learning, showcasing its transformative potential in enhancing model performance across diverse domains. This paper examines the efficiency and effectiveness of transfer learning in the context of wildfire detection. Three purpose-built models -- Visual Geometry Group (VGG)-7, VGG-10, and Convolutional Neural Network (CNN)-Support Vector Machine(SVM) CNN-SVM -- are rigorously compared with three pretrained models -- VGG-16, VGG-19, and Residual Neural Network (ResNet) ResNet101. We trained and evaluated these models using a dataset that captures the complexities of wildfires, incorporating variables such as varying lighting conditions, time of day, and diverse terrains. The objective is to discern how transfer learning performs against models trained from scratch in addressing the intricacies of the wildfire detection problem. By assessing the performance metrics, including accuracy, precision, recall, and F1 score, a comprehensive understanding of the advantages and disadvantages of transfer learning in this specific domain is obtained. This study contributes valuable insights to the ongoing discourse, guiding future directions in AI and ML research. Keywords: Wildfire prediction, deep learning, machine learning fire, detection

arXiv:2411.08785

Zero-shot Cross-lingual Transfer Learning with Multiple Source and Target Languages for Information Extraction: Language Selection and Adversarial Training

The majority of previous researches addressing multi-lingual IE are limited to zero-shot cross-lingual single-transfer (one-to-one) setting, with high-resource languages predominantly as source training data. As a result, these works provide little understanding and benefit for the realistic goal of developing a multi-lingual IE system that can generalize to as many languages as possible. Our study aims to fill this gap by providing a detailed analysis on Cross-Lingual Multi-Transferability (many-to-many transfer learning), for the recent IE corpora that cover a diverse set of languages. Specifically, we first determine the correlation between single-transfer performance and a wide range of linguistic-based distances. From the obtained insights, a combined language distance metric can be developed that is not only highly correlated but also robust across different tasks and model scales. Next, we investigate the more general zero-shot multi-lingual transfer settings where multiple languages are involved in the training and evaluation processes. Language clustering based on the newly defined distance can provide directions for achieving the optimal cost-performance trade-off in data (languages) selection problem. Finally, a relational-transfer setting is proposed to further incorporate multi-lingual unlabeled data based on adversarial training using the relation induced from the above linguistic distance.

arXiv:2411.08703

MVKTrans: Multi-View Knowledge Transfer for Robust Multiomics Classification

The distinct characteristics of multiomics data, including complex interactions within and across biological layers and disease heterogeneity (e.g., heterogeneity in etiology and clinical symptoms), drive us to develop novel designs to address unique challenges in multiomics prediction. In this paper, we propose the multi-view knowledge transfer learning (MVKTrans) framework, which transfers intra- and inter-omics knowledge in an adaptive manner by reviewing data heterogeneity and suppressing bias transfer, thereby enhancing classification performance. Specifically, we design a graph contrastive module that is trained on unlabeled data to effectively learn and transfer the underlying intra-omics patterns to the supervised task. This unsupervised pretraining promotes learning general and unbiased representations for each modality, regardless of the downstream tasks. In light of the varying discriminative capacities of modalities across different diseases and/or samples, we introduce an adaptive and bi-directional cross-omics distillation module. This module automatically identifies richer modalities and facilitates dynamic knowledge transfer from more informative to less informative omics, thereby enabling a more robust and generalized integration. Extensive experiments on four real biomedical datasets demonstrate the superior performance and robustness of MVKTrans compared to the state-of-the-art. Code and data are available at https://github.com/Yaolab-fantastic/MVKTrans.

arXiv:2411.07508

Feature Interaction Fusion Self-Distillation Network For CTR Prediction

Click-Through Rate (CTR) prediction plays a vital role in recommender systems, online advertising, and search engines. Most of the current approaches model feature interactions through stacked or parallel structures, with some employing knowledge distillation for model compression. However, we observe some limitations with these approaches: (1) In parallel structure models, the explicit and implicit components are executed independently and simultaneously, which leads to insufficient information sharing within the feature set. (2) The introduction of knowledge distillation technology brings about the problems of complex teacher-student framework design and low knowledge transfer efficiency. (3) The dataset and the process of constructing high-order feature interactions contain significant noise, which limits the model's effectiveness. To address these limitations, we propose FSDNet, a CTR prediction framework incorporating a plug-and-play fusion self-distillation module. Specifically, FSDNet forms connections between explicit and implicit feature interactions at each layer, enhancing the sharing of information between different features. The deepest fusion layer is then used as the teacher model, utilizing self-distillation to guide the training of shallow layers. Empirical evaluation across four benchmark datasets validates the framework's efficacy and generalization capabilities. The code is available on https://anonymous.4open.science/r/FSDNet.

arXiv:2411.07483

Quantifying Knowledge Distillation Using Partial Information Decomposition

Knowledge distillation deploys complex machine learning models in resource-constrained environments by training a smaller student model to emulate internal representations of a complex teacher model. However, the teacher's representations can also encode nuisance or additional information not relevant to the downstream task. Distilling such irrelevant information can actually impede the performance of a capacity-limited student model. This observation motivates our primary question: What are the information-theoretic limits of knowledge distillation? To this end, we leverage Partial Information Decomposition to quantify and explain the transferred knowledge and knowledge left to distill for a downstream task. We theoretically demonstrate that the task-relevant transferred knowledge is succinctly captured by the measure of redundant information about the task between the teacher and student. We propose a novel multi-level optimization to incorporate redundant information as a regularizer, leading to our framework of Redundant Information Distillation (RID). RID leads to more resilient and effective distillation under nuisance teachers as it succinctly quantifies task-relevant knowledge rather than simply aligning student and teacher representations.

arXiv:2411.06087

Cross-Domain Transfer Learning using Attention Latent Features for Multi-Agent Trajectory Prediction

With the advancements of sensor hardware, traffic infrastructure and deep learning architectures, trajectory prediction of vehicles has established a solid foundation in intelligent transportation systems. However, existing solutions are often tailored to specific traffic networks at particular time periods. Consequently, deep learning models trained on one network may struggle to generalize effectively to unseen networks. To address this, we proposed a novel spatial-temporal trajectory prediction framework that performs cross-domain adaption on the attention representation of a Transformer-based model. A graph convolutional network is also integrated to construct dynamic graph feature embeddings that accurately model the complex spatial-temporal interactions between the multi-agent vehicles across multiple traffic domains. The proposed framework is validated on two case studies involving the cross-city and cross-period settings. Experimental results show that our proposed framework achieves superior trajectory prediction and domain adaptation performances over the state-of-the-art models.

arXiv:2411.06506

CULL-MT: Compression Using Language and Layer pruning for Machine Translation

Multilingual machine translation models often outperform traditional bilingual models by leveraging translation knowledge transfer. Recent advancements have led to these models supporting hundreds of languages and achieving state-of-the-art results across various translation directions. However, as these models grow larger, their inference operations become increasingly costly. In many use cases, there is no need to support such a wide range of language pairs, as translation is typically needed in only a few selected directions. In this paper, we present CULL-MT, a compression method for machine translation models based on structural layer pruning and selected language directions. Our approach identifies and prunes unimportant layers using a greedy strategy, then mitigates the impact by applying knowledge distillation from the original model along with parameter-efficient fine-tuning. We apply CULL-MT to the NLLB-3.3B and LLaMA3.1-8B-Instruct models. In a multi-way translation scenario (Persian, French, and German to English), we find the NLLB-3.3B model to be robust, allowing 25% of layers to be pruned with only a 0.9 spBLEU drop. However, LLaMA3.1-8B-Instruct is more sensitive, with a 2.0 spBLEU drop after pruning 5 layers.

arXiv:2411.06491

MBL-CPDP: A Multi-objective Bilevel Method for Cross-Project Defect Prediction via Automated Machine Learning

Cross-project defect prediction (CPDP) leverages machine learning (ML) techniques to proactively identify software defects, especially where project-specific data is scarce. However, developing a robust ML pipeline with optimal hyperparameters that effectively use cross-project information and yield satisfactory performance remains challenging. In this paper, we resolve this bottleneck by formulating CPDP as a multi-objective bilevel optimization (MBLO) method, dubbed MBL-CPDP. It comprises two nested problems: the upper-level, a multi-objective combinatorial optimization problem, enhances robustness and efficiency in optimizing ML pipelines, while the lower-level problem is an expensive optimization problem that focuses on tuning their optimal hyperparameters. Due to the high-dimensional search space characterized by feature redundancy and inconsistent data distributions, the upper-level problem combines feature selection, transfer learning, and classification to leverage limited and heterogeneous historical data. Meanwhile, an ensemble learning method is proposed to capture differences in cross-project distribution and generalize across diverse datasets. Finally, a MBLO algorithm is presented to solve this problem while achieving high adaptability effectively. To evaluate the performance of MBL-CPDP, we compare it with five automated ML tools and $50$ CPDP techniques across $20$ projects. Extensive empirical results show that MBL-CPDPoutperforms the comparison methods, demonstrating its superior adaptability and comprehensive performance evaluation capability.

arXiv:2411.06140공식 코드 · Papers with Code

Deep Nonparametric Conditional Independence Tests for Images

Conditional independence tests (CITs) test for conditional dependence between random variables. As existing CITs are limited in their applicability to complex, high-dimensional variables such as images, we introduce deep nonparametric CITs (DNCITs). The DNCITs combine embedding maps, which extract feature representations of high-dimensional variables, with nonparametric CITs applicable to these feature representations. For the embedding maps, we derive general properties on their parameter estimators to obtain valid DNCITs and show that these properties include embedding maps learned through (conditional) unsupervised or transfer learning. For the nonparametric CITs, appropriate tests are selected and adapted to be applicable to feature representations. Through simulations, we investigate the performance of the DNCITs for different embedding maps and nonparametric CITs under varying confounder dimensions and confounder relationships. We apply the DNCITs to brain MRI scans and behavioral traits, given confounders, of healthy individuals from the UK Biobank (UKB), confirming null results from a number of ambiguous personality neuroscience studies with a larger data set and with our more powerful tests. In addition, in a confounder control study, we apply the DNCITs to brain MRI scans and a confounder set to test for sufficient confounder control, leading to a potential reduction in the confounder dimension under improved confounder control compared to existing state-of-the-art confounder control studies for the UKB. Finally, we provide an R package implementing the DNCITs.

arXiv:2411.07239

DeepONet as a Multi-Operator Extrapolation Model: Distributed Pretraining with Physics-Informed Fine-Tuning

We propose a novel fine-tuning method to achieve multi-operator learning through training a distributed neural operator with diverse function data and then zero-shot fine-tuning the neural network using physics-informed losses for downstream tasks. Operator learning effectively approximates solution operators for PDEs and various PDE-related problems, yet it often struggles to generalize to new tasks. To address this, we investigate fine-tuning a pretrained model, while carefully selecting an initialization that enables rapid adaptation to new tasks with minimal data. Our approach combines distributed learning to integrate data from various operators in pre-training, while physics-informed methods enable zero-shot fine-tuning, minimizing the reliance on downstream data. We investigate standard fine-tuning and Low-Rank Adaptation fine-tuning, applying both to train complex nonlinear target operators that are difficult to learn only using random initialization. Through comprehensive numerical examples, we demonstrate the advantages of our approach, showcasing significant improvements in accuracy. Our findings provide a robust framework for advancing multi-operator learning and highlight the potential of transfer learning techniques in this domain.

arXiv:2411.06397

A Hybrid Approach for COVID-19 Detection: Combining Wasserstein GAN with Transfer Learning

COVID-19 is extremely contagious and its rapid growth has drawn attention towards its early diagnosis. Early diagnosis of COVID-19 enables healthcare professionals and government authorities to break the chain of transition and flatten the epidemic curve. With the number of cases accelerating across the developed world, COVID-19 induced Viral Pneumonia cases is a big challenge. Overlapping of COVID-19 cases with Viral Pneumonia and other lung infections with limited dataset and long training hours is a serious problem to cater. Limited amount of data often results in over-fitting models and due to this reason, model does not predict generalized results. To fill this gap, we proposed GAN-based approach to synthesize images which later fed into the deep learning models to classify images of COVID-19, Normal, and Viral Pneumonia. Specifically, customized Wasserstein GAN is proposed to generate 19% more Chest X-ray images as compare to the real images. This expanded dataset is then used to train four proposed deep learning models: VGG-16, ResNet-50, GoogLeNet and MNAST. The result showed that expanded dataset utilized deep learning models to deliver high classification accuracies. In particular, VGG-16 achieved highest accuracy of 99.17% among all four proposed schemes. Rest of the models like ResNet-50, GoogLeNet and MNAST delivered 93.9%, 94.49% and 97.75% testing accuracies respectively. Later, the efficiency of these models is compared with the state of art models on the basis of accuracy. Further, our proposed models can be applied to address the issue of scant datasets for any problem of image analysis.