arXiv subjects

Min Wu

Publications and source records attributed to Min Wu.

At least 163 records · Page 9Linked to original sources

arXiv2023

Convex Bounds on the Softmax Function with Applications to Robustness Verification

The softmax function is a ubiquitous component at the output of neural networks and increasingly in intermediate layers as well. This paper provides convex lower bounds and concave upper bounds on the softmax function, which are compatible with convex optimization formulations for characterizing neural networks and other ML models. We derive bounds using both a natural exponential-reciprocal decomposition of the softmax as well as an alternative decomposition in terms of the log-sum-exp function. The new bounds are provably and/or numerically tighter than linear bounds obtained in previous work on robustness verification of transformers. As illustrations of the utility of the bounds, we apply them to verification of transformers as well as of the robustness of predictive uncertainty estimates of deep ensembles.

cs.LG↗

arXiv2023

Real-time Speech Emotion Recognition Based on Syllable-Level Feature Extraction

Speech emotion recognition systems have high prediction latency because of the high computational requirements for deep learning models and low generalizability mainly because of the poor reliability of emotional measurements across multiple corpora. To solve these problems, we present a speech emotion recognition system based on a reductionist approach of decomposing and analyzing syllable-level features. Mel-spectrogram of an audio stream is decomposed into syllable-level components, which are then analyzed to extract statistical features. The proposed method uses formant attention, noise-gate filtering, and rolling normalization contexts to increase feature processing speed and tolerance to adversity. A set of syllable-level formant features is extracted and fed into a single hidden layer neural network that makes predictions for each syllable as opposed to the conventional approach of using a sophisticated deep learner to make sentence-wide predictions. The syllable level predictions help to achieve the real-time latency and lower the aggregated error in utterance level cross-corpus predictions. The experiments on IEMOCAP (IE), MSP-Improv (MI), and RAVDESS (RA) databases show that the method archives real-time latency while predicting with state-of-the-art cross-corpus unweighted accuracy of 47.6% for IE to MI and 56.2% for MI to IE.

cs.SD↗

arXiv2022

Traceable Automatic Feature Transformation via Cascading Actor-Critic Agents

Feature transformation for AI is an essential task to boost the effectiveness and interpretability of machine learning (ML). Feature transformation aims to transform original data to identify an optimal feature space that enhances the performances of a downstream ML model. Existing studies either combines preprocessing, feature selection, and generation skills to empirically transform data, or automate feature transformation by machine intelligence, such as reinforcement learning. However, existing studies suffer from: 1) high-dimensional non-discriminative feature space; 2) inability to represent complex situational states; 3) inefficiency in integrating local and global feature information. To fill the research gap, we formulate the feature transformation task as an iterative, nested process of feature generation and selection, where feature generation is to generate and add new features based on original features, and feature selection is to remove redundant features to control the size of feature space. Finally, we present extensive experiments and case studies to illustrate 24.7\% improvements in F1 scores compared with SOTAs and robustness in high-dimensional data.

cs.LG↗

arXiv2022

Self-Optimizing Feature Transformation

Feature transformation aims to extract a good representation (feature) space by mathematically transforming existing features. It is crucial to address the curse of dimensionality, enhance model generalization, overcome data sparsity, and expand the availability of classic models. Current research focuses on domain knowledge-based feature engineering or learning latent representations; nevertheless, these methods are not entirely automated and cannot produce a traceable and optimal representation space. When rebuilding a feature space for a machine learning task, can these limitations be addressed concurrently? In this extension study, we present a self-optimizing framework for feature transformation. To achieve a better performance, we improved the preliminary work by (1) obtaining an advanced state representation for enabling reinforced agents to comprehend the current feature set better; and (2) resolving Q-value overestimation in reinforced agents for learning unbiased and effective policies. Finally, to make experiments more convincing than the preliminary work, we conclude by adding the outlier detection task with five datasets, evaluating various state representation approaches, and comparing different training strategies. Extensive experiments and case studies show that our work is more effective and superior.

cs.LG↗

arXiv2022

Leveraging Endo- and Exo-Temporal Regularization for Black-box Video Domain Adaptation

To enable video models to be applied seamlessly across video tasks in different environments, various Video Unsupervised Domain Adaptation (VUDA) methods have been proposed to improve the robustness and transferability of video models. Despite improvements made in model robustness, these VUDA methods require access to both source data and source model parameters for adaptation, raising serious data privacy and model portability issues. To cope with the above concerns, this paper firstly formulates Black-box Video Domain Adaptation (BVDA) as a more realistic yet challenging scenario where the source video model is provided only as a black-box predictor. While a few methods for Black-box Domain Adaptation (BDA) are proposed in image domain, these methods cannot apply to video domain since video modality has more complicated temporal features that are harder to align. To address BVDA, we propose a novel Endo and eXo-TEmporal Regularized Network (EXTERN) by applying mask-to-mix strategies and video-tailored regularizations: endo-temporal regularization and exo-temporal regularization, performed across both clip and temporal features, while distilling knowledge from the predictions obtained from the black-box predictor. Empirical results demonstrate the state-of-the-art performance of EXTERN across various cross-domain closed-set and partial-set action recognition benchmarks, which even surpassed most existing video domain adaptation methods with source data accessibility.

cs.CV↗

arXiv2022

Self-supervised Learning for Label-Efficient Sleep Stage Classification: A Comprehensive Evaluation

The past few years have witnessed a remarkable advance in deep learning for EEG-based sleep stage classification (SSC). However, the success of these models is attributed to possessing a massive amount of labeled data for training, limiting their applicability in real-world scenarios. In such scenarios, sleep labs can generate a massive amount of data, but labeling these data can be expensive and time-consuming. Recently, the self-supervised learning (SSL) paradigm has shined as one of the most successful techniques to overcome the scarcity of labeled data. In this paper, we evaluate the efficacy of SSL to boost the performance of existing SSC models in the few-labels regime. We conduct a thorough study on three SSC datasets, and we find that fine-tuning the pretrained SSC models with only 5% of labeled data can achieve competitive performance to the supervised training with full labels. Moreover, self-supervised pretraining helps SSC models to be more robust to data imbalance and domain shift problems. The code is publicly available at https://github.com/emadeldeen24/eval_ssl_ssc.

eess.SP↗

arXiv2022

High-performance chiral all-optical logic gate based on topological edge states of valley photonic crystal

For all-optical communication and information processing, it is necessary to develop all-optical logic gates based on photonic structures that can directly perform logic operations. All-optical logic gates have been demonstrated based on conventional waveguides and interferometry, as well as photonic crystal structures. Nonetheless, any defects in those structures will introduce high scattering loss, which compromises the fidelity and contrast ratio of the information process. Based on the spin-valley locking effect that can achieve defect-immune unidirectional transmission of topological edge states in valley photonic crystals (VPCs), we propose a high-performance all-optical logic OR gate based on a VPC structure. By tuning the working bandwidth of the two input channels, we prevent interference between the two channels to achieve a stable and high-fidelity output. The transmittance of both channels is higher than 0.8, and a high contrast ratio of 28.8 dB is achieved. Moreover, the chirality of the logic gate originated from the spin-valley locking effect allows using different circularly polarized light as inputs, representing "1" or "0", which is highly desired in quantum computing. The device's footprint is small, allowing high-density on-chip integration. In addition, this design can be experimentally fabricated using current nanofabrication techniques and will have potential applications in optical communication, information processing, and quantum computing.

physics.optics↗

arXiv2022

Full Poincaré polarimetry enabled through physical inference

While polarisation sensing is vital in many areas of research, with applications spanning from microscopy to aerospace, traditional approaches are limited by method-related error amplification or accumulation, placing fundamental limitations on precision and accuracy in single-shot polarimetry. Here, we put forward a new measurement paradigm to circumvent this, introducing the notion of a universal full Poincaré generator to map all polarisation analyser states into a single vectorially structured light field, allowing all vector components to be analysed in a single-shot with theoretically user-defined precision. To demonstrate the advantage of our approach, we use a common GRIN optic as our mapping device and show mean errors of <1% for each vector component, enhancing the sensitivity by around three times, allowing us to sense weak polarisation aberrations not measurable by traditional single-shot techniques. Our work paves the way for next-generation polarimetry, impacting a wide variety of applications relying on weak vector measurement.

physics.optics↗

arXiv2022

SwiftPruner: Reinforced Evolutionary Pruning for Efficient Ad Relevance

Ad relevance modeling plays a critical role in online advertising systems including Microsoft Bing. To leverage powerful transformers like BERT in this low-latency setting, many existing approaches perform ad-side computations offline. While efficient, these approaches are unable to serve cold start ads, resulting in poor relevance predictions for such ads. This work aims to design a new, low-latency BERT via structured pruning to empower real-time online inference for cold start ads relevance on a CPU platform. Our challenge is that previous methods typically prune all layers of the transformer to a high, uniform sparsity, thereby producing models which cannot achieve satisfactory inference speed with an acceptable accuracy. In this paper, we propose SwiftPruner - an efficient framework that leverages evolution-based search to automatically find the best-performing layer-wise sparse BERT model under the desired latency constraint. Different from existing evolution algorithms that conduct random mutations, we propose a reinforced mutator with a latency-aware multi-objective reward to conduct better mutations for efficiently searching the large space of layer-wise sparse models. Extensive experiments demonstrate that our method consistently achieves higher ROC AUC and lower latency than the uniform sparse baseline and state-of-the-art search methods. Remarkably, under our latency requirement of 1900us on CPU, SwiftPruner achieves a 0.86% higher AUC than the state-of-the-art uniform sparse baseline for BERT-Mini on a large scale real-world dataset. Online A/B testing shows that our model also achieves a significant 11.7% cut in the ratio of defective cold start ads with satisfactory real-time serving latency.

cs.IR↗

arXiv2022

Root-aligned SMILES: A Tight Representation for Chemical Reaction Prediction

Chemical reaction prediction, involving forward synthesis and retrosynthesis prediction, is a fundamental problem in organic synthesis. A popular computational paradigm formulates synthesis prediction as a sequence-to-sequence translation problem, where the typical SMILES is adopted for molecule representations. However, the general-purpose SMILES neglects the characteristics of chemical reactions, where the molecular graph topology is largely unaltered from reactants to products, resulting in the suboptimal performance of SMILES if straightforwardly applied. In this article, we propose the root-aligned SMILES (R-SMILES), which specifies a tightly aligned one-to-one mapping between the product and the reactant SMILES for more efficient synthesis prediction. Due to the strict one-to-one mapping and reduced edit distance, the computational model is largely relieved from learning the complex syntax and dedicated to learning the chemical knowledge for reactions. We compare the proposed R-SMILES with various state-of-the-art baselines and show that it significantly outperforms them all, demonstrating the superiority of the proposed method.

cs.LG↗

arXiv2022

Self-supervised Autoregressive Domain Adaptation for Time Series Data

Unsupervised domain adaptation (UDA) has successfully addressed the domain shift problem for visual applications. Yet, these approaches may have limited performance for time series data due to the following reasons. First, they mainly rely on large-scale dataset (i.e., ImageNet) for the source pretraining, which is not applicable for time-series data. Second, they ignore the temporal dimension on the feature space of the source and target domains during the domain alignment step. Last, most of prior UDA methods can only align the global features without considering the fine-grained class distribution of the target domain. To address these limitations, we propose a Self-supervised Autoregressive Domain Adaptation (SLARDA) framework. In particular, we first design a self-supervised learning module that utilizes forecasting as an auxiliary task to improve the transferability of the source features. Second, we propose a novel autoregressive domain adaptation technique that incorporates temporal dependency of both source and target features during domain alignment. Finally, we develop an ensemble teacher model to align the class-wise distribution in the target domain via a confident pseudo labeling approach. Extensive experiments have been conducted on three real-world time series applications with 30 cross-domain scenarios. Results demonstrate that our proposed SLARDA method significantly outperforms the state-of-the-art approaches for time series domain adaptation.

cs.LG↗

arXiv2022

ADAST: Attentive Cross-domain EEG-based Sleep Staging Framework with Iterative Self-Training

Sleep staging is of great importance in the diagnosis and treatment of sleep disorders. Recently, numerous data-driven deep learning models have been proposed for automatic sleep staging. They mainly train the model on a large public labeled sleep dataset and test it on a smaller one with subjects of interest. However, they usually assume that the train and test data are drawn from the same distribution, which may not hold in real-world scenarios. Unsupervised domain adaption (UDA) has been recently developed to handle this domain shift problem. However, previous UDA methods applied for sleep staging have two main limitations. First, they rely on a totally shared model for the domain alignment, which may lose the domain-specific information during feature extraction. Second, they only align the source and target distributions globally without considering the class information in the target domain, which hinders the classification performance of the model while testing. In this work, we propose a novel adversarial learning framework called ADAST to tackle the domain shift problem in the unlabeled target domain. First, we develop an unshared attention mechanism to preserve the domain-specific features in both domains. Second, we design an iterative self-training strategy to improve the classification performance on the target domain via target domain pseudo labels. We also propose dual distinct classifiers to increase the robustness and quality of the pseudo labels. The experimental results on six cross-domain scenarios validate the efficacy of our proposed framework and its advantage over state-of-the-art UDA methods. The source code is available at https://github.com/emadeldeen24/ADAST.

cs.LG↗

arXiv2022

Multi-Omic Data Integration and Feature Selection for Survival-based Patient Stratification via Supervised Concrete Autoencoders

Cancer is a complex disease with significant social and economic impact. Advancements in high-throughput molecular assays and the reduced cost for performing high-quality multi-omics measurements have fuelled insights through machine learning . Previous studies have shown promise on using multiple omic layers to predict survival and stratify cancer patients. In this paper, we developed a Supervised Autoencoder (SAE) model for survival-based multi-omic integration which improves upon previous work, and report a Concrete Supervised Autoencoder model (CSAE), which uses feature selection to jointly reconstruct the input features as well as predict survival. Our experiments show that our models outperform or are on par with some of the most commonly used baselines, while either providing a better survival separation (SAE) or being more interpretable (CSAE). We also perform a feature selection stability analysis on our models and notice that there is a power-law relationship with features which are commonly associated with survival. The code for this project is available at: https://github.com/phcavelar/coxae

cs.LG↗

arXiv2022

A Generalized Probabilistic Monitoring Model with Both Random and Sequential Data

Many multivariate statistical analysis methods and their corresponding probabilistic counterparts have been adopted to develop process monitoring models in recent decades. However, the insightful connections between them have rarely been studied. In this study, a generalized probabilistic monitoring model (GPMM) is developed with both random and sequential data. Since GPMM can be reduced to various probabilistic linear models under specific restrictions, it is adopted to analyze the connections between different monitoring methods. Using expectation maximization (EM) algorithm, the parameters of GPMM are estimated for both random and sequential cases. Based on the obtained model parameters, statistics are designed for monitoring different aspects of the process system. Besides, the distributions of these statistics are rigorously derived and proved, so that the control limits can be calculated accordingly. After that, contribution analysis methods are presented for identifying faulty variables once the process anomalies are detected. Finally, the equivalence between monitoring models based on classical multivariate methods and their corresponding probabilistic graphic models is further investigated. The conclusions of this study are verified using a numerical example and the Tennessee Eastman (TE) process. Experimental results illustrate that the proposed monitoring statistics are subject to their corresponding distributions, and they are equivalent to statistics in classical deterministic models under specific restrictions.

eess.SY↗

arXiv2022

FFConv: Fast Factorized Convolutional Neural Network Inference on Encrypted Data

Homomorphic Encryption (HE), allowing computations on encrypted data (ciphertext) without decrypting it first, enables secure but prohibitively slow Convolutional Neural Network (CNN) inference for privacy-preserving applications in clouds. To reduce the inference latency, one approach is to pack multiple messages into a single ciphertext in order to reduce the number of ciphertexts and support massive parallelism of Homomorphic Multiply-Accumulate (HMA) operations between ciphertexts. Despite the faster HECNN inference, the mainstream packing schemes Dense Packing (DensePack) and Convolution Packing (ConvPack) introduce expensive rotation overhead, which prolongs the inference latency of HECNN for deeper and wider CNN architectures. In this paper, we propose a low-rank factorization method named FFConv dedicated to efficient ciphertext packing for reducing both the rotation overhead and HMA operations. FFConv approximates a d x d convolution layer with low-rank factorized convolutions, in which a d x d low-rank convolution with fewer channels is followed by a 1 x 1 convolution to restore the channels. The d x d low-rank convolution with DensePack leads to significantly reduced rotation operations, while the rotation overhead of 1 x 1 convolution with ConvPack is close to zero. To our knowledge, FFConv is the first work that is capable of reducing the rotation overhead incurred by DensePack and ConvPack simultaneously, without introducing additional special blocks into the HECNN inference pipeline. Compared to prior art LoLa and Falcon, our method reduces the inference latency by up to 88% and 21%, respectively, with comparable accuracy on MNIST and CIFAR-10.

cs.CR↗

arXiv2022

Nebula Graph: An open source distributed graph database

This paper introduces the recent work of Nebula Graph, an open-source, distributed, scalable, and native graph database. We present a system design trade-off and a comprehensive overview of Nebula Graph internals, including graph data models, partitioning strategies, secondary indexes, optimizer rules, storage-side transactions, graph query languages, observability, graph processing frameworks, and visualization tool-kits. In addition, three sets of large-scale graph b

cs.DB↗

arXiv2022

Bispectrum-based Cross-frequency Functional Connectivity: Classification of Alzheimer's disease

Alzheimer's disease (AD) is a neurodegenerative disease known to affect brain functional connectivity (FC). Linear FC measures have been applied to study the differences in AD by splitting neurophysiological signals such as electroencephalography (EEG) recordings into discrete frequency bands and analysing them in isolation. We address this limitation by quantifying cross-frequency FC in addition to the traditional within-band approach. Cross-bispectrum, a higher-order spectral analysis, is used to measure the nonlinear FC and is compared with the cross-spectrum, which only measures the linear FC within bands. Each frequency coupling is then used to construct an FC network, which is in turn vectorised and used to train a classifier. We show that fusing features from networks improves classification accuracy. Although both within-frequency and cross-frequency networks can be used to predict AD with high accuracy, our results show that bispectrum-based FC outperforms cross-spectrum suggesting an important role of cross-frequency FC.

q-bio.NC↗

arXiv2022

Lumbar Bone Mineral Density Estimation from Chest X-ray Images: Anatomy-aware Attentive Multi-ROI Modeling

Osteoporosis is a common chronic metabolic bone disease often under-diagnosed and under-treated due to the limited access to bone mineral density (BMD) examinations, e.g. via Dual-energy X-ray Absorptiometry (DXA). This paper proposes a method to predict BMD from Chest X-ray (CXR), one of the most commonly accessible and low-cost medical imaging examinations. Our method first automatically detects Regions of Interest (ROIs) of local CXR bone structures. Then a multi-ROI deep model with transformer encoder is developed to exploit both local and global information in the chest X-ray image for accurate BMD estimation. Our method is evaluated on 13719 CXR patient cases with ground truth BMD measured by the gold standard DXA. The model predicted BMD has a strong correlation with the ground truth (Pearson correlation coefficient 0.894 on lumbar 1). When applied in osteoporosis screening, it achieves a high classification performance (average AUC of 0.968). As the first effort of using CXR scans to predict the BMD, the proposed algorithm holds strong potential for early osteoporosis screening and public health promotion.

eess.IV↗