arXiv subjects

Si Luo

Publications and source records attributed to Si Luo.

7 recordsLinked to original sources

arXiv2024

A purified input-output pseudomode model for structured open quantum systems

A full understanding of open quantum systems requires the characterization of both system and environmental properties. However, the complexity of the environmental statistics in the presence of strong system-bath hybridization and long memory effects usually prevents effective non-perturbative methods from going beyond the analysis of the reduced system dynamics. Here we present a model consisting of purified auxiliary bosonic modes to describe, alongside properties of the system, the dynamics of environmental observables for bosonic baths prepared in non-Gaussian initial states. We numerically exemplify this method by simulating non-Markovian multi-photon transfer processes on a coupled cavity waveguide system in the large time delay regime.

quant-ph↗

arXiv2024

Principle of least action for quasi-adiabatic state transfers with dissipation

We discuss a general formalism to optimize quasi-adiabatic state-transfer protocols, where high fidelity is achieved by maintaining the system in a dark subspace protected from the dominant dissipative channels. We cast the residual fidelity loss, induced by a combination of dissipation and non-adiabatic transitions, in the form of a classical action where the time-dependent control parameters act as coordinates. This allows us to apply the least action principle, yielding the fidelity upper-bound and the corresponding optimal transfer time. As an application, we analyze a system of two qubits subject to weak relaxation and dephasing, interacting through a strongly dissipative quantum bus. In this case, our formalism, we obtain a full characterization of the optimal state-transfer fidelity.

quant-ph↗

arXiv2023

Modeling the unphysical pseudomode model with physical ensembles: simulation, mitigation, and restructuring of non-Markovian quantum noise

The influence of a Gaussian environment on a quantum system can be described by effectively replacing the continuum with a discrete set of ancillary quantum and classical degrees of freedom. This defines a pseudomode model which can be used to classically simulate the reduced system dynamics. Here, we consider an alternative point of view and analyze the potential benefits of an analog or digital quantum simulation of the pseudomode model itself. Superficially, such a direct experimental implementation is, in general, impossible due to the unphysical properties of the effective degrees of freedom involved. However, we show that the effects of the unphysical pseudomode model can still be reproduced using measurement results over an ensemble of physical systems involving ancillary harmonic modes and an optional stochastic driving field. This is done by introducing an extrapolation technique whose efficiency is limited by stability against imprecision in the measurement data. We examine how such a simulation would allow us to (i) perform accurate quantum simulation of the effects of complex non-perturbative and non-Markovian environments in regimes that are challenging for classical simulation, (ii) conversely, mitigate potential unwanted non-Markovian noise present in quantum devices, and (iii) restructure some of some of the properties of a given physical bath, such as its temperature.

quant-ph↗

arXiv2023

A quantum-classical decomposition of Gaussian quantum environments: a stochastic pseudomode model

We show that the effect of a Gaussian Bosonic environment linearly coupled to a quantum system can be simulated by a stochastic Lindblad master equation characterized by a set of ancillary Bosonic modes initially at zero temperature and classical stochastic fields. We test the method for Ohmic environments with exponential and polynomial cut-offs against, respectively, the Hierarchical Equations of Motion and the deterministic pseudomode model with respect to which the number of ancillary quantum degrees of freedom is reduced. For a subset of rational spectral densities, all parameters are explicitly specified without the need of any fitting procedure, thereby simplifying the modeling strategy. Interestingly, the classical fields in this decomposition must sometimes be imaginary-valued, which can have counter-intuitive effects on the system properties which we demonstrate by showing that they can decrease the entropy of the system, in contrast to real-valued fields.

quant-ph↗

arXiv2021

Geometric Manipulation of a Decoherence-Free Subspace in Atomic Ensembles

We consider an ensemble of atoms with $\Lambda$-type level structure trapped in a single-mode cavity, and propose a geometric scheme of coherent manipulation of quantum states on the subspace of zero-energy states within the quantum Zeno subspace of the system. We find that the particular subspace inherits the decoherence-free nature of the quantum Zeno subspace and features a symmetry-protected degeneracy, fulfilling all the conditions for a universal scheme of arbitrary unitary operations on it.

quant-ph↗

arXiv2021

Data Augmentation for Object Detection via Differentiable Neural Rendering

It is challenging to train a robust object detector under the supervised learning setting when the annotated data are scarce. Thus, previous approaches tackling this problem are in two categories: semi-supervised learning models that interpolate labeled data from unlabeled data, and self-supervised learning approaches that exploit signals within unlabeled data via pretext tasks. To seamlessly integrate and enhance existing supervised object detection methods, in this work, we focus on addressing the data scarcity problem from a fundamental viewpoint without changing the supervised learning paradigm. We propose a new offline data augmentation method for object detection, which semantically interpolates the training data with novel views. Specifically, our new system generates controllable views of training images based on differentiable neural rendering, together with corresponding bounding box annotations which involve no human intervention. Firstly, we extract and project pixel-aligned image features into point clouds while estimating depth maps. We then re-project them with a target camera pose and render a novel-view 2d image. Objects in the form of keypoints are marked in point clouds to recover annotations in new views. Our new method is fully compatible with online data augmentation methods, such as affine transform, image mixup, etc. Extensive experiments show that our method, as a cost-free tool to enrich images and labels, can significantly boost the performance of object detection systems with scarce training data. Code is available at \url{https://github.com/Guanghan/DANR}.

cs.CV↗

arXiv2020

Efficient Pig Counting in Crowds with Keypoints Tracking and Spatial-aware Temporal Response Filtering

Pig counting is a crucial task for large-scale pig farming, which is usually completed by human visually. But this process is very time-consuming and error-prone. Few studies in literature developed automated pig counting method. Existing methods only focused on pig counting using single image, and its accuracy is challenged by several factors, including pig movements, occlusion and overlapping. Especially, the field of view of a single image is very limited, and could not meet the requirements of pig counting for large pig grouping houses. To that end, we presented a real-time automated pig counting system in crowds using only one monocular fisheye camera with an inspection robot. Our system showed that it produces accurate results surpassing human. Our pipeline began with a novel bottom-up pig detection algorithm to avoid false negatives due to overlapping, occlusion and deformation of pigs. A deep convolution neural network (CNN) is designed to detect keypoints of pig body part and associate the keypoints to identify individual pigs. After that, an efficient on-line tracking method is used to associate pigs across video frames. Finally, a novel spatial-aware temporal response filtering (STRF) method is proposed to predict the counts of pigs, which is effective to suppress false positives caused by pig or camera movements or tracking failures. The whole pipeline has been deployed in an edge computing device, and demonstrated the effectiveness.

cs.CV↗