arXiv subjects

Xiang Yin

Publications and source records attributed to Xiang Yin.

At least 127 records · Page 7Linked to original sources

arXiv2022

A Unified Framework for Verification of Observational Properties for Partially-Observed Discrete-Event Systems

In this paper, we investigate property verification problems in partially-observed discrete-event systems (DES). Particularly, we are interested in verifying observational properties that are related to the information-flow of the system. Observational properties considered here include diagnosability, predictability, detectability and opacity, which have drawn considerable attentions in the literature. However, in contrast to existing results, where different verification procedures are developed for different properties case-by-case, in this work, we provide a unified framework for verifying all these properties by reducing each of them as an instance of HyperLTL model checking. Our approach is based on the construction of a Kripke structure that effectively captures the issue of unobservability as well as the finite string semantics in partially-observed DES so that HyperLTL model checking techniques can be suitably applied. Then for each observational property considered, we explicitly provide the HyperLTL formula to be checked over the Kripke structure for the purpose of verification. Our approach is uniform in the sense that all different properties can be verified with the same model checking engine. Furthermore, our unified framework also brings new insights for classifying observational properties for partially-observed DES in terms of their verification complexity.

eess.SY↗

arXiv2022

A Novel Chinese Dialect TTS Frontend with Non-Autoregressive Neural Machine Translation

Chinese dialects are different variations of Chinese and can be considered as different languages in the same language family with Mandarin. Though they all use Chinese characters, the pronunciations, grammar and idioms can vary significantly, and even local speakers may find it hard to input correct written forms of dialect. Besides, using Mandarin text as text-to-speech inputs would generate speech with poor naturalness. In this paper, we propose a novel Chinese dialect TTS frontend with a translation module, which converts Mandarin text into dialectic expressions to improve the intelligibility and naturalness of synthesized speech. A non-autoregressive neural machine translation model with various tricks is proposed for the translation task. It is the first known work to incorporate translation with TTS frontend. Experiments on Cantonese show the proposed model improves 2.56 BLEU and TTS improves 0.27 MOS with Mandarin inputs.

cs.CL↗

arXiv2022

Direct Speech-to-speech Translation without Textual Annotation using Bottleneck Features

Speech-to-speech translation directly translates a speech utterance to another between different languages, and has great potential in tasks such as simultaneous interpretation. State-of-art models usually contains an auxiliary module for phoneme sequences prediction, and this requires textual annotation of the training dataset. We propose a direct speech-to-speech translation model which can be trained without any textual annotation or content information. Instead of introducing an auxiliary phoneme prediction task in the model, we propose to use bottleneck features as intermediate training objectives for our model to ensure the translation performance of the system. Experiments on Mandarin-Cantonese speech translation demonstrate the feasibility of the proposed approach and the performance can match a cascaded system with respect of translation and synthesis qualities.

cs.CL↗

arXiv2022

Explaining Random Forests using Bipolar Argumentation and Markov Networks (Technical Report)

Random forests are decision tree ensembles that can be used to solve a variety of machine learning problems. However, as the number of trees and their individual size can be large, their decision making process is often incomprehensible. In order to reason about the decision process, we propose representing it as an argumentation problem. We generalize sufficient and necessary argumentative explanations using a Markov network encoding, discuss the relevance of these explanations and establish relationships to families of abductive explanations from the literature. As the complexity of the explanation problems is high, we discuss a probabilistic approximation algorithm and present first experimental results.

cs.AI↗

arXiv2022

Model Predictive Control for Signal Temporal Logic Specifications with Time Interval Decomposition

In this paper, we investigate the problem of Model Predictive Control (MPC) of dynamic systems for high-level specifications described by Signal Temporal Logic (STL) formulae. Recent works show that MPC has the great potential in handling logical tasks in reactive environments. However, existing approaches suffer from the heavy computational burden, especially for tasks with large horizons. In this work, we propose a computationally more efficient MPC framework for STL tasks based on time interval decomposition. Specifically, we still use the standard shrink horizon MPC framework with Mixed Integer Linear Programming (MILP) techniques for open-loop optimization problems. However, instead of applying MPC directly for the entire task horizon, we decompose the STL formula into several sub-formulae with disjoint time horizons, and shrinking horizon MPC is applied for each short-horizon sub-formula iteratively. To guarantee the satisfaction of the entire STL formula and to ensure the recursive feasibility of the iterative process, we introduce new terminal constraints to connect each sub-formula. We show how these terminal constraints can be computed by an effective inner-approximation approach. The computational efficiency of our approach is illustrated by a case study.

eess.SY↗

arXiv2022

Abstraction-Based Verification of Approximate Pre-Opacity for Control Systems

In this paper, we consider the problem of verifying pre-opacity for discrete-time control systems. Pre-opacity is an important information-flow security property that secures the intention of a system to execute some secret behaviors in the future. Existing works on pre-opacity only consider non-metric discrete systems, where it is assumed that intruders can distinguish different output behaviors precisely. However, for continuous-space control systems whose output sets are equipped with metrics (which is the case for most real-world applications), it is too restrictive to assume precise measurements from outside observers. In this paper, we first introduce a concept of approximate pre-opacity by capturing the security level of control systems with respect to the measurement precision of the intruder. Based on this new notion of pre-opacity, we propose a verification approach for continuous-space control systems by leveraging abstraction-based techniques. In particular, a new concept of approximate pre-opacity preserving simulation relation is introduced to characterize the distance between two systems in terms of preserving pre-opacity. This new system relation allows us to verify pre-opacity of complex continuous-space control systems using their finite abstractions. We also present a method to construct pre-opacity preserving finite abstractions for a class of discrete-time control systems under certain stability assumptions.

eess.SY↗

arXiv2022

Towards a Theory of Faithfulness: Faithful Explanations of Differentiable Classifiers over Continuous Data

There is broad agreement in the literature that explanation methods should be faithful to the model that they explain, but faithfulness remains a rather vague term. We revisit faithfulness in the context of continuous data and propose two formal definitions of faithfulness for feature attribution methods. Qualitative faithfulness demands that scores reflect the true qualitative effect (positive vs. negative) of the feature on the model and quanitative faithfulness that the magnitude of scores reflect the true quantitative effect. We discuss under which conditions these requirements can be satisfied to which extent (local vs global). As an application of the conceptual idea, we look at differentiable classifiers over continuous data and characterize Gradient-scores as follows: every qualitatively faithful feature attribution method is qualitatively equivalent to Gradient-scores. Furthermore, if an attribution method is quantitatively faithful in the sense that changes of the output of the classifier are proportional to the scores of features, then it is either equivalent to gradient-scoring or it is based on an inferior approximation of the classifier. To illustrate the practical relevance of the theory, we experimentally demonstrate that popular attribution methods can fail to give faithful explanations in the setting where the data is continuous and the classifier differentiable.

cs.LG↗

arXiv2022

A Uniform Framework for Diagnosis of Discrete-Event Systems with Unreliable Sensors using Linear Temporal Logic

In this paper, we investigate the diagnosability verification problem of partially-observed discrete-event systems (DES) subject to unreliable sensors. In this setting, upon the occurrence of each event, the sensor reading may be non-deterministic due to measurement noises or possible sensor failures. Existing works on this topic mainly consider specific types of unreliable sensors such as the cases of intermittent sensors failures, permanent sensor failures or their combinations. In this work, we propose a novel \emph{uniform framework} for diagnosability of DES subject to, not only sensor failures, but also a very general class of unreliable sensors. Our approach is to use linear temporal logic (LTL) with semantics on infinite traces to describe the possible behaviors of the sensors. A new notion of $φ$-diagnosability is proposed as the necessary and sufficient condition for the existence of a diagnoser when the behaviors of sensors satisfy the LTL formula $φ$. Effective approach is provided to verify this notion. We show that, our new notion of $φ$-diagnosability subsumes all existing notions of robust diagnosability of DES subject to sensor failures. Furthermore, the proposed framework is user-friendly and flexible since it supports an arbitrary user-defined unreliable sensor type based on the specific scenario of the application. As examples, we provide two new notions of diagnosability, which have never been investigated in the literature, using our uniform framework.

eess.SY↗

arXiv2022

Fault Diagnosis of Discrete-Event Systems under Non-Deterministic Observations with Output Fairness

In this paper, we revisit the fault diagnosis problem of discrete-event systems (DES) under non-deterministic observations. Non-deterministic observation is a general observation model that includes the case of intermittent loss of observations. In this setting, upon the occurrence of an event, the sensor reading may be non-deterministic such that a set of output symbols are all possible. Existing works on fault diagnosis under non-deterministic observations require to consider all possible observation realizations. However, this approach includes the case where some possible outputs are permanently disabled. In this work, we introduce the concept of output fairness by requiring that, for any output symbols, if it has infinite chances to be generated, then it will indeed be generated infinite number of times. We use an assume-guarantee type of linear temporal logic formulas to formally describe this assumption. A new notion called output-fair diagnosability (OF-diagnosability) is proposed. An effective approach is provided for the verification of OF-diagnosability. We show that the proposed notion of OF-diagnosability is weaker than the standard definition of diagnosability under non-deterministic observations, and it better captures the physical scenario of observation non-determinism or intermittent loss of observations.

eess.SY↗

arXiv2022

You Don't Know What I Know: On Notion of High-Order Opacity in Discrete-Event Systems

In this paper, we investigate a class of information-flow security properties called opacity in partial-observed discrete-event systems. Roughly speaking, a system is said to be opaque if the intruder, which is modeled by a passive observer, can never determine the "secret" of the system for sure. Most of the existing notions of opacity consider secrets related to the actual behaviors of the system. In this paper, we consider a new type of secret related to the knowledge of the system user. Specifically, we assume that the system user also only has partial observation of the system and has to reason the actual behavior of the system. We say a system is high-order opaque if the intruder can never determine that the system user knows some information of importance based on its own incomparable information. We provide the formal definition of high-order opacity. Two algorithms are provided for the verification of this new notion: one with doubly-exponential complexity for the worst case and the other with single-exponential complexity. Illustrative examples are provided for the new notion of high-order opacity.

eess.SY↗

arXiv2022

Sensor Deception Attacks Against Initial-State Privacy in Supervisory Control Systems

This paper investigates the problem of synthesizing sensor deception attackers against privacy in the context of supervisory control of discrete-event systems (DES). We consider a DES plant controlled by a supervisor, which is subject to sensor deception attacks. Specifically, we consider an active attacker that can tamper with the observations received by the supervisor by, e.g., hacking on the communication channel between the sensors and the supervisor. The privacy requirement of the supervisory control system is to maintain initial-state opacity, i.e., it does not want to reveal the fact that it was initiated from a secret state during its operation. On the other hand, the attacker aims to deceive the supervisor, by tampering with its observations, such that initial-state opacity is violated due to incorrect control actions. In this work, we investigate from the attacker's point of view by presenting an effective approach for synthesizing sensor attack strategies threatening the privacy of the system. To this end, we propose the All Attack Structure (AAS) that records state estimates for both the supervisor and the attacker. This structure serves as a basis for synthesizing a sensor attack strategy. We also discuss how to simplify the synthesis complexity by leveraging the structural property of the initial-state privacy requirement. A running academic example is provided to illustrate the synthesis procedure.

eess.SY↗

arXiv2022

Online Monitoring of Dynamic Systems for Signal Temporal Logic Specifications with Model Information

Online monitoring aims to evaluate or to predict, at runtime, whether or not the behaviors of a system satisfy some desired specification. It plays a key role in safety-critical cyber-physical systems. In this work, we propose a new model-based approach for online monitoring for specifications described by Signal Temporal Logic (STL) formulae. Specifically, we assume that the observed state traces are generated by an underlying dynamic system whose model is known. The main idea is to consider the dynamic of the system when evaluating the satisfaction of the STL formulae. To this end, effective approaches for the computation of feasible sets for STL formulae are provided. We show that, by explicitly utilizing the model information of the dynamic system, the proposed online monitoring algorithm can falsify or certify of the specification in advance compared with existing algorithms, where no model information is used. We also demonstrate the proposed monitoring algorithm by case studies.

eess.SY↗

arXiv2022

Secure-by-Construction Synthesis of Cyber-Physical Systems

Correct-by-construction synthesis is a cornerstone of the confluence of formal methods and control theory towards designing safety-critical systems. Instead of following the time-tested, albeit laborious (re)design-verify-validate loop, correct-by-construction methodology advocates the use of continual refinements of formal requirements -- connected by chains of formal proofs -- to build a system that assures the correctness by design. A remarkable progress has been made in scaling the scope of applicability of correct-by-construction synthesis -- with a focus on cyber-physical systems that tie discrete-event control with continuous environment -- to enlarge control systems by combining symbolic approaches with principled state-space reduction techniques. Unfortunately, in the security-critical control systems, the security properties are verified ex post facto the design process in a way that undermines the correct-by-construction paradigm. We posit that, to truly realize the dream of correct-by-construction synthesis for security-critical systems, security considerations must take center-stage with the safety considerations. Moreover, catalyzed by the recent progress on the opacity sub-classes of security properties and the notion of hyperproperties capable of combining security with safety properties, we believe that the time is ripe for the research community to holistically target the challenge of secure-by-construction synthesis. This paper details our vision by highlighting the recent progress and open challenges that may serve as bricks for providing a solid foundation for secure-by-construction synthesis of cyber-physical systems.

cs.CR↗

arXiv2022

A Framework for Current-State Opacity under Dynamic Information Release Mechanism

Opacity is an important information-flow security property that characterizes the plausible deniability of a dynamic system for its "secret" against eavesdropping attacks. As an information-flow property, the underlying observation model is the key in the modeling and analysis of opacity. In this paper, we investigate the verification of current-state opacity for discrete-event systems under Orwellian-type observations, i.e., the system is allowed to re-interpret the observation of an event based on its future suffix. First, we propose a new Orwellian-type observation model called the dynamic information release mechanism (DIRM). In the DIRM, when to release previous "hold on" events is state-dependent. Then we propose a new definition of opacity based on the notion of history-equivalence rather than the standard projection-equivalence. This definition is more suitable for observations that are not prefix-closed. Finally, we show that by constructing a new structure called the DIRM-observer, current-state opacity can be effectively verified under the DIRM. Computational complexity analysis as well as illustrative examples for the proposed approach are also provided. Compared with the existing Orwellian-type observation model, the proposed framework is more general in the sense that the information-release-mechanism is state-dependent and the corresponding definition of opacity is more suitable for non-prefix-closed observations.

eess.SY↗

arXiv2022

Towards Realistic Visual Dubbing with Heterogeneous Sources

The task of few-shot visual dubbing focuses on synchronizing the lip movements with arbitrary speech input for any talking head video. Albeit moderate improvements in current approaches, they commonly require high-quality homologous data sources of videos and audios, thus causing the failure to leverage heterogeneous data sufficiently. In practice, it may be intractable to collect the perfect homologous data in some cases, for example, audio-corrupted or picture-blurry videos. To explore this kind of data and support high-fidelity few-shot visual dubbing, in this paper, we novelly propose a simple yet efficient two-stage framework with a higher flexibility of mining heterogeneous data. Specifically, our two-stage paradigm employs facial landmarks as intermediate prior of latent representations and disentangles the lip movements prediction from the core task of realistic talking head generation. By this means, our method makes it possible to independently utilize the training corpus for two-stage sub-networks using more available heterogeneous data easily acquired. Besides, thanks to the disentanglement, our framework allows a further fine-tuning for a given talking head, thereby leading to better speaker-identity preserving in the final synthesized results. Moreover, the proposed method can also transfer appearance features from others to the target speaker. Extensive experimental results demonstrate the superiority of our proposed method in generating highly realistic videos synchronized with the speech over the state-of-the-art.

cs.CV↗

arXiv2021

Opacity Enforcing Supervisory Control using Non-deterministic Supervisors

In this paper, we investigate the enforcement of opacity via supervisory control in the context of discrete-event systems. A system is said to be opaque if the intruder, which is modeled as a passive observer, can never infer confidently that the system is at a secret state. The design objective is to synthesize a supervisor such that the closed-loop system is opaque even when the control policy is publicly known. In this paper, we propose a new approach for enforcing opacity using non-deterministic supervisors. A non-deterministic supervisor is a decision mechanism that provides a set of control decisions at each instant, and randomly picks a specific control decision from the decision set to actually control the plant. Compared with the standard deterministic control mechanism, such a non-deterministic control mechanism can enhance the plausible deniability of the controlled system as the online control decision is a random realization and cannot be implicitly inferred from the control policy. We provide a sound and complete algorithm for synthesizing a non-deterministic opacity-enforcing supervisor. Furthermore, we show that non-deterministic supervisors are strictly more powerful than deterministic supervisors in the sense that there may exist a non-deterministic opacity-enforcing supervisor even when deterministic supervisors cannot enforce opacity.

eess.SY↗

arXiv2021

Data privacy protection in microscopic image analysis for material data mining

Recent progress in material data mining has been driven by high-capacity models trained on large datasets. However, collecting experimental data has been extremely costly owing to the amount of human effort and expertise required. Therefore, material researchers are often reluctant to easily disclose their private data, which leads to the problem of data island, and it is difficult to collect a large amount of data to train high-quality models. In this study, a material microstructure image feature extraction algorithm FedTransfer based on data privacy protection is proposed. The core contributions are as follows: 1) the federated learning algorithm is introduced into the polycrystalline microstructure image segmentation task to make full use of different user data to carry out machine learning, break the data island and improve the model generalization ability under the condition of ensuring the privacy and security of user data; 2) A data sharing strategy based on style transfer is proposed. By sharing style information of images that is not urgent for user confidentiality, it can reduce the performance penalty caused by the distribution difference of data among different users.

eess.IV↗

arXiv2021

Towards Using Clothes Style Transfer for Scenario-aware Person Video Generation

Clothes style transfer for person video generation is a challenging task, due to drastic variations of intra-person appearance and video scenarios. To tackle this problem, most recent AdaIN-based architectures are proposed to extract clothes and scenario features for generation. However, these approaches suffer from being short of fine-grained details and are prone to distort the origin person. To further improve the generation performance, we propose a novel framework with disentangled multi-branch encoders and a shared decoder. Moreover, to pursue the strong video spatio-temporal consistency, an inner-frame discriminator is delicately designed with input being cross-frame difference. Besides, the proposed framework possesses the property of scenario adaptation. Extensive experiments on the TEDXPeople benchmark demonstrate the superiority of our method over state-of-the-art approaches in terms of image quality and video coherence.

cs.CV↗