arXiv subjects

Rui Li

Publications and source records attributed to Rui Li.

At least 379 records · Page 21Linked to original sources

arXiv2023

Expanding Knowledge Graphs with Humans in the Loop

Curated knowledge graphs encode domain expertise and improve the performance of recommendation, segmentation, ad targeting, and other machine learning systems in several domains. As new concepts emerge in a domain, knowledge graphs must be expanded to preserve machine learning performance. Manually expanding knowledge graphs, however, is infeasible at scale. In this work, we propose a method for knowledge graph expansion with humans-in-the-loop. Concretely, given a knowledge graph, our method predicts the "parents" of new concepts to be added to this graph for further verification by human experts. We show that our method is both accurate and provably "human-friendly". Specifically, we prove that our method predicts parents that are "near" concepts' true parents in the knowledge graph, even when the predictions are incorrect. We then show, with a controlled experiment, that satisfying this property increases both the speed and the accuracy of the human-algorithm collaboration. We further evaluate our method on a knowledge graph from Pinterest and show that it outperforms competing methods on both accuracy and human-friendliness. Upon deployment in production at Pinterest, our method reduced the time needed for knowledge graph expansion by ~400% (compared to manual expansion), and contributed to a subsequent increase in ad revenue of 20%.

cs.LG↗

arXiv2023

Resilient conductive membrane synthesized by in-situ polymerisation for wearable non-invasive electronics on moving appendages of cyborg insect

By leveraging their high mobility and small size, insects have been combined with microcontrollers to build up cyborg insects for various practical applications. Unfortunately, all current cyborg insects rely on implanted electrodes to control their movement, which causes irreversible damage to their organs and muscles. Here, we develop a non-invasive method for cyborg insects to address above issues, using a conformal electrode with an in-situ polymerized ion-conducting layer and an electron-conducting layer. The neural and locomotion responses to the electrical inductions verify the efficient communication between insects and controllers by the non-invasive method. The precise "S" line following of the cyborg insect further demonstrates its potential in practical navigation. The conformal non-invasive electrodes keep the intactness of the insects used while controlling their motion. With the antennae, important olfactory organs of insects preserved, the cyborg insect, in the future, may be endowed with abilities to detect the surrounding environment.

cs.RO↗

arXiv2023

LSwinSR: UAV Imagery Super-Resolution based on Linear Swin Transformer

Super-resolution, which aims to reconstruct high-resolution images from low-resolution images, has drawn considerable attention and has been intensively studied in computer vision and remote sensing communities. The super-resolution technology is especially beneficial for Unmanned Aerial Vehicles (UAV), as the amount and resolution of images captured by UAV are highly limited by physical constraints such as flight altitude and load capacity. In the wake of the successful application of deep learning methods in the super-resolution task, in recent years, a series of super-resolution algorithms have been developed. In this paper, for the super-resolution of UAV images, a novel network based on the state-of-the-art Swin Transformer is proposed with better efficiency and competitive accuracy. Meanwhile, as one of the essential applications of the UAV is land cover and land use monitoring, simple image quality assessments such as the Peak-Signal-to-Noise Ratio (PSNR) and the Structural Similarity Index Measure (SSIM) are not enough to comprehensively measure the performance of an algorithm. Therefore, we further investigate the effectiveness of super-resolution methods using the accuracy of semantic segmentation. The code will be available at https://github.com/lironui/LSwinSR.

eess.IV↗

arXiv2023

Structural phase transitions and superconductivity in the Heusler intermetallics $X$Pd$_2$Sn ($X$ = Ti, Zr, Hf)

We report the discovery of structural phase transitions and superconductivity in the full Heusler compounds $X$Pd$_2$Sn ($X$ = Ti, Zr, Hf), by means of electrical transport, magnetic susceptibility, specific heat and x-ray diffraction measurements. TiPd$_2$Sn, ZrPd$_2$Sn and HfPd$_2$Sn undergo structural phase transitions from the room-temperature cubic MnCu$_2$Al-type structure (space group $Fm\bar{3}m$) to a low-temperature tetragonal structure at around 160 K, 110 K and 90 K, respectively, which are likely related charge density wave (CDW) instabilities. Low temperature single crystal x-ray diffraction measurements of ZrPd$_2$Sn demonstrate the emergence of a superstructure with multiple commensurate modulations below $T_s$. ZrPd$_2$Sn and HfPd$_2$Sn have bulk superconductivity (SC) with transition temperatures $T_c$ $\sim$ 1.2 K and 1.3 K, respectively. Density functional theory (DFT) calculations reveal evidence for structural and electronic instabilities which can give rise to CDW formation, suggesting that these $X$Pd$_2$Sn systems are good candidates for examining the interplay between CDW and SC.

cond-mat.supr-con↗

arXiv2023

Semantic-guided Disentangled Representation for Unsupervised Cross-modality Medical Image Segmentation

Disentangled representation is a powerful technique to tackle domain shift problem in medical image analysis in unsupervised domain adaptation setting.However, previous methods only focus on exacting domain-invariant feature and ignore whether exacted feature is meaningful for downstream tasks.We propose a novel framework, called semantic-guided disentangled representation (SGDR), an effective method to exact semantically meaningful feature for segmentation task to improve performance of cross modality medical image segmentation in unsupervised domain adaptation setting. To exact the meaningful domain-invariant features of different modality, we introduce a content discriminator to force the content representation to be embedded to the same space and a feature discriminator to exact the meaningful representation.We also use pixel-level annotations to guide the encoder to learn features that are meaningful for segmentation task.We validated our method on two public datasets and experiment results show that our approach outperforms the state of the art methods on two evaluation metrics by a significant margin.

eess.IV↗

arXiv2023

Learning on Large-scale Text-attributed Graphs via Variational Inference

This paper studies learning on text-attributed graphs (TAGs), where each node is associated with a text description. An ideal solution for such a problem would be integrating both the text and graph structure information with large language models and graph neural networks (GNNs). However, the problem becomes very challenging when graphs are large due to the high computational complexity brought by training large language models and GNNs together. In this paper, we propose an efficient and effective solution to learning on large text-attributed graphs by fusing graph structure and language learning with a variational Expectation-Maximization (EM) framework, called GLEM. Instead of simultaneously training large language models and GNNs on big graphs, GLEM proposes to alternatively update the two modules in the E-step and M-step. Such a procedure allows training the two modules separately while simultaneously allowing the two modules to interact and mutually enhance each other. Extensive experiments on multiple data sets demonstrate the efficiency and effectiveness of the proposed approach.

cs.LG↗

arXiv2023

Electrical manipulation of a hole `spin'-orbit qubit in nanowire quantum dot: the nontrivial magnetic field effects

Strong `spin'-orbit coupled one-dimensional hole gas is achievable in a Ge nanowire in the presence of a strong magnetic field. The strong magnetic field lifts the two-fold degeneracy in the hole subband dispersions, so that the effective low-energy subband dispersion exhibits strong `spin'-orbit coupling. Here, we study the electrical `spin' manipulation in a Ge nanowire quantum dot for both the lowest and second lowest hole subband dispersions. Using a finite square well to model the quantum dot confining potential, we calculate exactly the level splitting of the `spin'-orbit qubit and the Rabi frequency in the electric-dipole `spin' resonance. The `spin'-orbit coupling modulated longitudinal $g$-factor $g_{\rm so}$ is not only non-vanishing but also magnetic field dependent. Moreover, the `spin'-orbit couplings of the lowest and second lowest subband dispersions have opposite magnetic dependences, so that the results for these two subband dispersions are totally different.

cond-mat.mes-hall↗

arXiv2023

Two-band description of the strong `spin'-orbit coupled one-dimensional hole gas in a cylindrical Ge nanowire

The low-energy effective Hamiltonian of the strong `spin'-orbit coupled one-dimensional hole gas in a cylindrical Ge nanowire in the presence of a strong magnetic field is studied both numerically and analytically. Basing on the Luttinger-Kohn Hamiltonian in the spherical approximation, we show this strong `spin'-orbit coupled one-dimensional hole gas can be accurately described by an effective two-band Hamiltonian $H^{\rm ef}=\hbar^{2}k^{2}_{z}/(2m^{*}_{h})+ασ^{x}k_{z}+g^{*}_{h}μ_{B}Bσ^{z}/2$, as long as the magnetic field is purely longitudinal or purely transverse. The explicit magnetic field dependent expressions of the `spin'-orbit coupling $α\equivα(B)$ and the effective $g$-factor $g^{*}_{h}\equiv\,g^{*}_{h}(B)$ are given. When the magnetic field is applied in an arbitrary direction, the two-band Hamiltonian description is still a good approximation.

cond-mat.mes-hall↗

arXiv2023

Producing Usable Taxonomies Cheaply and Rapidly at Pinterest Using Discovered Dynamic $μ$-Topics

Creating a taxonomy of interests is expensive and human-effort intensive: not only do we need to identify nodes and interconnect them, in order to use the taxonomy, we must also connect the nodes to relevant entities such as users, pins, and queries. Connecting to entities is challenging because of ambiguities inherent to language but also because individual interests are dynamic and evolve. Here, we offer an alternative approach that begins with bottom-up discovery of $μ$-topics called pincepts. The discovery process itself connects these $μ$-topics dynamically with relevant queries, pins, and users at high precision, automatically adapting to shifting interests. Pincepts cover all areas of user interest and automatically adjust to the specificity of user interests and are thus suitable for the creation of various kinds of taxonomies. Human experts associate taxonomy nodes with $μ$-topics (on average, 3 $μ$-topics per node), and the $μ$-topics offer a high-level data layer that allows quick definition, immediate inspection, and easy modification. Even more powerfully, $μ$-topics allow easy exploration of nearby semantic space, enabling curators to spot and fill gaps. Curators' domain knowledge is heavily leveraged and we thus don't need untrained mechanical Turks, allowing further cost reduction. These $μ$-topics thus offer a satisfactory "symbolic" stratum over which to define taxonomies. We have successfully applied this technique for very rapidly iterating on and launching the home decor and fashion styles taxonomy for style-based personalization, prominently featured at the top of Pinterest search results, at 94% precision, improving search success rate by 34.8% as well as boosting long clicks and pin saves.

cs.CL↗

arXiv2023

Discovering strongly lensed quasar candidates with catalogue-based methods from DESI Legacy Surveys

The Hubble tension, revealed by a $\sim 5σ$ discrepancy between measurements of the Hubble-Lemaitre constant from early- and local-Universe observations, is one of the most significant problems in modern cosmology. In order to better understand the origin of this mismatch, independent techniques to measure $H_0$, such as strong lensing time delays, are required. Notably, the sample size of such systems is key to minimising statistical uncertainties and cosmic variance, which can be improved by exploring the datasets of large-scale sky surveys like DESI (Dark Energy Spectroscopic Instrument). We identify possible strong lensing time-delay systems within DESI by selecting candidate multiply imaged lensed quasars from a catalogue of 24,440,816 candidate QSOs contained in the 9th data release of the DESI Legacy Imaging Surveys (DESI-LS). Using a friend-of-friends-like algorithm on spatial co-ordinates, our method generates an initial list of compact quasar groups. This list is subsequently filtered using a measure of the similarity of colours of a group's members and the likelihood that they are quasars. A visual inspection finally selects candidate strong lensing systems based on the spatial configuration of the group members. We identify 620 new candidate multiply imaged lensed quasars (101 Grade-A, 214 Grade-B, 305 Grade-C). This number excludes 53 known spectroscopically confirmed systems and existing candidate systems identified in other similar catalogues. When available, these new candidates will be further checked by combining the spectroscopic and photometric data from DESI. The catalogues and images of the candidates in this work are available online (https://github.com/EigenHermit/lensed_qso_cand_catalogue_He-22/).

astro-ph.CO↗

arXiv2022

A Large-scale Friend Suggestion Architecture

Online social as an extension of traditional life plays an important role in our daily lives. Users often seek out new friends that have significant similarities such as interests and habits, motivating us to exploit such online information to suggest friends to users. In this work, we focus on friend suggestion in online game platforms because in-game social quality significantly correlates with player engagement, determining game experience. Unlike a typical recommendation system that depends on item-user interactions, in our setting, user-user interactions do not depend on each other. Meanwhile, user preferences change rapidly due to fast changing game environment. There has been little work on designing friend suggestion when facing these difficulties, and for the first time we aim to tackle this in large scale online games. Motivated by the fast changing online game environment, we formulate this problem as friend ranking by modeling the evolution of similarity among users, exploiting the long-term and short-term feature of users in games. Our experiments on large-scale game datasets with several million users demonstrate that our proposed model achieves superior performance over other competing baselines.

cs.SI↗

arXiv2022

NRTR: Neuron Reconstruction with Transformer from 3D Optical Microscopy Images

The neuron reconstruction from raw Optical Microscopy (OM) image stacks is the basis of neuroscience. Manual annotation and semi-automatic neuron tracing algorithms are time-consuming and inefficient. Existing deep learning neuron reconstruction methods, although demonstrating exemplary performance, greatly demand complex rule-based components. Therefore, a crucial challenge is designing an end-to-end neuron reconstruction method that makes the overall framework simpler and model training easier. We propose a Neuron Reconstruction Transformer (NRTR) that, discarding the complex rule-based components, views neuron reconstruction as a direct set-prediction problem. To the best of our knowledge, NRTR is the first image-to-set deep learning model for end-to-end neuron reconstruction. In experiments using the BigNeuron and VISoR-40 datasets, NRTR achieves excellent neuron reconstruction results for comprehensive benchmarks and outperforms competitive baselines. Results of extensive experiments indicate that NRTR is effective at showing that neuron reconstruction is viewed as a set-prediction problem, which makes end-to-end model training available.

cs.CV↗

arXiv2022

Towards Improved Learning in Gaussian Processes: The Best of Two Worlds

Gaussian process training decomposes into inference of the (approximate) posterior and learning of the hyperparameters. For non-Gaussian (non-conjugate) likelihoods, two common choices for approximate inference are Expectation Propagation (EP) and Variational Inference (VI), which have complementary strengths and weaknesses. While VI's lower bound to the marginal likelihood is a suitable objective for inferring the approximate posterior, it does not automatically imply it is a good learning objective for hyperparameter optimization. We design a hybrid training procedure where the inference leverages conjugate-computation VI and the learning uses an EP-like marginal likelihood approximation. We empirically demonstrate on binary classification that this provides a good learning objective and generalizes better.

cs.LG↗

arXiv2022

LeMoN: Lens Modelling with Neural networks -- I. Automated modelling of strong gravitational lenses with Bayesian Neural Networks

The unprecedented number of gravitational lenses expected from new-generation facilities such as the ESA Euclid telescope and the Vera Rubin Observatory makes it crucial to rethink our classical approach to lens-modelling. In this paper, we present LeMoN (Lens Modelling with Neural networks): a new machine-learning algorithm able to analyse hundreds of thousands of gravitational lenses in a reasonable amount of time. The algorithm is based on a Bayesian Neural Network: a new generation of neural networks able to associate a reliable confidence interval to each predicted parameter. We train the algorithm to predict the three main parameters of the Singular Isothermal Ellipsoid model (the Einstein radius and the two components of the ellipticity) by employing two simulated datasets built to resemble the imaging capabilities of the Hubble Space Telescope and the forthcoming Euclid satellite. In this work, we assess the accuracy of the algorithm and the reliability of the estimated uncertainties by applying the network to several simulated datasets of 10.000 images each. We obtain accuracies comparable to previous studies present in the current literature and an average modelling time of just 0.5s per lens. Finally, we apply the LeMoN algorithm to a pilot dataset of real lenses observed with HST during the SLACS program, obtaining unbiased estimates of their SIE parameters. The code is publicly available on GitHub (https://github.com/fab-gentile/LeMoN).

astro-ph.IM↗

arXiv2022

Test-Time Training for Graph Neural Networks

Graph Neural Networks (GNNs) have made tremendous progress in the graph classification task. However, a performance gap between the training set and the test set has often been noticed. To bridge such gap, in this work we introduce the first test-time training framework for GNNs to enhance the model generalization capacity for the graph classification task. In particular, we design a novel test-time training strategy with self-supervised learning to adjust the GNN model for each test graph sample. Experiments on the benchmark datasets have demonstrated the effectiveness of the proposed framework, especially when there are distribution shifts between training set and test set. We have also conducted exploratory studies and theoretical analysis to gain deeper understandings on the rationality of the design of the proposed graph test time training framework (GT3).

cs.LG↗

arXiv2022

A Policy-based Approach to the SpecAugment Method for Low Resource E2E ASR

SpecAugment is a very effective data augmentation method for both HMM and E2E-based automatic speech recognition (ASR) systems. Especially, it also works in low-resource scenarios. However, SpecAugment masks the spectrum of time or the frequency domain in a fixed augmentation policy, which may bring relatively less data diversity to the low-resource ASR. In this paper, we propose a policy-based SpecAugment (Policy-SpecAugment) method to alleviate the above problem. The idea is to use the augmentation-select policy and the augmentation-parameter changing policy to solve the fixed way. These policies are learned based on the loss of validation set, which is applied to the corresponding augmentation policies. It aims to encourage the model to learn more diverse data, which the model relatively requires. In experiments, we evaluate the effectiveness of our approach in low-resource scenarios, i.e., the 100 hours librispeech task. According to the results and analysis, we can see that the above issue can be obviously alleviated using our proposal. In addition, the experimental results show that, compared with the state-of-the-art SpecAugment, the proposed Policy-SpecAugment has a relative WER reduction of more than 10% on the Test/Dev-clean set, more than 5% on the Test/Dev-other set, and an absolute WER reduction of more than 1% on all test sets.

cs.SD↗

arXiv2022

Galaxy-galaxy lensing in the VOICE deep survey

The multi-band photometry of the VOICE imaging data, overlapping with 4.9 deg$^2$ of the Chandra Deep Field South (CDFS) area, enables both shape measurement and photometric redshift estimation to be the two essential quantities for weak lensing analysis. The depth of $mag_{AB}$ is up to 26.1 (5$σ$ limiting) in $r$-band. We estimate the Excess Surface Density (ESD; $ΔΣ$) based on galaxy-galaxy measurements around galaxies at lower redshift (0.10<$z_l$<0.35) while we select the background sources to be at higher redshift ranging from 0.3 to 1.5. The foreground galaxies are divided into two major categories according to their colour (blue/red), each of which has been further divided into high/low stellar mass bins. Then the halo masses of the samples are estimated by modelling the signals, and the posterior of the parameters are samples via Mote Carlo Markov Chain (MCMC) process. We compare our results with the existing Stellar-to-Halo Mass Relation (SHMR) and find that the blue low stellar mass bin (median $M_*=10^{8.31}M_\odot$) deviates from the SHMR relation whereas all other three samples agrees well with empirical curves. We interpret this discrepancy as the effect of a low star formation efficiency of the low-mass blue dwarf galaxy population dominated in the VOICE-CDFS area.

astro-ph.CO↗

arXiv2022

Learning Multi-Modal Brain Tumor Segmentation from Privileged Semi-Paired MRI Images with Curriculum Disentanglement Learning

Due to the difficulties of obtaining multimodal paired images in clinical practice, recent studies propose to train brain tumor segmentation models with unpaired images and capture complementary information through modality translation. However, these models cannot fully exploit the complementary information from different modalities. In this work, we thus present a novel two-step (intra-modality and inter-modality) curriculum disentanglement learning framework to effectively utilize privileged semi-paired images, i.e. limited paired images that are only available in training, for brain tumor segmentation. Specifically, in the first step, we propose to conduct reconstruction and segmentation with augmented intra-modality style-consistent images. In the second step, the model jointly performs reconstruction, unsupervised/supervised translation, and segmentation for both unpaired and paired inter-modality images. A content consistency loss and a supervised translation loss are proposed to leverage complementary information from different modalities in this step. Through these two steps, our method effectively extracts modality-specific style codes describing the attenuation of tissue features and image contrast, and modality-invariant content codes containing anatomical and functional information from the input images. Experiments on three brain tumor segmentation tasks show that our model outperforms competing segmentation models based on unpaired images.

cs.CV↗