Shuai Zheng

Jack FitzGerald, Shankar Ananthakrishnan, Konstantine Arkoudas et al. · amazon-science, gatech

We present results from a large-scale experiment on pretraining encoders with non-embedding parameter counts ranging from 700M to 9.3B, their subsequent distillation into smaller models ranging from 17M-170M parameters, and their application to the Natural Language Understanding (NLU) component of a virtual assistant system. Though we train using 70% spoken-form data, our teacher models perform comparably to XLM-R and mT5 when evaluated on the written-form Cross-lingual Natural Language Inference (XNLI) corpus. We perform a second stage of pretraining on our teacher models using in-domain data from our system, improving error rates by 3.86% relative for intent classification and 7.01% relative for slot filling. We find that even a 170M-parameter model distilled from our Stage 2 teacher model has 2.88% better intent classification and 7.69% better slot filling error rates when compared to the 2.3B-parameter teacher trained only on public data (Stage 1), emphasizing the importance of in-domain data for pretraining. When evaluated offline using labeled NLU data, our 17M-parameter Stage 2 distilled model outperforms both XLM-R Base (85M params) and DistillBERT (42M params) by 4.23% to 6.14%, respectively. Finally, we present results from a full virtual assistant experimentation platform, where we find that models trained using our pretraining and distillation pipeline outperform models distilled from 85M-parameter teachers by 3.74%-4.91% on an automatic measurement of full-system user dissatisfaction.

Muhao Xu, Zhenfeng Zhu, Youru Li et al.

Multimodal electronic health record (EHR) data can offer a holistic assessment of a patient's health status, supporting various predictive healthcare tasks. Recently, several studies have embraced the multitask learning approach in the healthcare domain, exploiting the inherent correlations among clinical tasks to predict multiple outcomes simultaneously. However, existing methods necessitate samples to possess complete labels for all tasks, which places heavy demands on the data and restricts the flexibility of the model. Meanwhile, within a multitask framework with multimodal inputs, how to comprehensively consider the information disparity among modalities and among tasks still remains a challenging problem. To tackle these issues, a unified healthcare prediction model, also named by \textbf{FlexCare}, is proposed to flexibly accommodate incomplete multimodal inputs, promoting the adaption to multiple healthcare tasks. The proposed model breaks the conventional paradigm of parallel multitask prediction by decomposing it into a series of asynchronous single-task prediction. Specifically, a task-agnostic multimodal information extraction module is presented to capture decorrelated representations of diverse intra- and inter-modality patterns. Taking full account of the information disparities between different modalities and different tasks, we present a task-guided hierarchical multimodal fusion module that integrates the refined modality-level representations into an individual patient-level representation. Experimental results on multiple tasks from MIMIC-IV/MIMIC-CXR/MIMIC-NOTE datasets demonstrate the effectiveness of the proposed method. Additionally, further analysis underscores the feasibility and potential of employing such a multitask strategy in the healthcare domain. The source code is available at https://github.com/mhxu1998/FlexCare.

Shuai Zheng, Zhenfeng Zhu, Zhizhe Liu et al.

A real-world graph has a complex topological structure, which is often formed by the interaction of different latent factors. However, most existing methods lack consideration of the intrinsic differences in relations between nodes caused by factor entanglement. In this paper, we propose an \underline{\textbf{A}}dversarial \underline{\textbf{D}}isentangled \underline{\textbf{G}}raph \underline{\textbf{C}}onvolutional \underline{\textbf{N}}etwork (ADGCN) for disentangled graph representation learning. To begin with, we point out two aspects of graph disentanglement that need to be considered, i.e., micro-disentanglement and macro-disentanglement. For them, a component-specific aggregation approach is proposed to achieve micro-disentanglement by inferring latent components that cause the links between nodes. On the basis of micro-disentanglement, we further propose a macro-disentanglement adversarial regularizer to improve the separability among component distributions, thus restricting the interdependence among components. Additionally, to reveal the topological graph structure, a diversity-preserving node sampling approach is proposed, by which the graph structure can be progressively refined in a way of local structure awareness. The experimental results on various real-world graph data verify that our ADGCN obtains more favorable performance over currently available alternatives. The source codes of ADGCN are available at \textit{\url{https://github.com/SsGood/ADGCN}}.

Shuai Zheng, Fan Yang, M. Hadi Kiapour et al.

Understanding clothes from a single image has strong commercial and cultural impacts on modern societies. However, this task remains a challenging computer vision problem due to wide variations in the appearance, style, brand and layering of clothing items. We present a new database called ModaNet, a large-scale collection of images based on Paperdoll dataset. Our dataset provides 55,176 street images, fully annotated with polygons on top of the 1 million weakly annotated street images in Paperdoll. ModaNet aims to provide a technical benchmark to fairly evaluate the progress of applying the latest computer vision techniques that rely on large data for fashion understanding. The rich annotation of the dataset allows to measure the performance of state-of-the-art algorithms for object detection, semantic segmentation and polygon prediction on street fashion images in detail. The polygon-based annotation dataset has been released https://github.com/eBay/modanet, we also host the leaderboard at EvalAI: https://evalai.cloudcv.org/featured-challenges/136/overview.

Nianyi Wang, Yu Chen, Shuai Zheng

Learning-based fluid simulation networks have been proven as viable alternatives to traditional numerical solvers for the Navier-Stokes equations. Existing neural methods follow Smoothed Particle Hydrodynamics (SPH) frameworks, which inherently rely only on local inter-particle interactions. However, we emphasize that global context integration is also essential for learning-based methods to stabilize complex fluid simulations. We propose the first Fluid Attention Block (FAB) with a local-global hierarchy, where continuous convolutions extract local features while self-attention captures global dependencies. This fusion suppresses the error accumulation and models long-range physical phenomena. Furthermore, we pioneer the first Transformer architecture specifically designed for continuous fluid simulation, seamlessly integrated within a dual-pipeline architecture. Our method establishes a new paradigm for neural fluid simulation by unifying convolution-based local features with attention-based global context modeling. FluidFormer demonstrates state-of-the-art performance, with stronger stability in complex fluid scenarios.

Xilong Xie, Liang Wang, Limin Xiao et al.

Large language models (LLMs) have significantly advanced the natural language processing paradigm but impose substantial demands on memory and computational resources. Quantization is one of the most effective ways to reduce memory consumption of LLMs. However, advanced single-precision quantization methods experience significant accuracy degradation when quantizing to ultra-low bits. Existing mixed-precision quantization methods are quantized by groups with coarse granularity. Employing high precision for group data leads to substantial memory overhead, whereas low precision severely impacts model accuracy. To address this issue, we propose FineQ, software-hardware co-design for low-bit fine-grained mixed-precision quantization of LLMs. First, FineQ partitions the weights into finer-grained clusters and considers the distribution of outliers within these clusters, thus achieving a balance between model accuracy and memory overhead. Then, we propose an outlier protection mechanism within clusters that uses 3 bits to represent outliers and introduce an encoding scheme for index and data concatenation to enable aligned memory access. Finally, we introduce an accelerator utilizing temporal coding that effectively supports the quantization algorithm while simplifying the multipliers in the systolic array. FineQ achieves higher model accuracy compared to the SOTA mixed-precision quantization algorithm at a close average bit-width. Meanwhile, the accelerator achieves up to 1.79x energy efficiency and reduces the area of the systolic array by 61.2%.

Zhenyu Guo, Shuai Zheng, Zhizhe Liu et al.

Treatment effect estimation, which refers to the estimation of causal effects and aims to measure the strength of the causal relationship, is of great importance in many fields but is a challenging problem in practice. As present, data-driven causal effect estimation faces two main challenges, i.e., selection bias and the missing of counterfactual. To address these two issues, most of the existing approaches tend to reduce the selection bias by learning a balanced representation, and then to estimate the counterfactual through the representation. However, they heavily rely on the finely hand-crafted metric functions when learning balanced representations, which generally doesn't work well for the situations where the original distribution is complicated. In this paper, we propose a CETransformer model for casual effect estimation via transformer based representation learning. To learn the representation of covariates(features) robustly, a self-supervised transformer is proposed, by which the correlation between covariates can be well exploited through self-attention mechanism. In addition, an adversarial network is adopted to balance the distribution of the treated and control groups in the representation space. Experimental results on three real-world datasets demonstrate the advantages of the proposed CETransformer, compared with the state-of-the-art treatment effect estimation methods.

Shuai Zheng, Zhenfeng Zhu, Zhizhe Liu et al.

Benefiting from the powerful expressive capability of graphs, graph-based approaches have achieved impressive performance in various biomedical applications. Most existing methods tend to define the adjacency matrix among samples manually based on meta-features, and then obtain the node embeddings for downstream tasks by Graph Representation Learning (GRL). However, it is not easy for these approaches to generalize to unseen samples. Meanwhile, the complex correlation between modalities is also ignored. As a result, these factors inevitably yield the inadequacy of providing valid information about the patient's condition for a reliable diagnosis. In this paper, we propose an end-to-end Multimodal Graph Learning framework (MMGL) for disease prediction. To effectively exploit the rich information across multi-modality associated with diseases, amodal-attentional multi-modal fusion is proposed to integrate the features of each modality by leveraging the correlation and complementarity between the modalities. Furthermore, instead of defining the adjacency matrix manually as existing methods, the latent graph structure can be captured through a novel way of adaptive graph learning. It could be jointly optimized with the prediction model, thus revealing the intrinsic connections among samples. Unlike the previous transductive methods, our model is also applicable to the scenario of inductive learning for those unseen data. An extensive group of experiments on two disease prediction problems is then carefully designed and presented, demonstrating that MMGL obtains more favorable performances. In addition, we also visualize and analyze the learned graph structure to provide more reliable decision support for doctors in real medical applications and inspiration for disease research.

Yuchen Zhong, Cong Xie, Shuai Zheng et al.

Communication overhead severely hinders the scalability of distributed machine learning systems. Recently, there has been a growing interest in using gradient compression to reduce the communication overhead of the distributed training. However, there is little understanding of applying gradient compression to adaptive gradient methods. Moreover, its performance benefits are often limited by the non-negligible compression overhead. In this paper, we first introduce a novel adaptive gradient method with gradient compression. We show that the proposed method has a convergence rate of $\mathcal{O}(1/\sqrt{T})$ for non-convex problems. In addition, we develop a scalable system called BytePS-Compress for two-way compression, where the gradients are compressed in both directions between workers and parameter servers. BytePS-Compress pipelines the compression and decompression on CPUs and achieves a high degree of parallelism. Empirical evaluations show that we improve the training time of ResNet50, VGG16, and BERT-base by 5.0%, 58.1%, 23.3%, respectively, without any accuracy loss with 25 Gb/s networking. Furthermore, for training the BERT models, we achieve a compression rate of 333x compared to the mixed-precision training.

Zhizhe Liu, Zhenfeng Zhu, Shuai Zheng et al.

To bridge the gap between the source and target domains in unsupervised domain adaptation (UDA), the most common strategy puts focus on matching the marginal distributions in the feature space through adversarial learning. However, such category-agnostic global alignment lacks of exploiting the class-level joint distributions, causing the aligned distribution less discriminative. To address this issue, we propose in this paper a novel margin preserving self-paced contrastive Learning (MPSCL) model for cross-modal medical image segmentation. Unlike the conventional construction of contrastive pairs in contrastive learning, the domain-adaptive category prototypes are utilized to constitute the positive and negative sample pairs. With the guidance of progressively refined semantic prototypes, a novel margin preserving contrastive loss is proposed to boost the discriminability of embedded representation space. To enhance the supervision for contrastive learning, more informative pseudo-labels are generated in target domain in a self-paced way, thus benefiting the category-aware distribution alignment for UDA. Furthermore, the domain-invariant representations are learned through joint contrastive learning between the two domains. Extensive experiments on cross-modal cardiac segmentation tasks demonstrate that MPSCL significantly improves semantic segmentation performance, and outperforms a wide variety of state-of-the-art methods by a large margin.

Zhizhe Liu, Xingxing Zhang, Zhenfeng Zhu et al.

Human identification is an important topic in event detection, person tracking, and public security. There have been numerous methods proposed for human identification, such as face identification, person re-identification, and gait identification. Typically, existing methods predominantly classify a queried image to a specific identity in an image gallery set (I2I). This is seriously limited for the scenario where only a textual description of the query or an attribute gallery set is available in a wide range of video surveillance applications (A2I or I2A). However, very few efforts have been devoted towards modality-free identification, i.e., identifying a query in a gallery set in a scalable way. In this work, we take an initial attempt, and formulate such a novel Modality-Free Human Identification (named MFHI) task as a generic zero-shot learning model in a scalable way. Meanwhile, it is capable of bridging the visual and semantic modalities by learning a discriminative prototype of each identity. In addition, the semantics-guided spatial attention is enforced on visual modality to obtain representations with both high global category-level and local attribute-level discrimination. Finally, we design and conduct an extensive group of experiments on two common challenging identification tasks, including face identification and person re-identification, demonstrating that our method outperforms a wide variety of state-of-the-art methods on modality-free human identification.

Cong Xie, Shuai Zheng, Oluwasanmi Koyejo et al.

The scalability of Distributed Stochastic Gradient Descent (SGD) is today limited by communication bottlenecks. We propose a novel SGD variant: Communication-efficient SGD with Error Reset, or CSER. The key idea in CSER is first a new technique called "error reset" that adapts arbitrary compressors for SGD, producing bifurcated local models with periodic reset of resulting local residual errors. Second we introduce partial synchronization for both the gradients and the models, leveraging advantages from them. We prove the convergence of CSER for smooth non-convex problems. Empirical results show that when combined with highly aggressive compressors, the CSER algorithms accelerate the distributed training by nearly 10x for CIFAR-100, and by 4.5x for ImageNet.

Shuai Zheng, Haibin Lin, Sheng Zha et al.

BERT has recently attracted a lot of attention in natural language understanding (NLU) and achieved state-of-the-art results in various NLU tasks. However, its success requires large deep neural networks and huge amount of data, which result in long training time and impede development progress. Using stochastic gradient methods with large mini-batch has been advocated as an efficient tool to reduce the training time. Along this line of research, LAMB is a prominent example that reduces the training time of BERT from 3 days to 76 minutes on a TPUv3 Pod. In this paper, we propose an accelerated gradient method called LANS to improve the efficiency of using large mini-batches for training. As the learning rate is theoretically upper bounded by the inverse of the Lipschitz constant of the function, one cannot always reduce the number of optimization iterations by selecting a larger learning rate. In order to use larger mini-batch size without accuracy loss, we develop a new learning rate scheduler that overcomes the difficulty of using large learning rate. Using the proposed LANS method and the learning rate scheme, we scaled up the mini-batch sizes to 96K and 33K in phases 1 and 2 of BERT pretraining, respectively. It takes 54 minutes on 192 AWS EC2 P3dn.24xlarge instances to achieve a target F1 score of 90.5 or higher on SQuAD v1.1, achieving the fastest BERT training time in the cloud.

Shuai Zheng, Zhenfeng Zhu, Xingxing Zhang et al.

Graph representation learning aims to encode all nodes of a graph into low-dimensional vectors that will serve as input of many compute vision tasks. However, most existing algorithms ignore the existence of inherent data distribution and even noises. This may significantly increase the phenomenon of over-fitting and deteriorate the testing accuracy. In this paper, we propose a Distribution-induced Bidirectional Generative Adversarial Network (named DBGAN) for graph representation learning. Instead of the widely used normal distribution assumption, the prior distribution of latent representation in our DBGAN is estimated in a structure-aware way, which implicitly bridges the graph and feature spaces by prototype learning. Thus discriminative and robust representations are generated for all nodes. Furthermore, to improve their generalization ability while preserving representation ability, the sample-level and distribution-level consistency is well balanced via a bidirectional adversarial learning framework. An extensive group of experiments are then carefully designed and presented, demonstrating that our DBGAN obtains remarkably more favorable trade-off between representation and robustness, and meanwhile is dimension-efficient, over currently available alternatives in various tasks.

Zhizhe Liu, Xingxing Zhang, Zhenfeng Zhu et al.

Zero-shot learning (ZSL) has received increasing attention in recent years especially in areas of fine-grained object recognition, retrieval, and image captioning. The key to ZSL is to transfer knowledge from the seen to the unseen classes via auxiliary class attribute vectors. However, the popularly learned projection functions in previous works cannot generalize well since they assume the distribution consistency between seen and unseen domains at sample-level.Besides, the provided non-visual and unique class attributes can significantly degrade the recognition performance in semantic space. In this paper, we propose a simple yet effective convolutional prototype learning (CPL) framework for zero-shot recognition. By assuming distribution consistency at task-level, our CPL is capable of transferring knowledge smoothly to recognize unseen samples.Furthermore, inside each task, discriminative visual prototypes are learned via a distance based training mechanism. Consequently, we can perform recognition in visual space, instead of semantic space. An extensive group of experiments are then carefully designed and presented, demonstrating that CPL obtains more favorable effectiveness, over currently available alternatives under various settings.

Jian Guo, He He, Tong He et al.

We present GluonCV and GluonNLP, the deep learning toolkits for computer vision and natural language processing based on Apache MXNet (incubating). These toolkits provide state-of-the-art pre-trained models, training scripts, and training logs, to facilitate rapid prototyping and promote reproducible research. We also provide modular APIs with flexible building blocks to enable efficient customization. Leveraging the MXNet ecosystem, the deep learning models in GluonCV and GluonNLP can be deployed onto a variety of platforms with different programming languages. The Apache 2.0 license has been adopted by GluonCV and GluonNLP to allow for software distribution, modification, and usage.

Shuai Zheng, Ziyue Huang, James T. Kwok

Communication overhead is a major bottleneck hampering the scalability of distributed machine learning systems. Recently, there has been a surge of interest in using gradient compression to improve the communication efficiency of distributed neural network training. Using 1-bit quantization, signSGD with majority vote achieves a 32x reduction on communication cost. However, its convergence is based on unrealistic assumptions and can diverge in practice. In this paper, we propose a general distributed compressed SGD with Nesterov's momentum. We consider two-way compression, which compresses the gradients both to and from workers. Convergence analysis on nonconvex problems for general gradient compressors is provided. By partitioning the gradient into blocks, a blockwise compressor is introduced such that each gradient block is compressed and transmitted in 1-bit format with a scaling factor, leading to a nearly 32x reduction on communication. Experimental results show that the proposed method converges as fast as full-precision distributed momentum SGD and achieves the same testing accuracy. In particular, on distributed ResNet training with 7 workers on the ImageNet, the proposed algorithm achieves the same testing accuracy as momentum SGD using full-precision gradients, but with $46\%$ less wall clock time.

Bryan A. Plummer, M. Hadi Kiapour, Shuai Zheng et al.

In this paper, we introduce an attribute-based interactive image search which can leverage human-in-the-loop feedback to iteratively refine image search results. We study active image search where human feedback is solicited exclusively in visual form, without using relative attribute annotations used by prior work which are not typically found in many datasets. In order to optimize the image selection strategy, a deep reinforcement model is trained to learn what images are informative rather than rely on hand-crafted measures typically leveraged in prior work. Additionally, we extend the recently introduced Conditional Similarity Network to incorporate global similarity in training visual embeddings, which results in more natural transitions as the user explores the learned similarity embeddings. Our experiments demonstrate the effectiveness of our approach, producing compelling results on both active image search and image attribute representation tasks.

Shuai Zheng, James T. Kwok

Variance reduction has been commonly used in stochastic optimization. It relies crucially on the assumption that the data set is finite. However, when the data are imputed with random noise as in data augmentation, the perturbed data set be- comes essentially infinite. Recently, the stochastic MISO (S-MISO) algorithm is introduced to address this expected risk minimization problem. Though it converges faster than SGD, a significant amount of memory is required. In this pa- per, we propose two SGD-like algorithms for expected risk minimization with random perturbation, namely, stochastic sample average gradient (SSAG) and stochastic SAGA (S-SAGA). The memory cost of SSAG does not depend on the sample size, while that of S-SAGA is the same as those of variance reduction methods on un- perturbed data. Theoretical analysis and experimental results on logistic regression and AUC maximization show that SSAG has faster convergence rate than SGD with comparable space requirement, while S-SAGA outperforms S-MISO in terms of both iteration complexity and storage.

Bryan A. Plummer, Paige Kordas, M. Hadi Kiapour et al.

This paper presents an approach for grounding phrases in images which jointly learns multiple text-conditioned embeddings in a single end-to-end model. In order to differentiate text phrases into semantically distinct subspaces, we propose a concept weight branch that automatically assigns phrases to embeddings, whereas prior works predefine such assignments. Our proposed solution simplifies the representation requirements for individual embeddings and allows the underrepresented concepts to take advantage of the shared representations before feeding them into concept-specific layers. Comprehensive experiments verify the effectiveness of our approach across three phrase grounding datasets, Flickr30K Entities, ReferIt Game, and Visual Genome, where we obtain a (resp.) 4%, 3%, and 4% improvement in grounding performance over a strong region-phrase embedding baseline.

Shuai Zheng, James T. Kwok

The alternating direction method of multipliers (ADMM) is a powerful optimization solver in machine learning. Recently, stochastic ADMM has been integrated with variance reduction methods for stochastic gradient, leading to SAG-ADMM and SDCA-ADMM that have fast convergence rates and low iteration complexities. However, their space requirements can still be high. In this paper, we propose an integration of ADMM with the method of stochastic variance reduced gradient (SVRG). Unlike another recent integration attempt called SCAS-ADMM, the proposed algorithm retains the fast convergence benefits of SAG-ADMM and SDCA-ADMM, but is more advantageous in that its storage requirement is very low, even independent of the sample size $n$. We also extend the proposed method for nonconvex problems, and obtain a convergence rate of $O(1/T)$. Experimental results demonstrate that it is as fast as SAG-ADMM and SDCA-ADMM, much faster than SCAS-ADMM, and can be used on much bigger data sets.