Showing 1–10 of 10 results for author: Trevisan, D

Student-t processes as infinite-width limits of posterior Bayesian neural networks

Authors: Francesco Caporali, Stefano Favaro, Dario Trevisan

Abstract: The asymptotic properties of Bayesian Neural Networks (BNNs) have been extensively studied, particularly regarding their approximations by Gaussian processes in the infinite-width limit. We extend these results by showing that posterior BNNs can be approximated by Student-t processes, which offer greater flexibility in modeling uncertainty. Specifically, we show that, if the parameters of a BNN fo… ▽ More The asymptotic properties of Bayesian Neural Networks (BNNs) have been extensively studied, particularly regarding their approximations by Gaussian processes in the infinite-width limit. We extend these results by showing that posterior BNNs can be approximated by Student-t processes, which offer greater flexibility in modeling uncertainty. Specifically, we show that, if the parameters of a BNN follow a Gaussian prior distribution, and the variance of both the last hidden layer and the Gaussian likelihood function follows an Inverse-Gamma prior distribution, then the resulting posterior BNN converges to a Student-t process in the infinite-width limit. Our proof leverages the Wasserstein metric to establish control over the convergence rate of the Student-t process approximation. △ Less

Submitted 6 February, 2025; originally announced February 2025.

A guideline proposal for minimizing cybersickness in VR-based serious games and applications

Authors: Thiago Porcino, Derek Reilly, Esteban Clua, Daniela Trevisan

Abstract: Head-mounted displays (HMDs) are popular immersive tools in general, not limited to entertainment but also for education, military, and serious games for health. While these displays have strong popularity, they still have user experience issues, triggering possible symptoms of discomfort to users. This condition is known as cybersickness (CS) and is one of the most popular research topics tied to… ▽ More Head-mounted displays (HMDs) are popular immersive tools in general, not limited to entertainment but also for education, military, and serious games for health. While these displays have strong popularity, they still have user experience issues, triggering possible symptoms of discomfort to users. This condition is known as cybersickness (CS) and is one of the most popular research topics tied to virtual reality (VR) issues. We first present the main strategies focused on minimizing cybersickness problems in virtual reality. Following this, we propose a guideline framework based on CS causes such as locomotion, acceleration, the field of view, depth of field, degree of freedom, exposition use time, latency-lag, static rest frame, and camera rotation. Additionally, serious games applications and broader categories of games can also adopt it. Additionally, we categorized the imminent challenges for CS minimization into four different items. Conclusively, this work contributes as a consulting reference to enable VR developers and designers to optimize their VR users' experience and VR serious games. △ Less

Submitted 13 July, 2022; originally announced July 2022.

Comments: 8 pages

Report number: Accepted at the IEEE 10th International Conference on Serious Games and Applications for Health (SeGAH) - SeGAH 2020

Quantitative Gaussian Approximation of Randomly Initialized Deep Neural Networks

Authors: Andrea Basteri, Dario Trevisan

Abstract: Given any deep fully connected neural network, initialized with random Gaussian parameters, we bound from above the quadratic Wasserstein distance between its output distribution and a suitable Gaussian process. Our explicit inequalities indicate how the hidden and output layers sizes affect the Gaussian behaviour of the network and quantitatively recover the distributional convergence results in… ▽ More Given any deep fully connected neural network, initialized with random Gaussian parameters, we bound from above the quadratic Wasserstein distance between its output distribution and a suitable Gaussian process. Our explicit inequalities indicate how the hidden and output layers sizes affect the Gaussian behaviour of the network and quantitatively recover the distributional convergence results in the wide limit, i.e., if all the hidden layers sizes become large. △ Less

Submitted 22 September, 2023; v1 submitted 14 March, 2022; originally announced March 2022.

MSC Class: 60F05; 60G15; 68T99

The Quantum Wasserstein Distance of Order 1

Authors: Giacomo De Palma, Milad Marvian, Dario Trevisan, Seth Lloyd

Abstract: We propose a generalization of the Wasserstein distance of order 1 to the quantum states of $n$ qudits. The proposal recovers the Hamming distance for the vectors of the canonical basis, and more generally the classical Wasserstein distance for quantum states diagonal in the canonical basis. The proposed distance is invariant with respect to permutations of the qudits and unitary operations acting… ▽ More We propose a generalization of the Wasserstein distance of order 1 to the quantum states of $n$ qudits. The proposal recovers the Hamming distance for the vectors of the canonical basis, and more generally the classical Wasserstein distance for quantum states diagonal in the canonical basis. The proposed distance is invariant with respect to permutations of the qudits and unitary operations acting on one qudit and is additive with respect to the tensor product. Our main result is a continuity bound for the von Neumann entropy with respect to the proposed distance, which significantly strengthens the best continuity bound with respect to the trace distance. We also propose a generalization of the Lipschitz constant to quantum observables. The notion of quantum Lipschitz constant allows us to compute the proposed distance with a semidefinite program. We prove a quantum version of Marton's transportation inequality and a quantum Gaussian concentration inequality for the spectrum of quantum Lipschitz observables. Moreover, we derive bounds on the contraction coefficients of shallow quantum circuits and of the tensor product of one-qudit quantum channels with respect to the proposed distance. We discuss other possible applications in quantum machine learning, quantum Shannon theory, and quantum many-body systems. △ Less

Submitted 13 January, 2022; v1 submitted 9 September, 2020; originally announced September 2020.

Journal ref: IEEE Transactions on Information Theory 67(10), 6627-6643 (2021)

Automatic Recommendation of Strategies for Minimizing Discomfort in Virtual Environments

Authors: Thiago Porcino, Esteban Clua, Daniela Trevisan, Érick Rodrigues, Alexandre Silva

Abstract: Virtual reality (VR) is an imminent trend in games, education, entertainment, military, and health applications, as the use of head-mounted displays is becoming accessible to the mass market. Virtual reality provides immersive experiences but still does not offer an entirely perfect situation, mainly due to Cybersickness (CS) issues. In this work, we first present a detailed review about possible… ▽ More Virtual reality (VR) is an imminent trend in games, education, entertainment, military, and health applications, as the use of head-mounted displays is becoming accessible to the mass market. Virtual reality provides immersive experiences but still does not offer an entirely perfect situation, mainly due to Cybersickness (CS) issues. In this work, we first present a detailed review about possible causes of CS. Following, we propose a novel CS prediction solution. Our system is able to suggest if the user may be entering in the next moments of the application into an illness situation. We use Random Forest classifiers, based on a dataset we have produced. The CSPQ (Cybersickness Profile Questionnaire) is also proposed, which is used to identify the player's susceptibility to CS and the dataset construction. In addition, we designed two immersive environments for empirical studies where participants are asked to complete the questionnaire and describe (orally) the degree of discomfort during their gaming experience. Our data was achieved through 84 individuals on different days, using VR devices. Our proposal also allows us to identify which are the most frequent attributes (causes) in the observed discomfort situations. △ Less

Submitted 27 June, 2020; originally announced June 2020.

Comments: Accepted at the IEEE 8th International Conference on Serious Games and Applications for Health (SeGAH) - SeGAH 2020

Gaussian optimizers for entropic inequalities in quantum information

Authors: Giacomo De Palma, Dario Trevisan, Vittorio Giovannetti, Luigi Ambrosio

Abstract: We survey the state of the art for the proof of the quantum Gaussian optimizer conjectures of quantum information theory. These fundamental conjectures state that quantum Gaussian input states are the solution to several optimization problems involving quantum Gaussian channels. These problems are the quantum counterpart of three fundamental results of functional analysis and probability: the Entr… ▽ More We survey the state of the art for the proof of the quantum Gaussian optimizer conjectures of quantum information theory. These fundamental conjectures state that quantum Gaussian input states are the solution to several optimization problems involving quantum Gaussian channels. These problems are the quantum counterpart of three fundamental results of functional analysis and probability: the Entropy Power Inequality, the sharp Young's inequality for convolutions, and the theorem "Gaussian kernels have only Gaussian maximizers." Quantum Gaussian channels play a key role in quantum communication theory: they are the quantum counterpart of Gaussian integral kernels and provide the mathematical model for the propagation of electromagnetic waves in the quantum regime. The quantum Gaussian optimizer conjectures are needed to determine the maximum communication rates over optical fibers and free space. The restriction of the quantum-limited Gaussian attenuator to input states diagonal in the Fock basis coincides with the thinning, which is the analog of the rescaling for positive integer random variables. Quantum Gaussian channels provide then a bridge between functional analysis and discrete probability. △ Less

Submitted 21 August, 2018; v1 submitted 6 March, 2018; originally announced March 2018.

Journal ref: Journal of Mathematical Physics 59(8), 081101 (2018)

The conditional Entropy Power Inequality for bosonic quantum systems

Authors: Giacomo De Palma, Dario Trevisan

Abstract: We prove the conditional Entropy Power Inequality for Gaussian quantum systems. This fundamental inequality determines the minimum quantum conditional von Neumann entropy of the output of the beam-splitter or of the squeezing among all the input states where the two inputs are conditionally independent given the memory and have given quantum conditional entropies. We also prove that, for any coupl… ▽ More We prove the conditional Entropy Power Inequality for Gaussian quantum systems. This fundamental inequality determines the minimum quantum conditional von Neumann entropy of the output of the beam-splitter or of the squeezing among all the input states where the two inputs are conditionally independent given the memory and have given quantum conditional entropies. We also prove that, for any couple of values of the quantum conditional entropies of the two inputs, the minimum of the quantum conditional entropy of the output given by the conditional Entropy Power Inequality is asymptotically achieved by a suitable sequence of quantum Gaussian input states. Our proof of the conditional Entropy Power Inequality is based on a new Stam inequality for the quantum conditional Fisher information and on the determination of the universal asymptotic behaviour of the quantum conditional entropy under the heat semigroup evolution. The beam-splitter and the squeezing are the central elements of quantum optics, and can model the attenuation, the amplification and the noise of electromagnetic signals. This conditional Entropy Power Inequality will have a strong impact in quantum information and quantum cryptography. Among its many possible applications there is the proof of a new uncertainty relation for the conditional Wehrl entropy. △ Less

Submitted 4 January, 2018; v1 submitted 1 June, 2017; originally announced June 2017.

Journal ref: Communications in Mathematical Physics 360(2), 639-662 (2018)

Minimizing cyber sickness in head mounted display systems: design guidelines and applications

Authors: Thiago M. Porcino, Esteban W. Clua, Cristina N. Vasconcelos, Daniela Trevisan, Luis Valente

Abstract: We are experiencing an upcoming trend of using head mounted display systems in games and serious games, which is likely to become an established practice in the near future. While these systems provide highly immersive experiences, many users have been reporting discomfort symptoms, such as nausea, sickness, and headaches, among others. When using VR for health applications, this is more critical,… ▽ More We are experiencing an upcoming trend of using head mounted display systems in games and serious games, which is likely to become an established practice in the near future. While these systems provide highly immersive experiences, many users have been reporting discomfort symptoms, such as nausea, sickness, and headaches, among others. When using VR for health applications, this is more critical, since the discomfort may interfere a lot in treatments. In this work we discuss possible causes of these issues, and present possible solutions as design guidelines that may mitigate them. In this context, we go deeper within a dynamic focus solution to reduce discomfort in immersive virtual environments, when using first-person navigation. This solution applies an heuristic model of visual attention that works in real time. This work also discusses a case study (as a first-person spatial shooter demo) that applies this solution and the proposed design guidelines. △ Less

Submitted 18 November, 2016; originally announced November 2016.

Comments: 11 pages, 3 figures, 3 tables

Gaussian states minimize the output entropy of one-mode quantum Gaussian channels

Authors: Giacomo De Palma, Dario Trevisan, Vittorio Giovannetti

Abstract: We prove the longstanding conjecture stating that Gaussian thermal input states minimize the output von Neumann entropy of one-mode phase-covariant quantum Gaussian channels among all the input states with a given entropy. Phase-covariant quantum Gaussian channels model the attenuation and the noise that affect any electromagnetic signal in the quantum regime. Our result is crucial to prove the co… ▽ More We prove the longstanding conjecture stating that Gaussian thermal input states minimize the output von Neumann entropy of one-mode phase-covariant quantum Gaussian channels among all the input states with a given entropy. Phase-covariant quantum Gaussian channels model the attenuation and the noise that affect any electromagnetic signal in the quantum regime. Our result is crucial to prove the converse theorems for both the triple trade-off region and the capacity region for broadcast communication of the Gaussian quantum-limited amplifier. Our result extends to the quantum regime the Entropy Power Inequality that plays a key role in classical information theory. Our proof exploits a completely new technique based on the recent determination of the p->q norms of the quantum-limited amplifier [De Palma et al., arXiv:1610.09967]. This technique can be applied to any quantum channel. △ Less

Submitted 22 April, 2017; v1 submitted 31 October, 2016; originally announced October 2016.

Journal ref: Phys. Rev. Lett. 118, 160503 (2017)

Gaussian States Minimize the Output Entropy of the One-Mode Quantum Attenuator

Authors: Giacomo De Palma, Dario Trevisan, Vittorio Giovannetti

Abstract: We prove that Gaussian thermal input states minimize the output von Neumann entropy of the one-mode Gaussian quantum-limited attenuator for fixed input entropy. The Gaussian quantum-limited attenuator models the attenuation of an electromagnetic signal in the quantum regime. The Shannon entropy of an attenuated real-valued classical signal is a simple function of the entropy of the original signal… ▽ More We prove that Gaussian thermal input states minimize the output von Neumann entropy of the one-mode Gaussian quantum-limited attenuator for fixed input entropy. The Gaussian quantum-limited attenuator models the attenuation of an electromagnetic signal in the quantum regime. The Shannon entropy of an attenuated real-valued classical signal is a simple function of the entropy of the original signal. A striking consequence of energy quantization is that the output von Neumann entropy of the quantum-limited attenuator is no more a function of the input entropy alone. The proof starts from the majorization result of De Palma et al., IEEE Trans. Inf. Theory 62, 2895 (2016), and is based on a new isoperimetric inequality. Our result implies that geometric input probability distributions minimize the output Shannon entropy of the thinning for fixed input entropy. Moreover, our result opens the way to the multimode generalization, that permits to determine both the triple trade-off region of the Gaussian quantum-limited attenuator and the classical capacity region of the Gaussian degraded quantum broadcast channel. △ Less

Submitted 21 December, 2016; v1 submitted 2 May, 2016; originally announced May 2016.

Journal ref: IEEE Transactions on Information Theory 63(1), 728-737 (2017)