arXiv subjects

Min Xu

Publications and source records attributed to Min Xu.

At least 181 records · Page 10Linked to original sources

arXiv2023

PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel

It is widely acknowledged that large models have the potential to deliver superior performance across a broad range of domains. Despite the remarkable progress made in the field of machine learning systems research, which has enabled the development and exploration of large models, such abilities remain confined to a small group of advanced users and industry leaders, resulting in an implicit technical barrier for the wider community to access and leverage these technologies. In this paper, we introduce PyTorch Fully Sharded Data Parallel (FSDP) as an industry-grade solution for large model training. FSDP has been closely co-designed with several key PyTorch core components including Tensor implementation, dispatcher system, and CUDA memory caching allocator, to provide non-intrusive user experiences and high training efficiency. Additionally, FSDP natively incorporates a range of techniques and settings to optimize resource utilization across a variety of hardware configurations. The experimental results demonstrate that FSDP is capable of achieving comparable performance to Distributed Data Parallel while providing support for significantly larger models with near-linear scalability in terms of TFLOPS.

cs.DC↗

arXiv2023

Choosing the $p$ in $L_p$ loss: rate adaptivity on the symmetric location problem

Given univariate random variables $Y_1, \ldots, Y_n$ with the $\text{Uniform}(θ_0 - 1, θ_0 + 1)$ distribution, the sample midrange $\frac{Y_{(n)}+Y_{(1)}}{2}$ is the MLE for $θ_0$ and estimates $θ_0$ with error of order $1/n$, which is much smaller compared with the $1/\sqrt{n}$ error rate of the usual sample mean estimator. However, the sample midrange performs poorly when the data has say the Gaussian $N(θ_0, 1)$ distribution, with an error rate of $1/\sqrt{\log n}$. In this paper, we propose an estimator of the location $θ_0$ with a rate of convergence that can, in many settings, adapt to the underlying distribution which we assume to be symmetric around $θ_0$ but is otherwise unknown. When the underlying distribution is compactly supported, we show that our estimator attains a rate of convergence of $n^{-\frac{1}α}$ up to polylog factors, where the rate parameter $α$ can take on any value in $(0, 2]$ and depends on the moments of the underlying distribution. Our estimator is formed by the $\ell^γ$-center of the data, for a $γ\geq2$ chosen in a data-driven way -- by minimizing a criterion motivated by the asymptotic variance. Our approach can be directly applied to the regression setting where $θ_0$ is a function of observed features and motivates the use of $\ell^γ$ loss function for $γ> 2$ in certain settings.

math.ST↗

arXiv2023

DenseMP: Unsupervised Dense Pre-training for Few-shot Medical Image Segmentation

Few-shot medical image semantic segmentation is of paramount importance in the domain of medical image analysis. However, existing methodologies grapple with the challenge of data scarcity during the training phase, leading to over-fitting. To mitigate this issue, we introduce a novel Unsupervised Dense Few-shot Medical Image Segmentation Model Training Pipeline (DenseMP) that capitalizes on unsupervised dense pre-training. DenseMP is composed of two distinct stages: (1) segmentation-aware dense contrastive pre-training, and (2) few-shot-aware superpixel guided dense pre-training. These stages collaboratively yield a pre-trained initial model specifically designed for few-shot medical image segmentation, which can subsequently be fine-tuned on the target dataset. Our proposed pipeline significantly enhances the performance of the widely recognized few-shot segmentation model, PA-Net, achieving state-of-the-art results on the Abd-CT and Abd-MRI datasets. Code will be released after acceptance.

cs.CV↗

arXiv2023

BiCro: Noisy Correspondence Rectification for Multi-modality Data via Bi-directional Cross-modal Similarity Consistency

As one of the most fundamental techniques in multimodal learning, cross-modal matching aims to project various sensory modalities into a shared feature space. To achieve this, massive and correctly aligned data pairs are required for model training. However, unlike unimodal datasets, multimodal datasets are extremely harder to collect and annotate precisely. As an alternative, the co-occurred data pairs (e.g., image-text pairs) collected from the Internet have been widely exploited in the area. Unfortunately, the cheaply collected dataset unavoidably contains many mismatched data pairs, which have been proven to be harmful to the model's performance. To address this, we propose a general framework called BiCro (Bidirectional Cross-modal similarity consistency), which can be easily integrated into existing cross-modal matching models and improve their robustness against noisy data. Specifically, BiCro aims to estimate soft labels for noisy data pairs to reflect their true correspondence degree. The basic idea of BiCro is motivated by that -- taking image-text matching as an example -- similar images should have similar textual descriptions and vice versa. Then the consistency of these two similarities can be recast as the estimated soft labels to train the matching model. The experiments on three popular cross-modal matching datasets demonstrate that our method significantly improves the noise-robustness of various matching models, and surpass the state-of-the-art by a clear margin.

cs.CV↗

arXiv2023

Vital Sign Monitoring in Dynamic Environment via mmWave Radar and Camera Fusion

Contact-free vital sign monitoring, which uses wireless signals for recognizing human vital signs (i.e, breath and heartbeat), is an attractive solution to health and security. However, the subject's body movement and the change in actual environments can result in inaccurate frequency estimation of heartbeat and respiratory. In this paper, we propose a robust mmWave radar and camera fusion system for monitoring vital signs, which can perform consistently well in dynamic scenarios, e.g., when some people move around the subject to be tracked, or a subject waves his/her arms and marches on the spot. Three major processing modules are developed in the system, to enable robust sensing. Firstly, we utilize a camera to assist a mmWave radar to accurately localize the subjects of interest. Secondly, we exploit the calculated subject position to form transmitting and receiving beamformers, which can improve the reflected power from the targets and weaken the impact of dynamic interference. Thirdly, we propose a weighted multi-channel Variational Mode Decomposition (WMC-VMD) algorithm to separate the weak vital sign signals from the dynamic ones due to subject's body movement. Experimental results show that, the 90${^{th}}$ percentile errors in respiration rate (RR) and heartbeat rate (HR) are less than 0.5 RPM (respirations per minute) and 6 BPM (beats per minute), respectively.

eess.SP↗

arXiv2023

Multimodal Hyperspectral Image Classification via Interconnected Fusion

Existing multiple modality fusion methods, such as concatenation, summation, and encoder-decoder-based fusion, have recently been employed to combine modality characteristics of Hyperspectral Image (HSI) and Light Detection And Ranging (LiDAR). However, these methods consider the relationship of HSI-LiDAR signals from limited perspectives. More specifically, they overlook the contextual information across modalities of HSI and LiDAR and the intra-modality characteristics of LiDAR. In this paper, we provide a new insight into feature fusion to explore the relationships across HSI and LiDAR modalities comprehensively. An Interconnected Fusion (IF) framework is proposed. Firstly, the center patch of the HSI input is extracted and replicated to the size of the HSI input. Then, nine different perspectives in the fusion matrix are generated by calculating self-attention and cross-attention among the replicated center patch, HSI input, and corresponding LiDAR input. In this way, the intra- and inter-modality characteristics can be fully exploited, and contextual information is considered in both intra-modality and inter-modality manner. These nine interrelated elements in the fusion matrix can complement each other and eliminate biases, which can generate a multi-modality representation for classification accurately. Extensive experiments have been conducted on three widely used datasets: Trento, MUUFL, and Houston. The IF framework achieves state-of-the-art results on these datasets compared to existing approaches.

cs.CV↗

arXiv2023

CoNIC Challenge: Pushing the Frontiers of Nuclear Detection, Segmentation, Classification and Counting

Nuclear detection, segmentation and morphometric profiling are essential in helping us further understand the relationship between histology and patient outcome. To drive innovation in this area, we setup a community-wide challenge using the largest available dataset of its kind to assess nuclear segmentation and cellular composition. Our challenge, named CoNIC, stimulated the development of reproducible algorithms for cellular recognition with real-time result inspection on public leaderboards. We conducted an extensive post-challenge analysis based on the top-performing models using 1,658 whole-slide images of colon tissue. With around 700 million detected nuclei per model, associated features were used for dysplasia grading and survival analysis, where we demonstrated that the challenge's improvement over the previous state-of-the-art led to significant boosts in downstream performance. Our findings also suggest that eosinophils and neutrophils play an important role in the tumour microevironment. We release challenge models and WSI-level results to foster the development of further methods for biomarker discovery.

cs.CV↗

arXiv2023

BenchDirect: A Directed Language Model for Compiler Benchmarks

The exponential increase of hardware-software complexity has made it impossible for compiler engineers to find the right optimization heuristics manually. Predictive models have been shown to find near optimal heuristics with little human effort but they are limited by a severe lack of diverse benchmarks to train on. Generative AI has been used by researchers to synthesize benchmarks into existing datasets. However, the synthetic programs are short, exceedingly simple and lacking diversity in their features. We develop BenchPress, the first ML compiler benchmark generator that can be directed within source code feature representations. BenchPress synthesizes executable functions by infilling code that conditions on the program's left and right context. BenchPress uses active learning to introduce new benchmarks with unseen features into the dataset of Grewe's et al. CPU vs GPU heuristic, improving its acquired performance by 50%. BenchPress targets features that has been impossible for other synthesizers to reach. In 3 feature spaces, we outperform human-written code from GitHub, CLgen, CLSmith and the SRCIROR mutator in targeting the features of Rodinia benchmarks. BenchPress steers generation with beam search over a feature-agnostic language model. We improve this with BenchDirect which utilizes a directed LM that infills programs by jointly observing source code context and the compiler features that are targeted. BenchDirect achieves up to 36% better accuracy in targeting the features of Rodinia benchmarks, it is 1.8x more likely to give an exact match and it speeds up execution time by up to 72% compared to BenchPress. Both our models produce code that is difficult to distinguish from human-written code. We conduct a Turing test which shows our models' synthetic benchmarks are labelled as 'human-written' as often as human-written code from GitHub.

cs.LG↗

arXiv2023

Dataset Pruning: Reducing Training Data by Examining Generalization Influence

The great success of deep learning heavily relies on increasingly larger training data, which comes at a price of huge computational and infrastructural costs. This poses crucial questions that, do all training data contribute to model's performance? How much does each individual training sample or a sub-training-set affect the model's generalization, and how to construct the smallest subset from the entire training data as a proxy training set without significantly sacrificing the model's performance? To answer these, we propose dataset pruning, an optimization-based sample selection method that can (1) examine the influence of removing a particular set of training samples on model's generalization ability with theoretical guarantee, and (2) construct the smallest subset of training data that yields strictly constrained generalization gap. The empirically observed generalization gap of dataset pruning is substantially consistent with our theoretical expectations. Furthermore, the proposed method prunes 40% training examples on the CIFAR-10 dataset, halves the convergence time with only 1.3% test accuracy decrease, which is superior to previous score-based sample selection methods.

cs.LG↗

arXiv2023

Root and community inference on the latent growth process of a network

Many existing statistical models for networks overlook the fact that many real world networks are formed through a growth process. To address this, we introduce the PAPER (Preferential Attachment Plus Erdős--Rényi) model for random networks, where we let a random network G be the union of a preferential attachment (PA) tree T and additional Erdős--Rényi (ER) random edges. The PA tree component captures the underlying growth/recruitment process of a network where vertices and edges are added sequentially, while the ER component can be regarded as random noise. Given only a single snapshot of the final network G, we study the problem of constructing confidence sets for the early history, in particular the root node, of the unobserved growth process; the root node can be patient zero in a disease infection network or the source of fake news in a social media network. We propose an inference algorithm based on Gibbs sampling that scales to networks with millions of nodes and provide theoretical analysis showing that the expected size of the confidence set is small so long as the noise level of the ER edges is not too large. We also propose variations of the model in which multiple growth processes occur simultaneously, reflecting the growth of multiple communities, and we use these models to provide a new approach to community detection.

stat.ME↗

arXiv2022

A generic framework for coded caching and distributed computation schemes

Several network communication problems are highly related such as coded caching and distributed computation. The centralized coded caching focuses on reducing the network burden in peak times in a wireless network system and the coded distributed computation studies the tradeoff between computation and communication in distributed system. In this paper, motivated by the study of the only rainbow $3$-term arithmetic progressions set, we propose a unified framework for constructing coded caching schemes. This framework builds bridges between coded caching schemes and lots of combinatorial objects due to the freedom of the choices of families and operations. We prove that any scheme based on a placement delivery array (PDA) can be represented by a rainbow scheme under this framework and lots of other known schemes can also be included in this framework. Moreover, we also present a new coded caching scheme with linear subpacketization and near constant rate using the only rainbow $3$-term arithmetic progressions set. Next, we modify the framework to be applicable to the distributed computing problem. We present a new transmission scheme in the shuffle phase and show that in certain cases it could have a lower communication load than the schemes based on PDAs or resolvable designs with the same number of files.

cs.IT↗

arXiv2022

BenchPress: A Deep Active Benchmark Generator

We develop BenchPress, the first ML benchmark generator for compilers that is steerable within feature space representations of source code. BenchPress synthesizes compiling functions by adding new code in any part of an empty or existing sequence by jointly observing its left and right context, achieving excellent compilation rate. BenchPress steers benchmark generation towards desired target features that has been impossible for state of the art synthesizers (or indeed humans) to reach. It performs better in targeting the features of Rodinia benchmarks in 3 different feature spaces compared with (a) CLgen - a state of the art ML synthesizer, (b) CLSmith fuzzer, (c) SRCIROR mutator or even (d) human-written code from GitHub. BenchPress is the first generator to search the feature space with active learning in order to generate benchmarks that will improve a downstream task. We show how using BenchPress, Grewe's et al. CPU vs GPU heuristic model can obtain a higher speedup when trained on BenchPress's benchmarks compared to other techniques. BenchPress is a powerful code generator: Its generated samples compile at a rate of 86%, compared to CLgen's 2.33%. Starting from an empty fixed input, BenchPress produces 10x more unique, compiling OpenCL benchmarks than CLgen, which are significantly larger and more feature diverse.

cs.AI↗

arXiv2022

Estimating Instance-dependent Bayes-label Transition Matrix using a Deep Neural Network

In label-noise learning, estimating the transition matrix is a hot topic as the matrix plays an important role in building statistically consistent classifiers. Traditionally, the transition from clean labels to noisy labels (i.e., clean-label transition matrix (CLTM)) has been widely exploited to learn a clean label classifier by employing the noisy data. Motivated by that classifiers mostly output Bayes optimal labels for prediction, in this paper, we study to directly model the transition from Bayes optimal labels to noisy labels (i.e., Bayes-label transition matrix (BLTM)) and learn a classifier to predict Bayes optimal labels. Note that given only noisy data, it is ill-posed to estimate either the CLTM or the BLTM. But favorably, Bayes optimal labels have less uncertainty compared with the clean labels, i.e., the class posteriors of Bayes optimal labels are one-hot vectors while those of clean labels are not. This enables two advantages to estimate the BLTM, i.e., (a) a set of examples with theoretically guaranteed Bayes optimal labels can be collected out of noisy data; (b) the feasible solution space is much smaller. By exploiting the advantages, we estimate the BLTM parametrically by employing a deep neural network, leading to better generalization and superior classification performance.

cs.LG↗

arXiv2022

Network Function Computation With Different Secure Conditions

In this paper, we investigate function computation problems under different secure conditions over a network with multiple source nodes and a single sink node which desires a function of all source messages without error. A wiretapper has access to some edges of the network. Based on different practical requirements, we consider two secure conditions named as secure and user secure respectively. The main parameter concerned here is the computing rate, which is the average times of the target function that can be computed securely or user securely without error for one use of the network. In the secure case, a new upper bound which is tighter than the previous one is provided for arithmetic sum functions and arbitrary networks. Moreover, we show that the improved upper bound is strictly tight for tree-like networks. In the user secure case, we give a sufficient and necessary condition for the existence of user secure network codes and obtain an upper bound for the computation capacity.

cs.IT↗

arXiv2022

An Efficient Multitask Neural Network for Face Alignment, Head Pose Estimation and Face Tracking

While Convolutional Neural Networks (CNNs) have significantly boosted the performance of face related algorithms, maintaining accuracy and efficiency simultaneously in practical use remains challenging. The state-of-the-art methods employ deeper networks for better performance, which makes it less practical for mobile applications because of more parameters and higher computational complexity. Therefore, we propose an efficient multitask neural network, Alignment & Tracking & Pose Network (ATPN) for face alignment, face tracking and head pose estimation. Specifically, to achieve better performance with fewer layers for face alignment, we introduce a shortcut connection between shallow-layer and deep-layer features. We find the shallow-layer features are highly correspond to facial boundaries that can provide the structural information of face and it is crucial for face alignment. Moreover, we generate a cheap heatmap based on the face alignment result and fuse it with features to improve the performance of the other two tasks. Based on the heatmap, the network can utilize both geometric information of landmarks and appearance information for head pose estimation. The heatmap also provides attention clues for face tracking. The face tracking task also saves us the face detection procedure for each frame, which also significantly boost the real-time capability for video-based tasks. We experimentally validate ATPN on four benchmark datasets, WFLW, 300VW, WIDER Face and 300W-LP. The experimental results demonstrate that it achieves better performance with much less parameters and lower computational complexity compared to other light models.

cs.CV↗

arXiv2022

Sparse Local Patch Transformer for Robust Face Alignment and Landmarks Inherent Relation Learning

Heatmap regression methods have dominated face alignment area in recent years while they ignore the inherent relation between different landmarks. In this paper, we propose a Sparse Local Patch Transformer (SLPT) for learning the inherent relation. The SLPT generates the representation of each single landmark from a local patch and aggregates them by an adaptive inherent relation based on the attention mechanism. The subpixel coordinate of each landmark is predicted independently based on the aggregated feature. Moreover, a coarse-to-fine framework is further introduced to incorporate with the SLPT, which enables the initial landmarks to gradually converge to the target facial landmarks using fine-grained features from dynamically resized local patches. Extensive experiments carried out on three popular benchmarks, including WFLW, 300W and COFW, demonstrate that the proposed method works at the state-of-the-art level with much less computational complexity by learning the inherent relation between facial landmarks. The code is available at the project website.

cs.CV↗

arXiv2022

Random Matrix Time Series

In this paper, a time series model with coefficients that take values from random matrix ensembles is proposed. Formal definitions, theoretical solutions, and statistical properties are derived. Estimation and forecast methodologies for random matrix time series are discussed with examples. Random matrix differential equations and potential applications of the time series model are suggested at the end.

stat.ME↗

arXiv2022

SHREC 2021: Classification in cryo-electron tomograms

Cryo-electron tomography (cryo-ET) is an imaging technique that allows three-dimensional visualization of macro-molecular assemblies under near-native conditions. Cryo-ET comes with a number of challenges, mainly low signal-to-noise and inability to obtain images from all angles. Computational methods are key to analyze cryo-electron tomograms. To promote innovation in computational methods, we generate a novel simulated dataset to benchmark different methods of localization and classification of biological macromolecules in tomograms. Our publicly available dataset contains ten tomographic reconstructions of simulated cell-like volumes. Each volume contains twelve different types of complexes, varying in size, function and structure. In this paper, we have evaluated seven different methods of finding and classifying proteins. Seven research groups present results obtained with learning-based methods and trained on the simulated dataset, as well as a baseline template matching (TM), a traditional method widely used in cryo-ET research. We show that learning-based approaches can achieve notably better localization and classification performance than TM. We also experimentally confirm that there is a negative relationship between particle size and performance for all methods.

eess.IV↗