arXiv subjects

Yuting Wang

Publications and source records attributed to Yuting Wang.

At least 73 records · Page 4Linked to original sources

arXiv2023

MISSRec: Pre-training and Transferring Multi-modal Interest-aware Sequence Representation for Recommendation

The goal of sequential recommendation (SR) is to predict a user's potential interested items based on her/his historical interaction sequences. Most existing sequential recommenders are developed based on ID features, which, despite their widespread use, often underperform with sparse IDs and struggle with the cold-start problem. Besides, inconsistent ID mappings hinder the model's transferability, isolating similar recommendation domains that could have been co-optimized. This paper aims to address these issues by exploring the potential of multi-modal information in learning robust and generalizable sequence representations. We propose MISSRec, a multi-modal pre-training and transfer learning framework for SR. On the user side, we design a Transformer-based encoder-decoder model, where the contextual encoder learns to capture the sequence-level multi-modal user interests while a novel interest-aware decoder is developed to grasp item-modality-interest relations for better sequence representation. On the candidate item side, we adopt a dynamic fusion module to produce user-adaptive item representation, providing more precise matching between users and items. We pre-train the model with contrastive learning objectives and fine-tune it in an efficient manner. Extensive experiments demonstrate the effectiveness and flexibility of MISSRec, promising a practical solution for real-world recommendation scenarios. Data and code are available on \url{https://github.com/gimpong/MM23-MISSRec}.

cs.IR↗

arXiv2023

ALWOD: Active Learning for Weakly-Supervised Object Detection

Object detection (OD), a crucial vision task, remains challenged by the lack of large training datasets with precise object localization labels. In this work, we propose ALWOD, a new framework that addresses this problem by fusing active learning (AL) with weakly and semi-supervised object detection paradigms. Because the performance of AL critically depends on the model initialization, we propose a new auxiliary image generator strategy that utilizes an extremely small labeled set, coupled with a large weakly tagged set of images, as a warm-start for AL. We then propose a new AL acquisition function, another critical factor in AL success, that leverages the student-teacher OD pair disagreement and uncertainty to effectively propose the most informative images to annotate. Finally, to complete the AL loop, we introduce a new labeling task delegated to human annotators, based on selection and correction of model-proposed detections, which is both rapid and effective in labeling the informative images. We demonstrate, across several challenging benchmarks, that ALWOD significantly narrows the gap between the ODs trained on few partially labeled but strategically selected image instances and those that rely on the fully-labeled data. Our code is publicly available on https://github.com/seqam-lab/ALWOD.

cs.CV↗

arXiv2023

Graph-based Topology Reasoning for Driving Scenes

Understanding the road genome is essential to realize autonomous driving. This highly intelligent problem contains two aspects - the connection relationship of lanes, and the assignment relationship between lanes and traffic elements, where a comprehensive topology reasoning method is vacant. On one hand, previous map learning techniques struggle in deriving lane connectivity with segmentation or laneline paradigms; or prior lane topology-oriented approaches focus on centerline detection and neglect the interaction modeling. On the other hand, the traffic element to lane assignment problem is limited in the image domain, leaving how to construct the correspondence from two views an unexplored challenge. To address these issues, we present TopoNet, the first end-to-end framework capable of abstracting traffic knowledge beyond conventional perception tasks. To capture the driving scene topology, we introduce three key designs: (1) an embedding module to incorporate semantic knowledge from 2D elements into a unified feature space; (2) a curated scene graph neural network to model relationships and enable feature interaction inside the network; (3) instead of transmitting messages arbitrarily, a scene knowledge graph is devised to differentiate prior knowledge from various types of the road genome. We evaluate TopoNet on the challenging scene understanding benchmark, OpenLane-V2, where our approach outperforms all previous works by a great margin on all perceptual and topological metrics. The code is released at https://github.com/OpenDriveLab/TopoNet

cs.CV↗

arXiv2022

D2DF2WOD: Learning Object Proposals for Weakly-Supervised Object Detection via Progressive Domain Adaptation

Weakly-supervised object detection (WSOD) models attempt to leverage image-level annotations in lieu of accurate but costly-to-obtain object localization labels. This oftentimes leads to substandard object detection and localization at inference time. To tackle this issue, we propose D2DF2WOD, a Dual-Domain Fully-to-Weakly Supervised Object Detection framework that leverages synthetic data, annotated with precise object localization, to supplement a natural image target domain, where only image-level labels are available. In its warm-up domain adaptation stage, the model learns a fully-supervised object detector (FSOD) to improve the precision of the object proposals in the target domain, and at the same time learns target-domain-specific and detection-aware proposal features. In its main WSOD stage, a WSOD model is specifically tuned to the target domain. The feature extractor and the object proposal generator of the WSOD model are built upon the fine-tuned FSOD model. We test D2DF2WOD on five dual-domain image benchmarks. The results show that our method results in consistently improved object detection and localization compared with state-of-the-art methods.

cs.CV↗

arXiv2022

Contrastive Masked Autoencoders for Self-Supervised Video Hashing

Self-Supervised Video Hashing (SSVH) models learn to generate short binary representations for videos without ground-truth supervision, facilitating large-scale video retrieval efficiency and attracting increasing research attention. The success of SSVH lies in the understanding of video content and the ability to capture the semantic relation among unlabeled videos. Typically, state-of-the-art SSVH methods consider these two points in a two-stage training pipeline, where they firstly train an auxiliary network by instance-wise mask-and-predict tasks and secondly train a hashing model to preserve the pseudo-neighborhood structure transferred from the auxiliary network. This consecutive training strategy is inflexible and also unnecessary. In this paper, we propose a simple yet effective one-stage SSVH method called ConMH, which incorporates video semantic information and video similarity relationship understanding in a single stage. To capture video semantic information for better hashing learning, we adopt an encoder-decoder structure to reconstruct the video from its temporal-masked frames. Particularly, we find that a higher masking ratio helps video understanding. Besides, we fully exploit the similarity relationship between videos by maximizing agreement between two augmented views of a video, which contributes to more discriminative and robust hash codes. Extensive experiments on three large-scale video datasets (i.e., FCVID, ActivityNet and YFCC) indicate that ConMH achieves state-of-the-art results. Code is available at https://github.com/huangmozhi9527/ConMH.

cs.CV↗

arXiv2022

DCDChain: A Credible Architecture of Digital Copyright Detection Based on Blockchain

Copyright detection is an effective method to prevent piracy. However, untrustworthy detection parties may lead to falsified detection results. Due to its credibility and tamper resistance, blockchain has been applied to copyright protection. Previous works mainly utilized blockchain for reliable copyright information storage or copyrighted digital media trading. As far as we know, the problem of credible copyright detection has not been addressed. In this paper, we propose a credible copyright detection architecture based on the blockchain, called DCDChain. In this architecture, the detection agency first detects copyrights off the chain, then uploads the detection records to the blockchain. Since data on the blockchain are publicly accessible, media providers can verify the correctness of the copyright detection, and appeal to a smart contract if there is any dissent. The smart contract then arbitrates the disputes by verifying the correctness of detection on the chain. The detect-verify-and-arbitrate mechanism guarantees the credibility of copyright detection. Security analysis and experimental simulations show that the digital copyright detection architecture is credible, secure and efficient. The proposed credible copyright detection scheme is highly important for copyright protection. The future work is to improve the scheme by designing more effective locality sensitive hash algorithms for various digital media.

cs.CR↗

arXiv2022

TENET: Transformer Encoding Network for Effective Temporal Flow on Motion Prediction

This technical report presents an effective method for motion prediction in autonomous driving. We develop a Transformer-based method for input encoding and trajectory prediction. Besides, we propose the Temporal Flow Header to enhance the trajectory encoding. In the end, an efficient K-means ensemble method is used. Using our Transformer network and ensemble method, we win the first place of Argoverse 2 Motion Forecasting Challenge with the state-of-the-art brier-minFDE score of 1.90.

cs.CV↗

arXiv2022

A new consistency test for $Λ$CDM cosmology using galaxy surveys

We propose a new consistency test for the $Λ$CDM cosmology using baryonic acoustic oscillations (BAO) and redshift space distortion (RSD) measurements from galaxy redshift surveys. Specifically, we determine the peak position of $fσ_8(z)$ in redshift $z$ offered by a RSD measurement, and compare it to the one predicted by the BAO observables assuming a flat $Λ$CDM cosmology. We demonstrate this new test using the simulated data for the DESI galaxy survey, and argue that this test complements those using the background observables alone, and is less subject to systematics in the RSD analysis, compared to traditional methods using values of $fσ_8(z)$ directly.

astro-ph.CO↗

arXiv2022

Determination of neutron-skin thickness using configurational information entropy

Configurational information entropy (CIE) theory was employed to determine the neutron skin thickness of neutron-rich calcium isotopes. The nuclear density distributions and fragment cross-sections in 350 MeV/u $^{40-60}$Ca + $^{9}$Be projectile fragmentation reactions were calculated using a modified statistical abrasion-ablation model. CIE quantities were determined from the nuclear density, isotopic, mass, and charge distributions. The linear correlations between the CIE determined using the isotopic, mass, and charge distributions and the neutron skin thickness of the projectile nucleus show that CIE provides new methods to extract the neutron skin thickness of neutron-rich nuclei.

nucl-th↗

arXiv2021

Perfect codes in vertex-transitive graphs

Given a graph $Γ$, a perfect code in $Γ$ is an independent set $C$ of vertices of $Γ$ such that every vertex outside of $C$ is adjacent to a unique vertex in $C$, and a total perfect code in $Γ$ is a set $C$ of vertices of $Γ$ such that every vertex of $Γ$ is adjacent to a unique vertex in $C$. To study (total) perfect codes in vertex-transitive graphs, we generalize the concept of subgroup (total) perfect code of a finite group introduced in \cite{HXZ18} as follows: Given a finite group $G$ and a subgroup $H$ of $G$, a subgroup $A$ of $G$ containing $H$ is called a subgroup (total) perfect code of the pair $(G,H)$ if there exists a coset graph $Cos(G,H,U)$ such that the set consisting of left cosets of $H$ in $A$ is a (total) perfect code in $Cos(G,H,U)$. We give a necessary and sufficient condition for a subgroup $A$ of $G$ containing $H$ to be a (total) perfect code of the pair $(G,H)$ and generalize a few known results of subgroup (total) perfect codes of groups. We also construct some examples of subgroup perfect codes of the pair $(G,H)$ and propose a few problems for further research.

math.CO↗

arXiv2021

The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: a multi-tracer analysis in Fourier space for measuring the cosmic structure growth and expansion rate

We perform a joint BAO and RSD analysis using the eBOSS DR16 LRG and ELG samples in the redshift range of $z\in[0.6,1.1]$, and detect a RSD signal from the cross power spectrum at a $\sim4σ$ confidence level, i.e., $fσ_8=0.317\pm0.080$ at $z_{\rm eff}=0.77$. Based on the chained power spectrum, which is a new development in this work to mitigate the angular systematics, we measurement the BAO distances and growth rate simultaneously at two effective redshifts, namely, $D_{\rm M}/r_{\rm d} \ (z=0.70)=17.96\pm0.51, \ D_{\rm H}/r_{\rm d} \ (z=0.70)=21.22\pm1.20, \ fσ_8 \ (z=0.70) =0.43\pm0.05$, and $D_{\rm M}/r_{\rm d} \ (z=0.845)=18.90\pm0.78, \ D_{\rm H}/r_{\rm d} \ (z=0.845)=20.91\pm2.86, \ fσ_8 \ (z=0.845) =0.30\pm0.08$. Combined with BAO measurements including those from the eBOSS DR16 QSO and Lyman-$α$ sample, our measurement has raised the significance level of a nonzero $Ω_{\rm Λ}$ to $\sim11σ$.

astro-ph.CO↗

arXiv2021

The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: measurement of the BAO and growth rate of structure of the emission line galaxy sample from the anisotropic power spectrum between redshift 0.6 and 1.1

We analyse the large-scale clustering in Fourier space of emission line galaxies (ELG) from the Data Release 16 of the Sloan Digital Sky Survey IV extended Baryon Oscillation Spectroscopic Survey. The ELG sample contains 173,736 galaxies covering 1,170 square degrees in the redshift range $0.6 < z < 1.1$. We perform a BAO measurement from the post-reconstruction power spectrum monopole, and study redshift space distortions (RSD) in the first three even multipoles. Photometric variations yield fluctuations of both the angular and radial survey selection functions. Those are directly inferred from data, imposing integral constraints which we model consistently. The full data set has only a weak preference for a BAO feature ($1.4σ$). At the effective redshift $z_{\rm eff} = 0.845$ we measure $D_{\rm V}(z_{\rm eff})/r_{\rm drag} = 18.33_{-0.62}^{+0.57}$, with $D_{\rm V}$ the volume-averaged distance and $r_{\rm drag}$ the comoving sound horizon at the drag epoch. In combination with the RSD measurement, at $z_{\rm eff} = 0.85$ we find $fσ_8(z_{\rm eff}) = 0.289_{-0.096}^{+0.085}$, with $f$ the growth rate of structure and $σ_8$ the normalisation of the linear power spectrum, $D_{\rm H}(z_{\rm eff})/r_{\rm drag} = 20.0_{-2.2}^{+2.4}$ and $D_{\rm M}(z_{\rm eff})/r_{\rm drag} = 19.17 \pm 0.99$ with $D_{\rm H}$ and $D_{\rm M}$ the Hubble and comoving angular distances, respectively. These results are in agreement with those obtained in configuration space, thus allowing a consensus measurement of $fσ_8(z_{\rm eff}) = 0.315 \pm 0.095$, $D_{\rm H}(z_{\rm eff})/r_{\rm drag} = 19.6_{-2.1}^{+2.2}$ and $D_{\rm M}(z_{\rm eff})/r_{\rm drag} = 19.5 \pm 1.0$. This measurement is consistent with a flat $Λ$CDM model with Planck parameters.

astro-ph.CO↗

arXiv2020

The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: measurement of the BAO and growth rate of structure of the luminous red galaxy sample from the anisotropic power spectrum between redshifts 0.6 and 1.0

We analyse the clustering of the Sloan Digital Sky Survey IV extended Baryon Oscillation Spectroscopic Survey Data Release 16 luminous red galaxy sample (DR16 eBOSS LRG) in combination with the high redshift tail of the Sloan Digital Sky Survey III Baryon Oscillation Spectroscopic Survey Data Release 12 (DR12 BOSS CMASS). We measure the redshift space distortions (RSD) and also extract the longitudinal and transverse baryonic acoustic oscillation (BAO) scale from the anisotropic power spectrum signal inferred from 377,458 galaxies between redshifts 0.6 and 1.0, with effective redshift of $z_{\rm eff}=0.698$ and effective comoving volume of $2.72\,{\rm Gpc}^3$. After applying reconstruction we measure the BAO scale and infer $D_H(z_{\rm eff})/r_{\rm drag} = 19.30\pm 0.56$ and $D_M(z_{\rm eff})/r_{\rm drag} =17.86 \pm 0.37$. When we perform a redshift space distortions analysis on the pre-reconstructed catalogue on the monopole, quadrupole and hexadecapole we find, $D_H(z_{\rm eff})/r_{\rm drag} = 20.18\pm 0.78$, $D_M(z_{\rm eff})/r_{\rm drag} =17.49 \pm 0.52$ and $fσ_8(z_{\rm eff})=0.454\pm0.046$. We combine both sets of results along with the measurements in configuration space of \cite{LRG_corr} and report the following consensus values: $D_H(z_{\rm eff})/r_{\rm drag} = 19.77\pm 0.47$, $D_M(z_{\rm eff})/r_{\rm drag} = 17.65\pm 0.30$ and $fσ_8(z_{\rm eff})=0.473\pm 0.044$, which are in full agreement with the standard $Λ$CDM and GR predictions. These results represent the most precise measurements within the redshift range $0.6\leq z \leq 1.0$ and are the culmination of more than 8 years of SDSS observations.

astro-ph.CO↗

arXiv2020

The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: measurement of the BAO and growth rate of structure of the luminous red galaxy sample from the anisotropic correlation function between redshifts 0.6 and 1

We present the cosmological analysis of the configuration-space anisotropic clustering in the completed Sloan Digital Sky Survey IV (SDSS-IV) extended Baryon Oscillation Spectroscopic Survey (eBOSS) DR16 galaxy sample. This sample consists of luminous red galaxies (LRGs) spanning the redshift range $0.6 < z < 1$, at an effective redshift of $z_{\rm eff}=0.698$. It combines 174 816 eBOSS LRGs and 202 642 BOSS CMASS galaxies. We extract and model the baryon acoustic oscillations (BAO) and redshift-space distortions (RSD) features from the galaxy two-point correlation function to infer geometrical and dynamical cosmological constraints. The adopted methodology is extensively tested on a set of realistic simulations. The correlations between the inferred parameters from the BAO and full-shape correlation function analyses are estimated. This allows us to derive joint constraints on the three cosmological parameter combinations: $D_M(z)/r_d$, $D_H(z)/r_d$ and $fσ_8(z)$, where $D_M$ is the comoving angular diameter distance, $D_H$ is Hubble distance, $r_d$ is the comoving BAO scale, $f$ is the linear growth rate of structure, and $σ_8$ is the amplitude of linear matter perturbations. After combining the results with those from the parallel power spectrum analysis of Gil-Marin et al. 2020, we obtain the constraints: $D_M/r_d = 17.65 \pm 0.30$, $D_H/r_d = 19.77 \pm 0.47$, $fσ_8 = 0.473 \pm 0.044$. These measurements are consistent with a flat $Λ$CDM model with standard gravity.

astro-ph.CO↗

arXiv2020

The clustering of the SDSS-IV extended Baryon Oscillation Spectroscopic Survey DR16 luminous red galaxy and emission line galaxy samples: cosmic distance and structure growth measurements using multiple tracers in configuration space

We perform a multi-tracer analysis using the complete Sloan Digital Sky Survey IV (SDSS-IV) extended Baryon Oscillation Spectroscopic Survey (eBOSS) DR16 luminous red galaxy (LRG) and the DR16 emission line galaxy (ELG) samples in the configuration space, and successfully detect a cross correlation between the two samples, and find the growth rate to be $fσ_8=0.342 \pm 0.085$ ($\sim25$ per cent accuracy) from the cross sample alone. We perform a joint measurement of the baryonic acoustic oscillation (BAO) and redshift space distortion (RSD) parameters at a single effective redshift of $z_{\rm eff}= 0.77$, using the auto- and cross-correlation functions of the LRG and ELG samples, and find that the comoving angular diameter distance $D_M(z_{\rm eff})/r_d = 18.85\pm 0.38$, the Hubble distance $D_H(z_{\rm eff})/r_d = 19.64 \pm 0.57$, and $fσ_8(z_{\rm eff}) = 0.432 \pm 0.038$, which is consistent with a $Λ$CDM model at $68\%$ CL. Compared to the single-tracer analysis on the LRG sample, the Figure of Merit (FoM) of $α_{\perp}, α_{||}$ and $fσ_8$ is improved by a factor of $1.11$ in our multi-tracer analysis, and in particular, the statistical uncertainty of $fσ_8$ is reduced by $11.6 \%$.

astro-ph.CO↗

arXiv2020

A brief review on cosmological analysis of galaxy surveys with multiple tracers

Galaxy redshift surveys are one of the key probes in modern cosmology. In the data analysis of galaxy surveys, the precision of the statistical measurement is primarily limited by the cosmic variance on large scales. Fortunately, this limitation can in principle be evaded by observing multiple types of biased tracers. In this brief review, we present the idea of the multi-tracer method, outline key steps in the data analysis, and show several worked examples based on the GAMA, BOSS and eBOSS galaxy surveys.

astro-ph.CO↗

arXiv2020

The Completed SDSS-IV extended Baryon Oscillation Spectroscopic Survey: Growth rate of structure measurement from anisotropic clustering analysis in configuration space between redshift 0.6 and 1.1 for the Emission Line Galaxy sample

We present the anisotropic clustering of emission line galaxies (ELGs) from the Sloan Digital Sky Survey IV (SDSS-IV) extended Baryon Oscillation Spectroscopic Survey (eBOSS) Data Release 16 (DR16). Our sample is composed of 173,736 ELGs covering an area of 1170 deg$^2$ over the redshift range $0.6 \leq z \leq 1.1$. We use the Convolution Lagrangian Perturbation Theory in addition to the Gaussian Streaming Redshift-Space Distortions to model the Legendre multipoles of the anisotropic correlation function. We show that the eBOSS ELG correlation function measurement is affected by the contribution of a radial integral constraint that needs to be modelled to avoid biased results. To mitigate the effect from unknown angular systematics, we adopt a modified correlation function estimator that cancels out the angular modes from the clustering. At the effective redshift, $z_{\rm eff}=0.85$, including statistical and systematical uncertainties, we measure the linear growth rate of structure $fσ_8(z_{\rm eff}) = 0.35\pm0.10$, the Hubble distance $D_H(z_{\rm eff})/r_{\rm drag} = 19.1^{+1.9}_{-2.1}$ and the comoving angular diameter distance $D_M(z_{\rm eff})/r_{\rm drag} = 19.9\pm1.0$. These results are in agreement with the Fourier space analysis, leading to consensus values of: $fσ_8(z_{\rm eff}) = 0.315\pm0.095$, $D_H(z_{\rm eff})/r_{\rm drag} = 19.6^{+2.2}_{-2.1}$ and $D_M(z_{\rm eff})/r_{\rm drag} = 19.5\pm1.0$, consistent with $Λ$CDM model predictions with Planck parameters.

astro-ph.CO↗

arXiv2020

Finite Size Scaling at the Topological Transition: Bilinear-Biquadratic Spin-1 Chain

We consider a finite size scaling function across a topological phase transition in 1D models. For models of non-interacting fermions it was shown to be universal for all topological symmetry classes and markedly asymmetric between trivial and topological sides of the transition (Gulden et al 2016). Here we verify its universality for the topological transition between dimerized and Haldane phases in bilinear-biquadratic spin-1 chain. To this end we perform high-accuracy variational matrix product state simulations. We show that the scaling function, expressed in terms of $L/ξ$, where $L$ is the chain length and $ξ$ is the correlation length, coincides with that of three species of non-interacting massive Majorana fermions. The latter is known to be a proper description of the conformal critical theory with central charge $c=3/2$. We have shown that it still holds away from the conformal point, including the finite size corrections. We have also observed peculiar differences between even and odd size chains, which may be fully accounted for by residual interactions of the edge states.

cond-mat.stat-mech↗