arXiv subjects

Da Ma

Publications and source records attributed to Da Ma.

36 records · Page 2Linked to original sources

arXiv2024

Evolving Subnetwork Training for Large Language Models

Large language models have ushered in a new era of artificial intelligence research. However, their substantial training costs hinder further development and widespread adoption. In this paper, inspired by the redundancy in the parameters of large language models, we propose a novel training paradigm: Evolving Subnetwork Training (EST). EST samples subnetworks from the layers of the large language model and from commonly used modules within each layer, Multi-Head Attention (MHA) and Multi-Layer Perceptron (MLP). By gradually increasing the size of the subnetworks during the training process, EST can save the cost of training. We apply EST to train GPT2 model and TinyLlama model, resulting in 26.7\% FLOPs saving for GPT2 and 25.0\% for TinyLlama without an increase in loss on the pre-training dataset. Moreover, EST leads to performance improvements in downstream tasks, indicating that it benefits generalization. Additionally, we provide intuitive theoretical studies based on training dynamics and Dropout theory to ensure the feasibility of EST. Our code is available at https://github.com/OpenDFM/EST.

cs.CL↗

arXiv2024

Sparsity-Accelerated Training for Large Language Models

Large language models (LLMs) have demonstrated proficiency across various natural language processing (NLP) tasks but often require additional training, such as continual pre-training and supervised fine-tuning. However, the costs associated with this, primarily due to their large parameter count, remain high. This paper proposes leveraging \emph{sparsity} in pre-trained LLMs to expedite this training process. By observing sparsity in activated neurons during forward iterations, we identify the potential for computational speed-ups by excluding inactive neurons. We address associated challenges by extending existing neuron importance evaluation metrics and introducing a ladder omission rate scheduler. Our experiments on Llama-2 demonstrate that Sparsity-Accelerated Training (SAT) achieves comparable or superior performance to standard training while significantly accelerating the process. Specifically, SAT achieves a $45\%$ throughput improvement in continual pre-training and saves $38\%$ training time in supervised fine-tuning in practice. It offers a simple, hardware-agnostic, and easily deployable framework for additional LLM training. Our code is available at https://github.com/OpenDFM/SAT.

cs.CL↗

arXiv2024

Hierarchical Multimodal Pre-training for Visually Rich Webpage Understanding

The growing prevalence of visually rich documents, such as webpages and scanned/digital-born documents (images, PDFs, etc.), has led to increased interest in automatic document understanding and information extraction across academia and industry. Although various document modalities, including image, text, layout, and structure, facilitate human information retrieval, the interconnected nature of these modalities presents challenges for neural networks. In this paper, we introduce WebLM, a multimodal pre-training network designed to address the limitations of solely modeling text and structure modalities of HTML in webpages. Instead of processing document images as unified natural images, WebLM integrates the hierarchical structure of document images to enhance the understanding of markup-language-based documents. Additionally, we propose several pre-training tasks to model the interaction among text, structure, and image modalities effectively. Empirical results demonstrate that the pre-trained WebLM significantly surpasses previous state-of-the-art pre-trained models across several webpage understanding tasks. The pre-trained models and code are available at https://github.com/X-LANCE/weblm.

cs.CL↗

arXiv2023

ASTormer: An AST Structure-aware Transformer Decoder for Text-to-SQL

Text-to-SQL aims to generate an executable SQL program given the user utterance and the corresponding database schema. To ensure the well-formedness of output SQLs, one prominent approach adopts a grammar-based recurrent decoder to produce the equivalent SQL abstract syntax tree (AST). However, previous methods mainly utilize an RNN-series decoder, which 1) is time-consuming and inefficient and 2) introduces very few structure priors. In this work, we propose an AST structure-aware Transformer decoder (ASTormer) to replace traditional RNN cells. The structural knowledge, such as node types and positions in the tree, is seamlessly incorporated into the decoder via both absolute and relative position embeddings. Besides, the proposed framework is compatible with different traversing orders even considering adaptive node selection. Extensive experiments on five text-to-SQL benchmarks demonstrate the effectiveness and efficiency of our structured decoder compared to competitive baselines.

cs.CL↗

arXiv2023

Anomalous skew-scattering nonlinear Hall effect and chiral photocurrents in {\it PT}-symmetric antiferromagnets

Berry curvature and skew-scattering play central roles in determining both the linear and nonlinear anomalous Hall effects. Yet in {\it PT}-symmetric antiferromagnetic metals, Hall effects from either intrinsic Berry curvature mediated anomalous velocity or the conventional skew-scattering process individually vanish. Here we reveal an unexpected nonlinear Hall effect that relies on both Berry curvature and skew-scattering working in cooperation. This anomalous skew-scattering nonlinear Hall effect (ASN) is {\it PT}-even and dominates the low-frequency nonlinear Hall effect for {\it PT}-symmetric antiferromagnetic metals. Surprisingly, we find that in addition to its Hall response, ASN produces helicity dependent photocurrents, in contrast to other known {\it PT}-even nonlinearities in metals which are helicity blind. This characteristic enables to isolate ASN and establishes new photocurrent tools to interrogate the antiferromagnetic order of {\it PT}-symmetric metals.

cond-mat.mes-hall↗

arXiv2023

Spectral Bandwidth Recovery of Optical Coherence Tomography Images using Deep Learning

Optical coherence tomography (OCT) captures cross-sectional data and is used for the screening, monitoring, and treatment planning of retinal diseases. Technological developments to increase the speed of acquisition often results in systems with a narrower spectral bandwidth, and hence a lower axial resolution. Traditionally, image-processing-based techniques have been utilized to reconstruct subsampled OCT data and more recently, deep-learning-based methods have been explored. In this study, we simulate reduced axial scan (A-scan) resolution by Gaussian windowing in the spectral domain and investigate the use of a learning-based approach for image feature reconstruction. In anticipation of the reduced resolution that accompanies wide-field OCT systems, we build upon super-resolution techniques to explore methods to better aid clinicians in their decision-making to improve patient outcomes, by reconstructing lost features using a pixel-to-pixel approach with an altered super-resolution generative adversarial network (SRGAN) architecture.

eess.IV↗

arXiv2022

DFM: Dialogue Foundation Model for Universal Large-Scale Dialogue-Oriented Task Learning

Building a universal conversational agent has been a long-standing goal of the dialogue research community. Most previous works only focus on a small set of dialogue tasks. In this work, we aim to build a unified dialogue foundation model (DFM) which can be used to solve massive diverse dialogue tasks. To achieve this goal, a large-scale well-annotated dialogue dataset with rich task diversity (DialogZoo) is collected. We introduce a framework to unify all dialogue tasks and propose novel auxiliary self-supervised tasks to achieve stable training of DFM on the highly diverse large scale DialogZoo corpus. Experiments show that, compared with models of the same size, DFM can achieve state-of-the-art or competitive performance on very rich cross-domain downstream dialogue tasks. This demonstrates that DFM largely extends the ability of unified dialogue pre-trained model.

cs.CL↗

arXiv2022

Segmentation-guided Domain Adaptation and Data Harmonization of Multi-device Retinal Optical Coherence Tomography using Cycle-Consistent Generative Adversarial Networks

Optical Coherence Tomography(OCT) is a non-invasive technique capturing cross-sectional area of the retina in micro-meter resolutions. It has been widely used as a auxiliary imaging reference to detect eye-related pathology and predict longitudinal progression of the disease characteristics. Retina layer segmentation is one of the crucial feature extraction techniques, where the variations of retinal layer thicknesses and the retinal layer deformation due to the presence of the fluid are highly correlated with multiple epidemic eye diseases like Diabetic Retinopathy(DR) and Age-related Macular Degeneration (AMD). However, these images are acquired from different devices, which have different intensity distribution, or in other words, belong to different imaging domains. This paper proposes a segmentation-guided domain-adaptation method to adapt images from multiple devices into single image domain, where the state-of-art pre-trained segmentation model is available. It avoids the time consumption of manual labelling for the upcoming new dataset and the re-training of the existing network. The semantic consistency and global feature consistency of the network will minimize the hallucination effect that many researchers reported regarding Cycle-Consistent Generative Adversarial Networks(CycleGAN) architecture.

eess.IV↗

arXiv2022

Predicting Time-to-conversion for Dementia of Alzheimer's Type using Multi-modal Deep Survival Analysis

Dementia of Alzheimer's Type (DAT) is a complex disorder influenced by numerous factors, but it is unclear how each factor contributes to disease progression. An in-depth examination of these factors may yield an accurate estimate of time-to-conversion to DAT for patients at various disease stages. We used 401 subjects with 63 features from MRI, genetic, and CDC (Cognitive tests, Demographic, and CSF) data modalities in the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. We used a deep learning-based survival analysis model that extends the classic Cox regression model to predict time-to-conversion to DAT. Our findings showed that genetic features contributed the least to survival analysis, while CDC features contributed the most. Combining MRI and genetic features improved survival prediction over using either modality alone, but adding CDC to any combination of features only worked as well as using only CDC features. Consequently, our study demonstrated that using the current clinical procedure, which includes gathering cognitive test results, can outperform survival analysis results produced using costly genetic or CSF data.

cs.LG↗

arXiv2022

Machine Learning Based Multimodal Neuroimaging Genomics Dementia Score for Predicting Future Conversion to Alzheimer's Disease

Background: The increasing availability of databases containing both magnetic resonance imaging (MRI) and genetic data allows researchers to utilize multimodal data to better understand the characteristics of dementia of Alzheimer's type (DAT). Objective: The goal of this study was to develop and analyze novel biomarkers that can help predict the development and progression of DAT. Methods: We used feature selection and ensemble learning classifier to develop an image/genotype-based DAT score that represents a subject's likelihood of developing DAT in the future. Three feature types were used: MRI only, genetic only, and combined multimodal data. We used a novel data stratification method to better represent different stages of DAT. Using a pre-defined 0.5 threshold on DAT scores, we predicted whether or not a subject would develop DAT in the future. Results: Our results on Alzheimer's Disease Neuroimaging Initiative (ADNI) database showed that dementia scores using genetic data could better predict future DAT progression for currently normal control subjects (Accuracy=0.857) compared to MRI (Accuracy=0.143), while MRI can better characterize subjects with stable mild cognitive impairment (Accuracy=0.614) compared to genetics (Accuracy=0.356). Combining MRI and genetic data showed improved classification performance in the remaining stratified groups. Conclusion: MRI and genetic data can contribute to DAT prediction in different ways. MRI data reflects anatomical changes in the brain, while genetic data can detect the risk of DAT progression prior to the symptomatic onset. Combining information from multimodal data in the right way can improve prediction performance.

cs.LG↗

arXiv2021

Differential Diagnosis of Frontotemporal Dementia and Alzheimer's Disease using Generative Adversarial Network

Frontotemporal dementia and Alzheimer's disease are two common forms of dementia and are easily misdiagnosed as each other due to their similar pattern of clinical symptoms. Differentiating between the two dementia types is crucial for determining disease-specific intervention and treatment. Recent development of Deep-learning-based approaches in the field of medical image computing are delivering some of the best performance for many binary classification tasks, although its application in differential diagnosis, such as neuroimage-based differentiation for multiple types of dementia, has not been explored. In this study, a novel framework was proposed by using the Generative Adversarial Network technique to distinguish FTD, AD and normal control subjects, using volumetric features extracted at coarse-to-fine structural scales from Magnetic Resonance Imaging scans. Experiments of 10-folds cross-validation on 1,954 images achieved high accuracy. With the proposed framework, we have demonstrated that the combination of multi-scale structural features and synthetic data augmentation based on generative adversarial network can improve the performance of challenging tasks such as differentiating Dementia sub-types.

eess.IV↗

arXiv2021

Comprehensive Validation of Automated Whole Body Skeletal Muscle, Adipose Tissue, and Bone Segmentation from 3D CT images for Body Composition Analysis: Towards Extended Body Composition

The latest advances in computer-assisted precision medicine are making it feasible to move from population-wide models that are useful to discover aggregate patterns that hold for group-based analysis to patient-specific models that can drive patient-specific decisions with regard to treatment choices, and predictions of outcomes of treatment. Body Composition is recognized as an important driver and risk factor for a wide variety of diseases, as well as a predictor of individual patient-specific clinical outcomes to treatment choices or surgical interventions. 3D CT images are routinely acquired in the oncological worklows and deliver accurate rendering of internal anatomy and therefore can be used opportunistically to assess the amount of skeletal muscle and adipose tissue compartments. Powerful tools of artificial intelligence such as deep learning are making it feasible now to segment the entire 3D image and generate accurate measurements of all internal anatomy. These will enable the overcoming of the severe bottleneck that existed previously, namely, the need for manual segmentation, which was prohibitive to scale to the hundreds of 2D axial slices that made up a 3D volumetric image. Automated tools such as presented here will now enable harvesting whole-body measurements from 3D CT or MRI images, leading to a new era of discovery of the drivers of various diseases based on individual tissue, organ volume, shape, and functional status. These measurements were hitherto unavailable thereby limiting the field to a very small and limited subset. These discoveries and the potential to perform individual image segmentation with high speed and accuracy are likely to lead to the incorporation of these 3D measures into individual specific treatment planning models related to nutrition, aging, chemotoxicity, surgery and survival after the onset of a major disease such as cancer.

cs.CV↗

arXiv2021

Domain Adaptation via CycleGAN for Retina Segmentation in Optical Coherence Tomography

With the FDA approval of Artificial Intelligence (AI) for point-of-care clinical diagnoses, model generalizability is of the utmost importance as clinical decision-making must be domain-agnostic. A method of tackling the problem is to increase the dataset to include images from a multitude of domains; while this technique is ideal, the security requirements of medical data is a major limitation. Additionally, researchers with developed tools benefit from the addition of open-sourced data, but are limited by the difference in domains. Herewith, we investigated the implementation of a Cycle-Consistent Generative Adversarial Networks (CycleGAN) for the domain adaptation of Optical Coherence Tomography (OCT) volumes. This study was done in collaboration with the Biomedical Optics Research Group and Functional & Anatomical Imaging & Shape Analysis Lab at Simon Fraser University. In this study, we investigated a learning-based approach of adapting the domain of a publicly available dataset, UK Biobank dataset (UKB). To evaluate the performance of domain adaptation, we utilized pre-existing retinal layer segmentation tools developed on a different set of RETOUCH OCT data. This study provides insight on state-of-the-art tools for domain adaptation compared to traditional processing techniques as well as a pipeline for adapting publicly available retinal data to the domains previously used by our collaborators.

eess.IV↗

arXiv2021

Decoupled Dialogue Modeling and Semantic Parsing for Multi-Turn Text-to-SQL

Recently, Text-to-SQL for multi-turn dialogue has attracted great interest. Here, the user input of the current turn is parsed into the corresponding SQL query of the appropriate database, given all previous dialogue history. Current approaches mostly employ end-to-end models and consequently face two challenges. First, dialogue history modeling and Text-to-SQL parsing are implicitly combined, hence it is hard to carry out interpretable analysis and obtain targeted improvement. Second, SQL annotation of multi-turn dialogue is very expensive, leading to training data sparsity. In this paper, we propose a novel decoupled multi-turn Text-to-SQL framework, where an utterance rewrite model first explicitly solves completion of dialogue context, and then a single-turn Text-to-SQL parser follows. A dual learning approach is also proposed for the utterance rewrite model to address the data sparsity problem. Compared with end-to-end approaches, the proposed decoupled method can achieve excellent performance without any annotated in-domain data. With just a few annotated rewrite cases, the decoupled method outperforms the released state-of-the-art end-to-end models on both SParC and CoSQL datasets.

cs.CL↗

arXiv2020

Unconventional Hall Effect induced by Berry Curvature

Berry phase and Berry curvature play a key role in the development of topology in physics and do contribute to the transport properties in solid state systems. In this paper, we report the finding of novel nonzero Hall effect in topological material ZrTe5 flakes when in-plane magnetic field is parallel and perpendicular to the current. Surprisingly, both symmetric and antisymmetric components with respect to magnetic field are detected in the in-plane Hall resistivity. Further theoretical analysis suggests that the magnetotransport properties originate from the anomalous velocity induced by Berry curvature in a tilted Weyl semimetal. Our work not only enriches the Hall family but also provides new insights into the Berry phase effect in topological materials.

cond-mat.mes-hall↗

arXiv2019

Cascaded Deep Neural Networks for Retinal Layer Segmentation of Optical Coherence Tomography with Fluid Presence

Optical coherence tomography (OCT) is a non-invasive imaging technology which can provide micrometer-resolution cross-sectional images of the inner structures of the eye. It is widely used for the diagnosis of ophthalmic diseases with retinal alteration, such as layer deformation and fluid accumulation. In this paper, a novel framework was proposed to segment retinal layers with fluid presence. The main contribution of this study is two folds: 1) we developed a cascaded network framework to incorporate the prior structural knowledge; 2) we proposed a novel deep neural network based on U-Net and fully convolutional network, termed LF-UNet. Cross validation experiments proved that the proposed LF-UNet has superior performance comparing with the state-of-the-art methods, and incorporating the relative distance map structural prior information could further improve the performance regardless the network.

eess.IV↗

arXiv2019

Grey matter sublayer thickness estimation in themouse cerebellum

The cerebellar grey matter morphology is an important feature to study neurodegenerative diseases such as Alzheimer's disease or Down's syndrome. Its volume or thickness is commonly used as a surrogate imaging biomarker for such diseases. Most studies about grey matter thickness estimation focused on the cortex, and little attention has been drawn on the morphology of the cerebellum. Using ex vivo high-resolution MRI, it is now possible to visualise the different cell layers in the mouse cerebellum. In this work, we introduce a framework to extract the Purkinje layer within the grey matter, enabling the estimation of the thickness of the cerebellar grey matter, the granular layer and molecular layer from gadolinium-enhanced ex vivo mouse brain MRI. Application to mouse model of Down's syndrome found reduced cortical and layer thicknesses in the transchromosomic group.

cs.CV↗

arXiv2017

Kondo Effect with Weyl Semimetal Fermi Arcs

We investigate the Kondo effect of the Fermi arcs in a time-reversal-invariant Weyl semimetal with the variational method. To show the consequence brought out by the nontrivial spin texture, we calculate the spatial spin-spin correlation functions. The correlation functions exhibit high anisotropy. The diagonal correlation functions are dominated by the antiferromagnetic correlation while the off-diagonal part has more complicated pattern. The correlation functions obey the same symmetry as the spin texture. Tuning chemical potential changes the pattern of the correlation functions and the correlation length. The correlation functions of the Weyl semimetal Fermi arcs and that from a Dirac semimetal show discrepancy.

cond-mat.str-el↗