Searcharxiv⌕ Search

arXiv subjects

Jiaxuan Wang

Publications and source records attributed to Jiaxuan Wang.

34 records · Page 2Linked to original sources
arXiv2023

Quantum advantage of time-reversed ancilla-based metrology of absorption parameters

Quantum estimation of parameters defining open-system dynamics may be enhanced by using ancillas that are entangled with the probe but are not submitted to the dynamics. Here we consider the important problem of estimation of transmission of light by a sample, with losses due to absorption and scattering. We show, through the determination of the quantum Fisher information, that the ancilla strategy leads to the best possible precision in single-mode estimation, the one obtained for a Fock state input, through joint photon-counting of probe and ancilla, which are modes of a bimodal squeezed state produced by an optical parametric amplifier. This proposal overcomes the challenge of producing and detecting high photon-number Fock states, and it is quite robust against additional noise: we show that it is immune to phase noise and the precision does not change if the incoming state gets disentangled. Furthermore, the quantum gain is still present under moderate photon losses of the input beams. We also discuss an alternative to joint photon counting, which is readily implementable with present technology, and approaches the quantum Fisher information result for weak absorption, even with moderate photons losses of the input beams before the sample is probed: a time-reversal procedure, placing the sample between two optical parametric amplifiers, with the second undoing the squeezing produced by the first one. The precision of estimation of the loss parameter is obtained from the average outgoing total photon number and its variance. In both procedures, the state of the probe and the detection procedure are independent of the value of the parameter.

quant-ph↗
arXiv2023

Two-photon Hong-Ou-Mandel interference and quantum entanglement between the frequency-converted idler photon and the signal photon

Quantum frequency up-conversion is a cutting-edge technique that leverages the interaction between photons and quantum systems to shift the frequency of single photons from a lower frequency to a higher frequency. If the photon before up-conversion was one of the entangled pair, then it is important to understand how much entanglement is preserved after up-conversion. In this study, we present a theoretical analysis of the transformation of the time-dependent second-order quantum correlations in photon pairs and find the preservation of such correlations under fairly general conditions. We also analyze the two-photon Hong-Ou-Mandel interference between the frequency-converted idler photon and the signal photon. The visibility of the two-photon interference is sensitive to the magnitude of the frequency conversion, and it improves when the frequency separation between two photons goes down.

quant-ph↗
arXiv2023

SoK: Training Machine Learning Models over Multiple Sources with Privacy Preservation

Nowadays, gathering high-quality training data from multiple data sources with privacy preservation is a crucial challenge to training high-performance machine learning models. The potential solutions could break the barriers among isolated data corpus, and consequently enlarge the range of data available for processing. To this end, both academic researchers and industrial vendors are recently strongly motivated to propose two main-stream folders of solutions mainly based on software constructions: 1) Secure Multi-party Learning (MPL for short); and 2) Federated Learning (FL for short). The above two technical folders have their advantages and limitations when we evaluate them according to the following five criteria: security, efficiency, data distribution, the accuracy of trained models, and application scenarios. Motivated to demonstrate the research progress and discuss the insights on the future directions, we thoroughly investigate these protocols and frameworks of both MPL and FL. At first, we define the problem of Training machine learning Models over Multiple data sources with Privacy Preservation (TMMPP for short). Then, we compare the recent studies of TMMPP from the aspects of the technical routes, the number of parties supported, data partitioning, threat model, and machine learning models supported, to show their advantages and limitations. Next, we investigate and evaluate five popular FL platforms. Finally, we discuss the potential directions to resolve the problem of TMMPP in the future.

cs.CR↗
arXiv2023

Structure PLP-SLAM: Efficient Sparse Mapping and Localization using Point, Line and Plane for Monocular, RGB-D and Stereo Cameras

This paper presents a visual SLAM system that uses both points and lines for robust camera localization, and simultaneously performs a piece-wise planar reconstruction (PPR) of the environment to provide a structural map in real-time. One of the biggest challenges in parallel tracking and mapping with a monocular camera is to keep the scale consistent when reconstructing the geometric primitives. This further introduces difficulties in graph optimization of the bundle adjustment (BA) step. We solve these problems by proposing several run-time optimizations on the reconstructed lines and planes. Our system is able to run with depth and stereo sensors in addition to the monocular setting. Our proposed SLAM tightly incorporates the semantic and geometric features to boost both frontend pose tracking and backend map optimization. We evaluate our system exhaustively on various datasets, and show that we outperform state-of-the-art methods in terms of trajectory precision. The code of PLP-SLAM has been made available in open-source for the research community (https://github.com/PeterFWS/Structure-PLP-SLAM).

cs.CV↗
arXiv2022

pMPL: A Robust Multi-Party Learning Framework with a Privileged Party

In order to perform machine learning among multiple parties while protecting the privacy of raw data, privacy-preserving machine learning based on secure multi-party computation (MPL for short) has been a hot spot in recent. The configuration of MPL usually follows the peer-to-peer architecture, where each party has the same chance to reveal the output result. However, typical business scenarios often follow a hierarchical architecture where a powerful, usually privileged party, leads the tasks of machine learning. Only the privileged party can reveal the final model even if other assistant parties collude with each other. It is even required to avoid the abort of machine learning to ensure the scheduled deadlines and/or save used computing resources when part of assistant parties drop out. Motivated by the above scenarios, we propose pMPL, a robust MPL framework with a privileged part}. pMPL supports three-party training in the semi-honest setting. By setting alternate shares for the privileged party, pMPL is robust to tolerate one of the rest two parties dropping out during the training. With the above settings, we design a series of efficient protocols based on vector space secret sharing for pMPL to bridge the gap between vector space secret sharing and machine learning. Finally, the experimental results show that the performance of pMPL is promising when we compare it with the state-of-the-art MPL frameworks. Especially, in the LAN setting, pMPL is around $16\times$ and $5\times$ faster than TF-encrypted (with ABY3 as the back-end framework) for the linear regression, and logistic regression, respectively. Besides, the accuracy of trained models of linear regression, logistic regression, and BP neural networks can reach around 97%, 99%, and 96% on MNIST dataset respectively.

cs.CR↗
arXiv2022

Probing Ultra-Fast Dephasing via Entangled Photon Pairs

We demonstrate how the Hong-Ou-Mandel (HOM) interference with polarization-entangled photons can be used to probe ultrafast dephasing. We can infer the optical properties like the real and imaginary parts of the complex susceptibility of the medium from changes in the position and the shape of the HOM dip. From the shift of the HOM dip, we are able to measure 22 fs dephasing time using a continuous-wave (CW) laser even with optical loss > 97%, while the HOM dip visibility is maintained at 92.3~\% (which can be as high as 96.7%). The experimental observations, which are explained in terms of a rigorous theoretical model, demonstrate the utility of HOM interference in probing ultrafast dephasing.

quant-ph↗
arXiv2021

Quantum Fisher Information Bounds on Precision Limits of Circular Dichroism

Circular dichroism (CD) is a widely used technique for investigating optically chiral molecules, especially for biomolecules. It is thus of great importance that these parameters be estimated precisely so that the molecules with desired functionalities can be designed. In order to surpass the limits of classical measurements, we need to probe the system with quantum light. We develop quantum Fisher information matrix (QFIM) for precision estimates of the circular dichroism and the optical rotary dispersion for a variety of input quantum states of light. The Cramer-Rao bounds, for all four chirality parameters are obtained, from QFIM for (a) single photon input states with a specific linear polarization and for (b) NOON states having two photons with both either left polarized or right polarized. The QFIM bounds, using quantum light, are compared with bounds obtained for classical light beams i.e., beams in coherent states. Quite generally, both the single photon state and the NOON state exhibit superior precision in the estimation of absorption and phase shift in relation to a coherent source of comparable intensity, especially in the weak absorption regime. In particular, the NOON state naturally offers the best precision among the three. We compare QFIM bounds with the error sensitivity bounds, as the latter are relatively easier to measure whereas the QFIM bounds require full state tomography. We also outline an empirical scheme for estimating the measurement sensitivities by projective measurements with single-photon detectors.

quant-ph↗
arXiv2021

Shapley Flow: A Graph-based Approach to Interpreting Model Predictions

Many existing approaches for estimating feature importance are problematic because they ignore or hide dependencies among features. A causal graph, which encodes the relationships among input variables, can aid in assigning feature importance. However, current approaches that assign credit to nodes in the causal graph fail to explain the entire graph. In light of these limitations, we propose Shapley Flow, a novel approach to interpreting machine learning models. It considers the entire causal graph, and assigns credit to \textit{edges} instead of treating nodes as the fundamental unit of credit assignment. Shapley Flow is the unique solution to a generalization of the Shapley value axioms to directed acyclic graphs. We demonstrate the benefit of using Shapley Flow to reason about the impact of a model's input on its output. In addition to maintaining insights from existing approaches, Shapley Flow extends the flat, set-based, view prevalent in game theory based explanation methods to a deeper, \textit{graph-based}, view. This graph-based view enables users to understand the flow of importance through a system, and reason about potential interventions.

cs.LG↗
arXiv2020

Quantum sensing of open systems: Estimation of damping constants and temperature

We determine quantum precision limits for estimation of damping constants and temperature of lossy bosonic channels. A direct application would be the use of light for estimation of the absorption and the temperature of a transparent slab. Analytic lower bounds are obtained for the uncertainty in the estimation, through a purification procedure that replaces the master equation description by a unitary evolution involving the system and ad hoc environments. For zero temperature, Fock states are shown to lead to the minimal uncertainty in the estimation of damping, with boson-counting being the best measurement procedure. In both damping and temperature estimates, sequential pre-thermalization measurements, through a stream of single bosons, may lead to huge gain in precision.

quant-ph↗
arXiv2020

AdaSGD: Bridging the gap between SGD and Adam

In the context of stochastic gradient descent(SGD) and adaptive moment estimation (Adam),researchers have recently proposed optimization techniques that transition from Adam to SGD with the goal of improving both convergence and generalization performance. However, precisely how each approach trades off early progress and generalization is not well understood; thus, it is unclear when or even if, one should transition from one approach to the other. In this work, by first studying the convex setting, we identify potential contributors to observed differences in performance between SGD and Adam. In particular,we provide theoretical insights for when and why Adam outperforms SGD and vice versa. We ad-dress the performance gap by adapting a single global learning rate for SGD, which we refer to as AdaSGD. We justify this proposed approach with empirical analyses in non-convex settings. On several datasets that span three different domains,we demonstrate how AdaSGD combines the benefits of both SGD and Adam, eliminating the need for approaches that transition from Adam to SGD.

cs.LG↗
arXiv2020

Relaxed Parameter Sharing: Effectively Modeling Time-Varying Relationships in Clinical Time-Series

Recurrent neural networks (RNNs) are commonly applied to clinical time-series data with the goal of learning patient risk stratification models. Their effectiveness is due, in part, to their use of parameter sharing over time (i.e., cells are repeated hence the name recurrent). We hypothesize, however, that this trait also contributes to the increased difficulty such models have with learning relationships that change over time. Conditional shift, i.e., changes in the relationship between the input X and the output y, arises when risk factors associated with the event of interest change over the course of a patient admission. While in theory, RNNs and gated RNNs (e.g., LSTMs) in particular should be capable of learning time-varying relationships, when training data are limited, such models often fail to accurately capture these dynamics. We illustrate the advantages and disadvantages of complete parameter sharing (RNNs) by comparing an LSTM with shared parameters to a sequential architecture with time-varying parameters on prediction tasks involving three clinically-relevant outcomes: acute respiratory failure (ARF), shock, and in-hospital mortality. In experiments using synthetic data, we demonstrate how parameter sharing in LSTMs leads to worse performance in the presence of conditional shift. To improve upon the dichotomy between complete parameter sharing and no parameter sharing, we propose a novel RNN formulation based on a mixture model in which we relax parameter sharing over time. The proposed method outperforms standard LSTMs and other state-of-the-art baselines across all tasks. In settings with limited data, relaxed parameter sharing can lead to improved patient risk stratification performance.

cs.LG↗
arXiv2018

Learning to Exploit Invariances in Clinical Time-Series Data using Sequence Transformer Networks

Recently, researchers have started applying convolutional neural networks (CNNs) with one-dimensional convolutions to clinical tasks involving time-series data. This is due, in part, to their computational efficiency, relative to recurrent neural networks and their ability to efficiently exploit certain temporal invariances, (e.g., phase invariance). However, it is well-established that clinical data may exhibit many other types of invariances (e.g., scaling). While preprocessing techniques, (e.g., dynamic time warping) may successfully transform and align inputs, their use often requires one to identify the types of invariances in advance. In contrast, we propose the use of Sequence Transformer Networks, an end-to-end trainable architecture that learns to identify and account for invariances in clinical time-series data. Applied to the task of predicting in-hospital mortality, our proposed approach achieves an improvement in the area under the receiver operating characteristic curve (AUROC) relative to a baseline CNN (AUROC=0.851 vs. AUROC=0.838). Our results suggest that a variety of valuable invariances can be learned directly from the data.

cs.LG↗
arXiv2018

Learning Credible Models

In many settings, it is important that a model be capable of providing reasons for its predictions (i.e., the model must be interpretable). However, the model's reasoning may not conform with well-established knowledge. In such cases, while interpretable, the model lacks \textit{credibility}. In this work, we formally define credibility in the linear setting and focus on techniques for learning models that are both accurate and credible. In particular, we propose a regularization penalty, expert yielded estimates (EYE), that incorporates expert knowledge about well-known relationships among covariates and the outcome of interest. We give both theoretical and empirical results comparing our proposed method to several other regularization techniques. Across a range of settings, experiments on both synthetic and real data show that models learned using the EYE penalty are significantly more credible than those learned using other penalties. Applied to a large-scale patient risk stratification task, our proposed technique results in a model whose top features overlap significantly with known clinical risk factors, while still achieving good predictive performance.

cs.LG↗
arXiv2018

The Advantage of Doubling: A Deep Reinforcement Learning Approach to Studying the Double Team in the NBA

During the 2017 NBA playoffs, Celtics coach Brad Stevens was faced with a difficult decision when defending against the Cavaliers: "Do you double and risk giving up easy shots, or stay at home and do the best you can?" It's a tough call, but finding a good defensive strategy that effectively incorporates doubling can make all the difference in the NBA. In this paper, we analyze double teaming in the NBA, quantifying the trade-off between risk and reward. Using player trajectory data pertaining to over 643,000 possessions, we identified when the ball handler was double teamed. Given these data and the corresponding outcome (i.e., was the defense successful), we used deep reinforcement learning to estimate the quality of the defensive actions. We present qualitative and quantitative results summarizing our learned defensive strategy for defending. We show that our policy value estimates are predictive of points per possession and win percentage. Overall, the proposed framework represents a step toward a more comprehensive understanding of defensive strategies in the NBA.

cs.LG↗
arXiv2018

Magnetic dipole-dipole interaction induced by the electromagnetic field

We give a derivation for the indirect interaction between two magnetic dipoles induced by the quantized electromagnetic field. It turns out that the interaction between permanent dipoles directly returns to the classical form; the interaction between transition dipoles does not directly return to the classical result, yet returns in the short-distance limit. In a finite volume, the field modes are highly discrete, and both the permanent and transition dipole-dipole interactions are changed. For transition dipoles, the changing mechanism is similar with the Purcell effect, since only a few number of nearly resonant modes take effect in the interaction mediation; for permanent dipoles, the correction comes from the boundary effect: if the dipoles are placed close to the boundary, the influence is strong, otherwise, their interaction does not change too much from the free space case.

quant-ph↗