arXiv subjects

Zi Wang

Publications and source records attributed to Zi Wang.

At least 127 records · Page 7Linked to original sources

arXiv2023

Efficient Symbolic Reasoning for Neural-Network Verification

The neural network has become an integral part of modern software systems. However, they still suffer from various problems, in particular, vulnerability to adversarial attacks. In this work, we present a novel program reasoning framework for neural-network verification, which we refer to as symbolic reasoning. The key components of our framework are the use of the symbolic domain and the quadratic relation. The symbolic domain has very flexible semantics, and the quadratic relation is quite expressive. They allow us to encode many verification problems for neural networks as quadratic programs. Our scheme then relaxes the quadratic programs to semidefinite programs, which can be efficiently solved. This framework allows us to verify various neural-network properties under different scenarios, especially those that appear challenging for non-symbolic domains. Moreover, it introduces new representations and perspectives for the verification tasks. We believe that our framework can bring new theoretical insights and practical tools to verification problems for neural networks.

cs.AI↗

arXiv2023

TBP-Former: Learning Temporal Bird's-Eye-View Pyramid for Joint Perception and Prediction in Vision-Centric Autonomous Driving

Vision-centric joint perception and prediction (PnP) has become an emerging trend in autonomous driving research. It predicts the future states of the traffic participants in the surrounding environment from raw RGB images. However, it is still a critical challenge to synchronize features obtained at multiple camera views and timestamps due to inevitable geometric distortions and further exploit those spatial-temporal features. To address this issue, we propose a temporal bird's-eye-view pyramid transformer (TBP-Former) for vision-centric PnP, which includes two novel designs. First, a pose-synchronized BEV encoder is proposed to map raw image inputs with any camera pose at any time to a shared and synchronized BEV space for better spatial-temporal synchronization. Second, a spatial-temporal pyramid transformer is introduced to comprehensively extract multi-scale BEV features and predict future BEV states with the support of spatial-temporal priors. Extensive experiments on nuScenes dataset show that our proposed framework overall outperforms all state-of-the-art vision-based prediction methods.

cs.CV↗

arXiv2022

Bridging the Domain Gap in Satellite Pose Estimation: a Self-Training Approach based on Geometrical Constraints

Recently, unsupervised domain adaptation in satellite pose estimation has gained increasing attention, aiming at alleviating the annotation cost for training deep models. To this end, we propose a self-training framework based on the domain-agnostic geometrical constraints. Specifically, we train a neural network to predict the 2D keypoints of a satellite and then use PnP to estimate the pose. The poses of target samples are regarded as latent variables to formulate the task as a minimization problem. Furthermore, we leverage fine-grained segmentation to tackle the information loss issue caused by abstracting the satellite as sparse keypoints. Finally, we iteratively solve the minimization problem in two steps: pseudo-label generation and network training. Experimental results show that our method adapts well to the target domain. Moreover, our method won the 1st place on the sunlamp task of the second international Satellite Pose Estimation Competition.

cs.CV↗

arXiv2022

A Paired Phase and Magnitude Reconstruction for Advanced Diffusion-Weighted Imaging

Objective: Multi-shot interleaved echo planer imaging can obtain diffusion-weighted images (DWI) with high spatial resolution and low distortion, but suffers from ghost artifacts introduced by phase variations between shots. In this work, we aim at solving the challenging reconstructions under inter-shot motions between shots and a low signal-to-noise ratio. Methods: An explicit phase model with paired phase and magnitude priors is proposed to regularize the reconstruction (PAIR). The former prior is derived from the smoothness of the shot phase and enforced with low-rankness in the k-space domain. The latter explores similar edges among multi-b-value and multi-direction DWI with weighted total variation in the image domain. Results: Extensive simulation and in vivo results show that PAIR can remove ghost artifacts very well under a high number of shots (8 shots) and significantly suppress the noise under the ultra-high b-value (4000 s/mm2). Conclusion: The explicit phase model PAIR with complementary priors has a good performance on challenging reconstructions under inter-shot motions between shots and a low signal-to-noise ratio. Significance: PAIR has great potential in advanced clinical DWI applications and brain function research.

eess.SP↗

arXiv2022

Magnetic Resonance Spectroscopy Deep Learning Denoising Using Few In Vivo Data

Magnetic Resonance Spectroscopy (MRS) is a noninvasive tool to reveal metabolic information. One challenge of 1H-MRS is the low Signal-Noise Ratio (SNR). To improve the SNR, a typical approach is to perform Signal Averaging (SA) with M repeated samples. The data acquisition time, however, is increased by M times accordingly, and a complete clinical MRS scan takes approximately 10 minutes at a common setting M=128. Recently, deep learning has been introduced to improve the SNR but most of them use the simulated data as the training set. This may hinder the MRS applications since some potential differences, such as acquisition system imperfections, and physiological and psychologic conditions may exist between the simulated and in vivo data. Here, we proposed a new scheme that purely used the repeated samples of realistic data. A deep learning model, Refusion Long Short-Term Memory (ReLSTM), was designed to learn the mapping from the low SNR time-domain data (24 SA) to the high SNR one (128 SA). Experiments on the in vivo brain spectra of 7 healthy subjects, 2 brain tumor patients and 1 cerebral infarction patient showed that only using 20% repeated samples, the denoised spectra by ReLSTM could provide comparable estimated concentrations of metabolites to 128 SA. Compared with the state-of-the-art low-rank denoising method, the ReLSTM achieved the lower relative error and the Cramér-Rao lower bounds in quantifying some important biomarkers. In summary, ReLSTM can perform high-fidelity denoising of the spectra under fast acquisition (24 SA), which would be valuable to MRS clinical studies.

physics.med-ph↗

arXiv2022

Towards Learning Universal Hyperparameter Optimizers with Transformers

Meta-learning hyperparameter optimization (HPO) algorithms from prior experiments is a promising approach to improve optimization efficiency over objective functions from a similar distribution. However, existing methods are restricted to learning from experiments sharing the same set of hyperparameters. In this paper, we introduce the OptFormer, the first text-based Transformer HPO framework that provides a universal end-to-end interface for jointly learning policy and function prediction when trained on vast tuning data from the wild, such as Google's Vizier database, one of the world's largest HPO datasets. Our extensive experiments demonstrate that the OptFormer can simultaneously imitate at least 7 different HPO algorithms, which can be further improved via its function uncertainty estimates. Compared to a Gaussian Process, the OptFormer also learns a robust prior distribution for hyperparameter response functions, and can thereby provide more accurate and better calibrated predictions. This work paves the path to future extensions for training a Transformer-based model as a general HPO optimizer.

cs.LG↗

arXiv2022

A Quantitative Geometric Approach to Neural-Network Smoothness

Fast and precise Lipschitz constant estimation of neural networks is an important task for deep learning. Researchers have recently found an intrinsic trade-off between the accuracy and smoothness of neural networks, so training a network with a loose Lipschitz constant estimation imposes a strong regularization and can hurt the model accuracy significantly. In this work, we provide a unified theoretical framework, a quantitative geometric approach, to address the Lipschitz constant estimation. By adopting this framework, we can immediately obtain several theoretical results, including the computational hardness of Lipschitz constant estimation and its approximability. Furthermore, the quantitative geometric perspective can also provide some insights into recent empirical observations that techniques for one norm do not usually transfer to another one. We also implement the algorithms induced from this quantitative geometric approach in a tool GeoLIP. These algorithms are based on semidefinite programming (SDP). Our empirical evaluation demonstrates that GeoLIP is more scalable and precise than existing tools on Lipschitz constant estimation for $\ell_\infty$-perturbations. Furthermore, we also show its intricate relations with other recent SDP-based techniques, both theoretically and empirically. We believe that this unified quantitative geometric perspective can bring new insights and theoretical tools to the investigation of neural-network smoothness and robustness.

cs.LG↗

arXiv2022

Multitask quantum thermal machines and cooperative effects

Including phonon-assisted inelastic process in thermoelectric devices is able to enhance the performance of nonequilibrium work extraction. In this work, we demonstrate that inelastic phonon-thermoelectric devices have a fertile functionality diagram, where particle current and phononic heat currents are coupled and fueled by chemical potential difference. Such devices can simultaneously perform multiple tasks, e.g., heat engines, refrigerators, and heat pumps. Guided by the entropy production, we mainly study the efficiencies and coefficients of performance of multitask quantum thermal machines, where the roles of the inelastic scattering process and multiple biases in multiterminal setups are emphasized. Specifically, in a three-terminal double-quantum-dot setup with a tunable gate, we show that it efficiently performs two useful tasks due to the phonon-assisted inelastic process. Moreover, the cooperation between the longitudinal and transverse thermoelectric effects in the three-terminal thermoelectric systems leads to markedly improved performance of the thermal machines. While for the four-terminal four-quantum-dot thermoelectric setup, we find that additional thermodynamic affinity furnishes the system with both enriched functionality and enhanced efficiency. Our work provides insights into optimizing phonon-thermoelectric devices.

cond-mat.mes-hall↗

arXiv2022

Plex: Towards Reliability using Pretrained Large Model Extensions

A recent trend in artificial intelligence is the use of pretrained models for language and vision tasks, which have achieved extraordinary performance but also puzzling failures. Probing these models' abilities in diverse ways is therefore critical to the field. In this paper, we explore the reliability of models, where we define a reliable model as one that not only achieves strong predictive performance but also performs well consistently over many decision-making tasks involving uncertainty (e.g., selective prediction, open set recognition), robust generalization (e.g., accuracy and proper scoring rules such as log-likelihood on in- and out-of-distribution datasets), and adaptation (e.g., active learning, few-shot uncertainty). We devise 10 types of tasks over 40 datasets in order to evaluate different aspects of reliability on both vision and language domains. To improve reliability, we developed ViT-Plex and T5-Plex, pretrained large model extensions for vision and language modalities, respectively. Plex greatly improves the state-of-the-art across reliability tasks, and simplifies the traditional protocol as it improves the out-of-the-box performance and does not require designing scores or tuning the model for each task. We demonstrate scaling effects over model sizes up to 1B parameters and pretraining dataset sizes up to 4B examples. We also demonstrate Plex's capabilities on challenging tasks including zero-shot open set recognition, active learning, and uncertainty in conversational language understanding.

cs.LG↗

arXiv2022

Pre-training helps Bayesian optimization too

Bayesian optimization (BO) has become a popular strategy for global optimization of many expensive real-world functions. Contrary to a common belief that BO is suited to optimizing black-box functions, it actually requires domain knowledge on characteristics of those functions to deploy BO successfully. Such domain knowledge often manifests in Gaussian process priors that specify initial beliefs on functions. However, even with expert knowledge, it is not an easy task to select a prior. This is especially true for hyperparameter tuning problems on complex machine learning models, where landscapes of tuning objectives are often difficult to comprehend. We seek an alternative practice for setting these functional priors. In particular, we consider the scenario where we have data from similar functions that allow us to pre-train a tighter distribution a priori. To verify our approach in realistic model training setups, we collected a large multi-task hyperparameter tuning dataset by training tens of thousands of configurations of near-state-of-the-art models on popular image and text datasets, as well as a protein sequence dataset. Our results show that on average, our method is able to locate good hyperparameters at least 3 times more efficiently than the best competing methods.

cs.LG↗

arXiv2022

Physics-driven Synthetic Data Learning for Biomedical Magnetic Resonance

Deep learning has innovated the field of computational imaging. One of its bottlenecks is unavailable or insufficient training data. This article reviews an emerging paradigm, imaging physics-based data synthesis (IPADS), that can provide huge training data in biomedical magnetic resonance without or with few real data. Following the physical law of magnetic resonance, IPADS generates signals from differential equations or analytical solution models, making the learning more scalable, explainable, and better protecting privacy. Key components of IPADS learning, including signal generation models, basic deep learning network structures, enhanced data generation, and learning methods are discussed. Great potentials of IPADS have been demonstrated by representative applications in fast imaging, ultrafast signal reconstruction and accurate parameter quantification. Finally, open questions and future work have been discussed.

cs.LG↗

arXiv2022

Ear Wearable (Earable) User Authentication via Acoustic Toothprint

Earables (ear wearables) is rapidly emerging as a new platform encompassing a diverse range of personal applications. The traditional authentication methods hence become less applicable and inconvenient for earables due to their limited input interface. Nevertheless, earables often feature rich around-the-head sensing capability that can be leveraged to capture new types of biometrics. In this work, we proposeToothSonic which leverages the toothprint-induced sonic effect produced by users performing teeth gestures for earable authentication. In particular, we design representative teeth gestures that can produce effective sonic waves carrying the information of the toothprint. To reliably capture the acoustic toothprint, it leverages the occlusion effect of the ear canal and the inward-facing microphone of the earables. It then extracts multi-level acoustic features to reflect the intrinsic toothprint information for authentication. The key advantages of ToothSonic are that it is suitable for earables and is resistant to various spoofing attacks as the acoustic toothprint is captured via the user's private teeth-ear channel that modulates and encrypts the sonic waves. Our experiment studies with 25 participants show that ToothSonic achieves up to 95% accuracy with only one of the users' tooth gestures.

cs.CR↗

arXiv2022

Enhancing Innate and Adaptive Immune Systems by Cold Atmospheric Plasma (CAP) and Its Antitumor Immunity

Cold atmospheric plasma (CAP) is a near room temperature ionized gas, generated under non-equilibrium discharge conditions. Here we show that a short exposure of rat peritoneal exudate macrophages and T-cells to CAP in vitro, triggered an inflammatory phenotype leading to better antigen-presenting and effector cell function respectively. Different from previous studies mainly using immortalized cell lines, both macrophage and T-cells in this study were primary cells isolated from mice. Furthermore, ex-vivo exposure of T-cells to CAP, followed by their adoptive transfer into tumor-bearing mice resulted in a strong antitumor effect in vivo. Mechanistically, CAP seems to disrupt tolerogenic pathways leading to enhanced production of pro-inflammatory cytokines while limiting the production of anti-inflammatory cytokines and the expression of inhibitory molecules such as programmed death-ligand 1 (PD-L1). CAP represents therefore a novel, non-toxic and easy to deliver technology to augment the function of immune cells and enhance antitumor responses when used as a component of T-cell adoptive immunotherapies strategies or, potentially in combination with other cancer immunotherapeutic approaches.

physics.med-ph↗

arXiv2022

A Sparse Model-inspired Deep Thresholding Network for Exponential Signal Reconstruction -- Application in Fast Biological Spectroscopy

The non-uniform sampling is a powerful approach to enable fast acquisition but requires sophisticated reconstruction algorithms. Faithful reconstruction from partial sampled exponentials is highly expected in general signal processing and many applications. Deep learning has shown astonishing potential in this field but many existing problems, such as lack of robustness and explainability, greatly limit its applications. In this work, by combining merits of the sparse model-based optimization method and data-driven deep learning, we propose a deep learning architecture for spectra reconstruction from undersampled data, called MoDern. It follows the iterative reconstruction in solving a sparse model to build the neural network and we elaborately design a learnable soft-thresholding to adaptively eliminate the spectrum artifacts introduced by undersampling. Extensive results on both synthetic and biological data show that MoDern enables more robust, high-fidelity, and ultra-fast reconstruction than the state-of-the-art methods. Remarkably, MoDern has a small number of network parameters and is trained on solely synthetic data while generalizing well to biological data in various scenarios. Furthermore, we extend it to an open-access and easy-to-use cloud computing platform (XCloud-MoDern), contributing a promising strategy for further development of biological applications.

cs.LG↗

arXiv2022

Geometric thermodynamic uncertainty relation in periodically driven thermoelectric heat engine

Thermodynamic uncertainty relation, quantifying a trade-off among average current, the associated fluctuation (precision), and entropy production (cost), has been formulated in nonequilibrium steady state and various stochastic systems. Herein, we study the thermodynamic uncertainty relation in generic thermoelectric heat engines under a periodic control protocol, by uncovering the underlying Berry-phase-like contribution. We show that our thermodynamic uncertainty relation breaks the seminal steady-state results, originating from the non-vanishing geometric effect. Furthermore, by deriving the consequent trade-off relation binding efficiency, power, and constancy, we prove that the periodically driven thermoelectric heat engines can generally outperform the steady-state analogies. The general bounds are illustrated by an analytically solvable two-terminal single quantum dot heat engine under the periodic modulation. Our work provides a geometric framework in bounding and optimizing a wide range of periodically driven thermoelectric thermal machines.

cond-mat.mes-hall↗

arXiv2021

Onsite Non-Line-of-Sight Imaging via Online Calibrations

There has been an increasing interest in deploying non-line-of-sight (NLOS) imaging systems for recovering objects behind an obstacle. Existing solutions generally pre-calibrate the system before scanning the hidden objects. Onsite adjustments of the occluder, object and scanning pattern require re-calibration. We present an online calibration technique that directly decouples the acquired transients at onsite scanning into the LOS and hidden components. We use the former to directly (re)calibrate the system upon changes of scene/obstacle configurations, scanning regions, and scanning patterns whereas the latter for hidden object recovery via spatial, frequency or learning based techniques. Our technique avoids using auxiliary calibration apparatus such as mirrors or checkerboards and supports both laboratory validations and real-world deployments.

eess.IV↗

arXiv2021

Exponential Signal Reconstruction with Deep Hankel Matrix Factorization

Exponential is a basic signal form, and how to fast acquire this signal is one of the fundamental problems and frontiers in signal processing. To achieve this goal, partial data may be acquired but result in the severe artifacts in its spectrum, which is the Fourier transform of exponentials. Thus, reliable spectrum reconstruction is highly expected in the fast sampling in many applications, such as chemistry, biology, and medical imaging. In this work, we propose a deep learning method whose neural network structure is designed by unrolling the iterative process in the model-based state-of-the-art exponentials reconstruction method with low-rank Hankel matrix factorization. With the experiments on synthetic data and realistic biological magnetic resonance signals, we demonstrate that the new method yields much lower reconstruction errors and preserves the low-intensity signals much better.

eess.SP↗

arXiv2021

One-dimensional Deep Low-rank and Sparse Network for Accelerated MRI

Deep learning has shown astonishing performance in accelerated magnetic resonance imaging (MRI). Most state-of-the-art deep learning reconstructions adopt the powerful convolutional neural network and perform 2D convolution since many magnetic resonance images or their corresponding k-space are in 2D. In this work, we present a new approach that explores the 1D convolution, making the deep network much easier to be trained and generalized. We further integrate the 1D convolution into the proposed deep network, named as One-dimensional Deep Low-rank and Sparse network (ODLS), which unrolls the iteration procedure of a low-rank and sparse reconstruction model. Extensive results on in vivo knee and brain datasets demonstrate that, the proposed ODLS is very suitable for the case of limited training subjects and provides improved reconstruction performance than state-of-the-art methods both visually and quantitatively. Additionally, ODLS also shows nice robustness to different undersampling scenarios and some mismatches between the training and test data. In summary, our work demonstrates that the 1D deep learning scheme is memory-efficient and robust in fast MRI.

eess.IV↗