arXiv subjects

Xinyi Jiang

Publications and source records attributed to Xinyi Jiang.

25 records · Page 2Linked to original sources

arXiv2021

Virufy: Global Applicability of Crowdsourced and Clinical Datasets for AI Detection of COVID-19 from Cough

Rapid and affordable methods of testing for COVID-19 infections are essential to reduce infection rates and prevent medical facilities from becoming overwhelmed. Current approaches of detecting COVID-19 require in-person testing with expensive kits that are not always easily accessible. This study demonstrates that crowdsourced cough audio samples recorded and acquired on smartphones from around the world can be used to develop an AI-based method that accurately predicts COVID-19 infection with an ROC-AUC of 77.1% (75.2%-78.3%). Furthermore, we show that our method is able to generalize to crowdsourced audio samples from Latin America and clinical samples from South Asia, without further training using the specific samples from those regions. As more crowdsourced data is collected, further development can be implemented using various respiratory audio samples to create a cough analysis-based machine learning (ML) solution for COVID-19 detection that can likely generalize globally to all demographic groups in both clinical and non-clinical settings.

cs.SD↗

arXiv2020

Efficient Receive Beamformers for Secure Spatial Modulation against a Malicious Full-duplex Attacker with Eavesdropping Ability

In this paper, we consider a new secure spatial modulation scenario with a full-duplex (FD) malicious attacker Mallory owning eavesdropping capacity, where Mallory works on FD model and transmits a malicious jamming such as artificial noise (AN) to interfere with Bob. To suppress the malicious jamming on Bob from Mallory, a conventional maximum receive power (Max-RP) at Bob is presented firstly. Subsequently, to exploit the colored property of noise plus interference at Bob, a whitening-filter-based Max-RP (Max-WFRP) is proposed with an obvious performance enhancement over Max-RP. To completely remove the malicious jamming from Mallory, a Max-RP with a constraint of forcing the malicious jamming from Mallory to zero at Bob is proposed. To further improve secrecy rate (SR) by removing the ZF contraint (ZFC), the maximum signal-to-jamming-plus-noise ratio (Max-SJNR) is proposed. Our proposed methods have closed-form expressions. From simulation results, the four receive beamforming methods have an increasing order in performance: Max-RP, Max-RP with ZFC and Max-SJNR$\approx$Max-WFRP. Additionally, the latter two harvest a substantial performance gains over Max-RP and Max-RP with ZFC in the low and medium signal-to-noise ratio regions.

cs.IT↗

arXiv2020

Incremental Sense Weight Training for the Interpretation of Contextualized Word Embeddings

We present a novel online algorithm that learns the essence of each dimension in word embeddings by minimizing the within-group distance of contextualized embedding groups. Three state-of-the-art neural-based language models are used, Flair, ELMo, and BERT, to generate contextualized word embeddings such that different embeddings are generated for the same word type, which are grouped by their senses manually annotated in the SemCor dataset. We hypothesize that not all dimensions are equally important for downstream tasks so that our algorithm can detect unessential dimensions and discard them without hurting the performance. To verify this hypothesis, we first mask dimensions determined unessential by our algorithm, apply the masked word embeddings to a word sense disambiguation task (WSD), and compare its performance against the one achieved by the original embeddings. Several KNN approaches are experimented to establish strong baselines for WSD. Our results show that the masked word embeddings do not hurt the performance and can improve it by 3%. Our work can be used to conduct future research on the interpretability of contextualized embeddings.

cs.CL↗

arXiv2020

Precoding and Transmit Antenna Subarray Selection for Secure Hybrid Spatial Modulation

Spatial modulation (SM) is a particularly important form of multiple-input-multiple-output (MIMO). Unlike traditional MIMO, it uses both modulation symbols and antenna indices to carry information. In this paper, to avoid the high cost and circuit complexity of fully-digital SM, we mainly consider the hybrid SM system with a hybrid precoding transmitter architecture, combining a digital precoder and an analog precoder. Here, the partially-connected structure is adopted with each radio frequency chain (RF) being connected to a transmit antenna subarray (TAS). In such a system, we made an investigation of secure hybrid precoding and transmit antenna subarray selection (TASS) methods. Two hybrid precoding methods, called maximizing the approximate secrecy rate (SR) via gradient ascent (Max-ASR-GA) and maximizing the approximate SR via alternating direction method of multipliers (Max-ASR-ADMM), are proposed to improve the SR performance. As for TASS, a high-performance method of maximizing the approximate SR (Max-ASR) TASS method is first presented. To reduce its high complexity, two low-complexity TASS methods, namely maximizing the eigenvalue (Max-EV) and maximizing the product of signal-to-interference-plus-noise ratio and artificial noise-to-signal-plus-noise ratio (Max-P-SINR-ANSNR), are proposed. Simulation results will demonstrate that the proposed Max-ASR-GA and Max-ASR-ADMM hybrid precoders harvest substantial SR performance gains over existing method. For TASS, the proposed three methods Max-ASR, Max-EV, and Max-P-SINR-ANSNR perform better than existing leakage method. Particularly, the proposed Max-EV and Max-P-SINR-ANSNR is low-complexity at the expense of a little performance loss compared with Max-ASR.

eess.SP↗

arXiv2019

How to make a giant bubble

Using mixtures of soap, water, and long chain polymers, free-floating soap bubbles can be formed with volumes approaching 100 m$^3$. Here we investigate how such thin films are created and maintained over time. We show how the extensional rheology is the most important factor in creating the bubble, and how polydispersity in molecular weight of the solvated polymers leads to better performance at lower concentrations. Additionally, using IR absorption, we measure soap film thickness profiles and film lifetimes. Although the initial thickness mostly depends on the choice of detergent, polymers can dramatically increase film lifetime at high molecular weights and high concentrations, although such high concentrations can inhibit the initial film formation. Thus, the ideal concentration of polymer additives for making giant bubbles requires a robust viscoelastic rheology during extension, and is aided by long film lifetimes during gravitational drainage and evaporation.

physics.flu-dyn↗

arXiv2017

Generalized bipyramids and hyperbolic volumes of tiling links

We present explicit geometric decompositions of the complement of tiling links, which are alternating links whose projection graphs are uniform tilings of the 2-sphere, the Euclidean plane or the hyperbolic plane. This requires generalizing the angle structures program of Casson and Rivin for triangulations with a mixture of finite, ideal, and truncated (i.e. ultra-ideal) vertices. A consequence of this decomposition is that the volumes of spherical tiling links are precisely twice the maximal volumes of the ideal Archimedean solids of the same combinatorial description. In the case of hyperbolic tiling links, we are led to consider links embedded in thickened surfaces S_g x I with genus g at least 2. We generalize the bipyramid construction of Adams to truncated bipyramids and use them to prove that the set of possible volume densities for links in S_g x I, ranging over all g at least 2, is a dense subset of the interval [0, 2v_{oct}], where v_{oct}, approximately 3.66386, is the volume of the regular ideal octahedron.

math.GT↗

arXiv2015

Volume and Determinant Densities of Hyperbolic Rational Links

The volume density of a hyperbolic link is defined as the ratio of hyperbolic volume to crossing number. We study its properties and a closely-related invariant called the determinant density. It is known that the sets of volume densities and determinant densities of links are dense in the interval [0,v_{oct}]. We construct sequences of alternating knots whose volume and determinant densities both converge to any x in [0,v_{oct}]. We also investigate the distributions of volume and determinant densities for hyperbolic rational links, and establish upper bounds and density results for these invariants.

math.GT↗