Song Fang

Guanchong Huang, Song Fang

Video identification attacks pose a significant privacy threat that can reveal videos that victims watch, which may disclose their hobbies, religious beliefs, political leanings, sexual orientation, and health status. Also, video watching history can be used for user profiling or advertising and may result in cyberbullying, discrimination, or blackmail. Existing extensive video inference techniques usually depend on analyzing network traffic generated by streaming online videos. In this work, we observe that the content of a subtitle determines its silhouette displayed on the screen, and identifying each subtitle silhouette also derives the temporal difference between two consecutive subtitles. We then propose SilhouetteTell, a novel video identification attack that combines the spatial and time domain information into a spatiotemporal feature of subtitle silhouettes. SilhouetteTell explores the spatiotemporal correlation between recorded subtitle silhouettes of a video and its subtitle file. It can infer both online and offline videos. Comprehensive experiments on off-the-shelf smartphones confirm the high efficacy of SilhouetteTell for inferring video titles and clips under various settings, including from a distance of up to 40 meters.

Song Fang, Quanyan Zhu

In this paper, we relate the feedback capacity of parallel additive colored Gaussian noise (ACGN) channels to a variant of the Kalman filter. By doing so, we obtain lower bounds on the feedback capacity of such channels, as well as the corresponding feedback (recursive) coding schemes, which are essentially power allocation policies with feedback, to achieve the bounds. The results are seen to reduce to existing lower bounds in the case of a single ACGN feedback channel, whereas when it comes to parallel additive white Gaussian noise (AWGN) channels with feedback, the recursive coding scheme reduces to a feedback "water-filling" power allocation policy.

Song Fang, Quanyan Zhu

In this short note, we investigate the feedback control of relativistic dynamics propelled by mass ejection, modeling, e.g., the relativistic rocket control or the relativistic (space-travel) flight control. As an extreme case, we also examine the control of relativistic photon rockets which are propelled by ejecting photons.

Song Fang, Quanyan Zhu

In this paper, we examine the fundamental performance limitations in the control of stochastic dynamical systems; more specifically, we derive generic $\mathcal{L}_p$ bounds that hold for any causal (stabilizing) controllers and any stochastic disturbances, by an information-theoretic analysis. We first consider the scenario where the plant (i.e., the dynamical system to be controlled) is linear time-invariant, and it is seen in general that the lower bounds are characterized by the unstable poles (or nonminimum-phase zeros) of the plant as well as the conditional entropy of the disturbance. We then analyze the setting where the plant is assumed to be (strictly) causal, for which case the lower bounds are determined by the conditional entropy of the disturbance. We also discuss the special cases of $p = 2$ and $p = \infty$, which correspond to minimum-variance control and controlling the maximum deviations, respectively. In addition, we investigate the power-spectral characterization of the lower bounds as well as its relation to the Kolmogorov-Szegö formula.

Song Fang, Quanyan Zhu

In this short note, we introduce the spectral-domain $\mathcal{W}_2$ Wasserstein distance for elliptical stochastic processes in terms of their power spectra. We also introduce the spectral-domain Gelbrich bound for processes that are not necessarily elliptical.

Song Fang, Quanyan Zhu

This short note is on a property of the $\mathcal{W}_2$ Wasserstein distance which indicates that independent elliptical distributions minimize their $\mathcal{W}_2$ Wasserstein distance from given independent elliptical distributions with the same density generators. Furthermore, we examine the implications of this property in the Gelbrich bound when the distributions are not necessarily elliptical. Meanwhile, we also generalize the results to the cases when the distributions are not independent. The primary purpose of this note is for the referencing of papers that need to make use of this property or its implications.

Song Fang, Quanyan Zhu

In this paper, we analyze the fundamental stealthiness-distortion tradeoffs of linear Gaussian dynamical systems under data injection attacks using a power spectral analysis, whereas the Kullback-Leibler (KL) divergence is employed as the stealthiness measure. Particularly, we obtain explicit formulas in terms of power spectra that characterize analytically the stealthiness-distortion tradeoffs as well as the properties of the worst-case attacks. Furthermore, it is seen in general that the attacker only needs to know the input-output behaviors of the systems in order to carry out the worst-case attacks.

Song Fang, Quanyan Zhu

This short note is on a property of the Kullback-Leibler (KL) divergence which indicates that independent Gaussian distributions minimize the KL divergence from given independent Gaussian distributions. The primary purpose of this note is for the referencing of papers that need to make use of this property entirely or partially.

Song Fang, Quanyan Zhu

In this paper, we study the fundamental limits of obfuscation in terms of privacy-distortion tradeoffs for linear Gaussian dynamical systems via an information-theoretic approach. Particularly, we obtain analytical formulas that capture the fundamental privacy-distortion tradeoffs when privacy masks are to be added to the outputs of the dynamical systems, while indicating explicitly how to design the privacy masks in an optimal way: The privacy masks should be colored Gaussian with power spectra shaped specifically based upon the system and noise properties.

Song Fang, Quanyan Zhu

In this paper, we first introduce the notion of channel leakage as the minimum mutual information between the channel input and channel output. As its name indicates, channel leakage quantifies the minimum information leakage to the malicious receiver. In a broad sense, it can be viewed as a dual concept of channel capacity, which characterizes the maximum information transmission to the targeted receiver. We obtain explicit formulas of channel leakage for the white Gaussian case, the colored Gaussian case, and the fading case. We then utilize this notion to investigate the fundamental limitations of obfuscation in terms of privacy-distortion tradeoffs (as well as privacy-power tradeoffs) for streaming data; particularly, we derive analytical tradeoff equations for the stationary case, the non-stationary case, and the finite-time case. Our results also indicate explicitly how to design the privacy masks in an optimal way.

Song Fang, Quanyan Zhu

In this paper, we examine the fundamental performance limits of prediction, with or without side information. More specifically, we derive generic lower bounds on the $\mathcal{L}_p$ norms of the prediction errors that are valid for any prediction algorithms and for any data distributions. Meanwhile, we combine the entropic analysis from information theory and the innovations approach from prediction/estimation theory to characterize the conditions (in terms of, e.g., directed information or mutual information) to achieve the bounds. We also investigate the implications of the results in analyzing the fundamental limits of generalization in fitting (learning) problems from the perspective of prediction with side information, as well as the fundamental limits of recursive algorithms by viewing them as generalized prediction problems.

Song Fang, Quanyan Zhu

In this paper, we relate a feedback channel with any finite-order autoregressive moving-average (ARMA) Gaussian noises to a variant of the Kalman filter. In light of this, we obtain relatively explicit lower bounds on the feedback capacity for such colored Gaussian noises, and the bounds are seen to be consistent with various existing results in the literature. Meanwhile, this variant of the Kalman filter also leads to explicit recursive coding schemes with clear structures to achieve the lower bounds. In general, our results provide an alternative perspective while pointing to potentially tighter bounds for the feedback capacity problem.

Song Fang, Quanyan Zhu

In this paper, we utilize information theory to study the fundamental performance limitations of generic feedback systems, where both the controller and the plant may be any causal functions/mappings while the disturbance can be with any distributions. More specifically, we obtain fundamental $\mathcal{L}_p$ bounds on the control error, which are shown to be completely characterized by the conditional entropy of the disturbance, based upon the entropic laws that are inherent in any feedback systems. We also discuss the generality and implications (in, e.g., fundamental limits of learning-based control) of the obtained bounds.

Song Fang, Quanyan Zhu

Strictly speaking, Newton's second law of motion is only an approximation of the so-called relativistic dynamics, i.e., Einstein's modification of the second law based on his theory of special relativity. Although the approximation is almost exact when the velocity of the dynamical system is far less than the speed of light, the difference will become larger and larger (and will eventually go to infinity) as the velocity approaches the speed of light. Correspondingly, feedback control of such dynamics should also take this modification into consideration (though it will render the system nonlinear), especially when the velocity is relatively large. Towards this end, we start this note by studying the state-space representation of the relativistic dynamics. We then investigate on how to employ the feedback linearization approach for such relativistic dynamics, based upon which an additional linear controller may then be designed. As such, the feedback linearization together with the linear controller compose the overall relativistic feedback control law. We also provide discussions on, e.g., controllability, state feedback and output feedback, as well as PID control, in the relativistic setting.

Song Fang, Quanyan Zhu

In this paper, we obtain fundamental $\mathcal{L}_{p}$ bounds in sequential prediction and recursive algorithms via an entropic analysis. Both classes of problems are examined by investigating the underlying entropic relationships of the data and/or noises involved, and the derived lower bounds may all be quantified in a conditional entropy characterization. We also study the conditions to achieve the generic bounds from an innovations' viewpoint.

Song Fang, Quanyan Zhu

In this paper, we derive generic bounds on the maximum deviations in prediction errors for sequential prediction via an information-theoretic approach. The fundamental bounds are shown to depend only on the conditional entropy of the data point to be predicted given the previous data points. In the asymptotic case, the bounds are achieved if and only if the prediction error is white and uniformly distributed.

Song Fang, Karl Henrik Johansson, Mikael Skoglund et al.

In this paper, we introduce the concept of two-way coding, which originates in communication theory characterizing coding schemes for two-way channels, into control theory, particularly to facilitate the analysis and design of feedback control systems under injection attacks. Moreover, we propose the notion of attack decoupling, and show how the controller and the two-way coding can be co-designed to nullify the transfer function from attack to plant, rendering the attack effect zero both in transient phase and in steady state.

Song Fang, Mikael Skoglund, Karl Henrik Johansson et al.

In this paper, we obtain generic bounds on the variances of estimation and prediction errors in time series analysis via an information-theoretic approach. It is seen in general that the error bounds are determined by the conditional entropy of the data point to be estimated or predicted given the side information or past observations. Additionally, we discover that in order to achieve the prediction error bounds asymptotically, the necessary and sufficient condition is that the "innovation" is asymptotically white Gaussian. When restricted to Gaussian processes and 1-step prediction, our bounds are shown to reduce to the Kolmogorov-Szegö formula and Wiener-Masani formula known from linear prediction theory.

Song Fang, Hideaki Ishii, Jie Chen et al.

In this paper, we discover that the trace of the division of the optimal output estimation error covariance over the noise covariance attained by the Kalman-Bucy filter can be explicitly expressed in terms of the plant dynamics and noise statistics in a frequency-domain integral characterization. Towards this end, we examine the algebraic Riccati equation associated with Kalman-Bucy filtering using analytic function theory and relate it to the Bode integral. Our approach features an alternative, frequency-domain framework for analyzing algebraic Riccati equations and reduces to various existing related results.