Searcharxiv⌕ Search

arXiv subjects

Yang Gao

Publications and source records attributed to Yang Gao.

At least 469 records · Page 26Linked to original sources
arXiv2023

Study of the doubly Cabibbo-suppressed decays $D^+_s\to K^+K^+π^-$ and $D^+_s\to K^+K^+π^-π^0$

Based on 7.33 fb$^{-1}$ of $e^+e^-$ collision data collected at center-of-mass energies between 4.128 and 4.226 GeV with the BESIII detector, the experimental studies of the doubly Cabibbo-suppressed decays $D^+_s\to K^+K^+π^-$ and $D^+_s\to K^+K^+π^-π^0$ are reported. We determine the absolute branching fraction of $D^+_s\to K^+K^+π^-$ to be (${1.23^{+0.28}_{-0.25}}({\rm stat})\pm0.06({\rm syst})$) $\times 10^{-4}$. No significant signal of $D^+_s\to K^+K^+π^-π^0$ is observed and the upper limit on its decay branching fraction at 90\% confidence level is set to be $1.7\times10^{-4}$.

hep-ex↗
arXiv2023

Observation of the $ψ(3686)$ decays into $Σ^{+}\barΣ^{-}ω$ and $Σ^{+}\barΣ^{-}{\mathcalϕ}$

Based on $(27.08\pm 0.14)\times10^{8}$ $ψ(3686)$ events collected with the BESIII detector operating at the BEPCII collider, the $ψ(3686)\toΣ^{+}\barΣ^{-}ω$ and $Σ^{+}\barΣ^{-}ϕ$ decays are observed for the first time with statistical significances of 13.8$σ$ and 7.6$σ$, respectively. The corresponding branching fractions are measured to be $\mathcal{B}(ψ(3686)\toΣ^{+}\barΣ^{-}ω)=(1.90 \pm 0.18 \pm 0.21) \times 10^{-5}$ and $\mathcal{B}(ψ(3686)\toΣ^{+}\barΣ^{-}ϕ)=(2.96 \pm 0.54 \pm 0.41) \times 10^{-6}$, where the first uncertainties are statistical and the second systematic.

hep-ex↗
arXiv2023

Measurement of the branching fractions of the singly Cabibbo-suppressed decays $Λ_{c}^{+}\to pη$ and $Λ_{c}^{+}\to pω$

Based on 4.5 $\mbox{fb$^{-1}$}$ $e^{+}e^{-}$ collision data collected with BESIII detector at seven energy points between 4.600 and 4.699 GeV, the branching fractions for $Λ_{c}^{+}\to pη$ and $Λ_{c}^{+}\to pω$ were measured by means of single-tag method. The branching fractions of $Λ_{c}^{+}\to pη$ and $Λ_{c}^{+}\to pω$ are determined to be $(1.57\pm0.11_{\rm {stat}}\pm0.04_{\rm{syst}})\times10^{-3}$ and $(1.11\pm0.20_{\rm{stat}}\pm0.07_{\rm{syst}})\times10^{-3}$, with a statistical significance of greater than 10 $σ$ and 5.7 $σ$, respectively. These results are consistent with the previous measurements by BESIII, LHCb and Belle, and the result of $Λ_{c}^{+}\to pη$ is the most precise to date.

hep-ex↗
arXiv2023

A Policy Optimization Method Towards Optimal-time Stability

In current model-free reinforcement learning (RL) algorithms, stability criteria based on sampling methods are commonly utilized to guide policy optimization. However, these criteria only guarantee the infinite-time convergence of the system's state to an equilibrium point, which leads to sub-optimality of the policy. In this paper, we propose a policy optimization technique incorporating sampling-based Lyapunov stability. Our approach enables the system's state to reach an equilibrium point within an optimal time and maintain stability thereafter, referred to as "optimal-time stability". To achieve this, we integrate the optimization method into the Actor-Critic framework, resulting in the development of the Adaptive Lyapunov-based Actor-Critic (ALAC) algorithm. Through evaluations conducted on ten robotic tasks, our approach outperforms previous studies significantly, effectively guiding the system to generate stable patterns.

cs.RO↗
arXiv2023

A Universal Semantic-Geometric Representation for Robotic Manipulation

Robots rely heavily on sensors, especially RGB and depth cameras, to perceive and interact with the world. RGB cameras record 2D images with rich semantic information while missing precise spatial information. On the other side, depth cameras offer critical 3D geometry data but capture limited semantics. Therefore, integrating both modalities is crucial for learning representations for robotic perception and control. However, current research predominantly focuses on only one of these modalities, neglecting the benefits of incorporating both. To this end, we present $\textbf{Semantic-Geometric Representation} (\textbf{SGR})$, a universal perception module for robotics that leverages the rich semantic information of large-scale pre-trained 2D models and inherits the merits of 3D spatial reasoning. Our experiments demonstrate that SGR empowers the agent to successfully complete a diverse range of simulated and real-world robotic manipulation tasks, outperforming state-of-the-art methods significantly in both single-task and multi-task settings. Furthermore, SGR possesses the capability to generalize to novel semantic attributes, setting it apart from the other methods. Project website: https://semantic-geometric-representation.github.io.

cs.RO↗
arXiv2023

Search for $J/ψ$ weak decays containing $D$ meson

Using a sample of about 10 billion $J/ψ$ events with the BESIII detector, we search for the weak decays of $J/ψ\to \bar{D}^0π^0 + c.c.$, $J/ψ\to \bar{D}^0η+ c.c.$, $J/ψ\to \bar{D}^0ρ^0 + c.c.$, $J/ψ\to D^-π^+ + c.c.$, and $J/ψ\to D^-ρ^+ + c.c.$. Since no significant signal is observed, we set the upper limits of the branching fractions of these decays to be $\mathcal{B}(J/ψ\to \bar{D}^0π^0 + c.c.) < 4.7 \times 10^{-7}$, $\mathcal{B}(J/ψ\to \bar{D}^0η+ c.c.) < 6.8 \times 10^{-7}$, $\mathcal{B}(J/ψ\to \bar{D}^0ρ^0 + c.c.) < 5.2 \times 10^{-7}$, $\mathcal{B}(J/ψ\to D^-π^+ + c.c.) < 7.0 \times 10^{-8}$, and $\mathcal{B}(J/ψ\to D^-ρ^+ + c.c.) < 6.0 \times 10^{-7}$ at the 90\% confidence level.

hep-ex↗
arXiv2023

Measurement of $e^{+}e^{-}\rightarrowηJ/ψ$ Cross Section from $\sqrt{s}=$ 3.808 GeV to 4.951 GeV

Using data samples with an integrated luminosity of 22.42 fb$^{-1}$ collected by the BESIII detector operating at the BEPCII storage ring, we measure the cross sections of the $e^{+}e^{-}\rightarrow\etaJ/ψ$ process at center-of-mass energies from 3.808 to 4.951 GeV. Three structures are observed in the line shape of the measured cross sections. A maximum-likelihood fit with $ψ(4040)$, two additional resonances, and a non-resonant component is performed. The mass and width of the first additional state are $(4219.7\pm2.5\pm4.5) \rm{MeV}/\rm{c}^2$ and $(80.7\pm4.4\pm1.4) \rm{MeV}$, respectively, consistent with the $ψ(4230)$. For the second state, the mass and width are $(4386\pm13\pm17) \rm{MeV}/\rm{c}^2$ and $(177\pm32\pm13) \rm{MeV}$, respectively, consistent with the $ψ(4360)$. The first uncertainties are statistical and the second ones are systematic. The statistical significance of $ψ(4040)$ is $8.0σ$ and those for $ψ(4230)$ and $ψ(4360)$ are more than $10.0σ$.

hep-ex↗
arXiv2023

First measurement of $ΛN$ inelastic scattering with $Λ$ from $e^{+} e^{-} \rightarrow J/ψ\to Λ\barΛ$

Using an $e^+ e^-$ collision data sample of $(10087 \pm 44)\times10^6 ~J/ψ$ events taken at the center-of-mass energy of $3.097~\rm{GeV}$ by the BESIII detector at the BEPCII collider, the process $Λ+N \rightarrow Σ^+ + X$ is studied for the first time employing a novel method. The $Σ^{+}$ hyperons are produced by the collisions of $Λ$ hyperons from $J/ψ$ decays with nuclei in the material of the BESIII detector. The total cross section of $Λ+ ^{9}{\rm Be} \rightarrow Σ^+ + X$ is measured to be $σ= (37.3 \pm 4.7 \pm 3.5)~{\rm mb}$ at $Λ$ beam momenta within $[1.057, 1.091]~{\rm GeV}/c$, where the uncertainties are statistical and systematic, respectively. This analysis is the first study of $Λ$-nucleon interactions at an $e^+ e^-$ collider, providing information and constraints relevant for the strong-interaction potential, the origin of color confinement, the unified model for baryon-baryon interactions, and the internal structure of neutron stars.

nucl-ex↗
arXiv2023

DNA-Rendering: A Diverse Neural Actor Repository for High-Fidelity Human-centric Rendering

Realistic human-centric rendering plays a key role in both computer vision and computer graphics. Rapid progress has been made in the algorithm aspect over the years, yet existing human-centric rendering datasets and benchmarks are rather impoverished in terms of diversity, which are crucial for rendering effect. Researchers are usually constrained to explore and evaluate a small set of rendering problems on current datasets, while real-world applications require methods to be robust across different scenarios. In this work, we present DNA-Rendering, a large-scale, high-fidelity repository of human performance data for neural actor rendering. DNA-Rendering presents several alluring attributes. First, our dataset contains over 1500 human subjects, 5000 motion sequences, and 67.5M frames' data volume. Second, we provide rich assets for each subject -- 2D/3D human body keypoints, foreground masks, SMPLX models, cloth/accessory materials, multi-view images, and videos. These assets boost the current method's accuracy on downstream rendering tasks. Third, we construct a professional multi-view system to capture data, which contains 60 synchronous cameras with max 4096 x 3000 resolution, 15 fps speed, and stern camera calibration steps, ensuring high-quality resources for task training and evaluation. Along with the dataset, we provide a large-scale and quantitative benchmark in full-scale, with multiple tasks to evaluate the existing progress of novel view synthesis, novel pose animation synthesis, and novel identity rendering methods. In this manuscript, we describe our DNA-Rendering effort as a revealing of new observations, challenges, and future directions to human-centric rendering. The dataset, code, and benchmarks will be publicly available at https://dna-rendering.github.io/

cs.CV↗
arXiv2023

CasIL: Cognizing and Imitating Skills via a Dual Cognition-Action Architecture

Enabling robots to effectively imitate expert skills in longhorizon tasks such as locomotion, manipulation, and more, poses a long-standing challenge. Existing imitation learning (IL) approaches for robots still grapple with sub-optimal performance in complex tasks. In this paper, we consider how this challenge can be addressed within the human cognitive priors. Heuristically, we extend the usual notion of action to a dual Cognition (high-level)-Action (low-level) architecture by introducing intuitive human cognitive priors, and propose a novel skill IL framework through human-robot interaction, called Cognition-Action-based Skill Imitation Learning (CasIL), for the robotic agent to effectively cognize and imitate the critical skills from raw visual demonstrations. CasIL enables both cognition and action imitation, while high-level skill cognition explicitly guides low-level primitive actions, providing robustness and reliability to the entire skill IL process. We evaluated our method on MuJoCo and RLBench benchmarks, as well as on the obstacle avoidance and point-goal navigation tasks for quadrupedal robot locomotion. Experimental results show that our CasIL consistently achieves competitive and robust skill imitation capability compared to other counterparts in a variety of long-horizon robotic tasks.

cs.RO↗
arXiv2023

Updated measurements of the M1 transition $ψ(3686) \to γη_{c}(2S)$ with $η_{c}(2S) \to K \bar{K} π$

Based on a data sample of $(27.08 \pm 0.14 ) \times 10^8~ψ(3686)$ events collected with the BESIII detector at the BEPCII collider, the M1 transition $ψ(3686) \to γη_{c}(2S)$ with $η_{c}(2S) \to K\bar{K}π$ is studied, where $K\bar{K}π$ is $K^{+} K^{-} π^{0}$ or $K_{S}^{0}K^{\pm}π^{\mp}$. The mass and width of the $η_{c}(2S)$ are measured to be $(3637.8 \pm 0.8 (\rm {stat}) \pm 0.2 (\rm {syst}))$ MeV/$c^{2}$ and $(10.5 \pm 1.7 (\rm {stat}) \pm 3.5 (\rm {syst}))$ MeV, respectively. The product branching fraction $\mathcal{B}\left(ψ(3686) \rightarrow γη_{c}(2 S)\right) \times \mathcal{B}(η_{c}(2 S) \rightarrow K \bar{K} π)$ is determined to be $(0.97 \pm 0.06 (\rm {stat}) \pm 0.09 (\rm {syst})) \times 10^{-5}$. Using $\mathcal{BR}(η_{c}(2S)\to K\bar{K}π)=(1.86^{+0.68}_{-0.49})\%$, we obtain the branching fraction of the radiative transition to be $\mathcal{BR}(ψ(3686) \to γη_{c}(2S)) = (5.2 \pm 0.3 (\rm {stat}) \pm 0.5 (\rm {syst}) ^{+1.9}_{-1.4} (extr)) \times 10^{-4}$, where the third uncertainty is due to the quoted $\mathcal{BR}(η_{c}(2S) \to K\bar{K}π)$.

hep-ex↗
arXiv2023

OpenMSD: Towards Multilingual Scientific Documents Similarity Measurement

We develop and evaluate multilingual scientific documents similarity measurement models in this work. Such models can be used to find related works in different languages, which can help multilingual researchers find and explore papers more efficiently. We propose the first multilingual scientific documents dataset, Open-access Multilingual Scientific Documents (OpenMSD), which has 74M papers in 103 languages and 778M citation pairs. With OpenMSD, we pretrain science-specialized language models, and explore different strategies to derive "related" paper pairs to fine-tune the models, including using a mixture of citation, co-citation, and bibliographic-coupling pairs. To further improve the models' performance for non-English papers, we explore the use of generative language models to enrich the non-English papers with English summaries. This allows us to leverage the models' English capabilities to create better representations for non-English papers. Our best model significantly outperforms strong baselines by 7-16% (in mean average precision).

cs.CL↗
arXiv2023

Observation of $J/ψ$ decays to $e^{+}e^{-}e^{+}e^{-}$ and $e^{+}e^{-}μ^{+}μ^{-}$

Using a data sample of $4.481\times 10^8 ψ^\prime$ events collected with the BESIII detector, we report the first observation of the four-lepton-decays $J/ψ\to e^+e^-e^+e^-$ and $J/ψ\to e^+e^-μ^+μ^-$ utilizing the process $ψ^\prime\to π^+π^- J/ψ$. The branching fractions are determined to be $[5.48\pm0.31~(\rm stat)\pm0.45~(\rm syst)]\times 10^{-5}$ and $[3.53~\pm0.22~(\rm stat)\pm0.13~(\rm syst)]\times 10^{-5}$, respectively. The results are consistent with theoretical predictions. No significant signal is observed for $J/ψ\to μ^+μ^-μ^+μ^-$, and an upper limit on the branching fraction is set at $1.6\times 10^{-6}$ at the 90$\%$ confidence level. A CP asymmetry observable is constructed for the first two channels, which is measured to be $(-0.012\pm0.054\pm0.010)$ and $(0.062\pm0.059\pm0.006)$, respectively. No evidence for CP violation is observed in this process.

hep-ex↗
arXiv2023

PLACE dropout: A Progressive Layer-wise and Channel-wise Dropout for Domain Generalization

Domain generalization (DG) aims to learn a generic model from multiple observed source domains that generalizes well to arbitrary unseen target domains without further training. The major challenge in DG is that the model inevitably faces a severe overfitting issue due to the domain gap between source and target domains. To mitigate this problem, some dropout-based methods have been proposed to resist overfitting by discarding part of the representation of the intermediate layers. However, we observe that most of these methods only conduct the dropout operation in some specific layers, leading to an insufficient regularization effect on the model. We argue that applying dropout at multiple layers can produce stronger regularization effects, which could alleviate the overfitting problem on source domains more adequately than previous layer-specific dropout methods. In this paper, we develop a novel layer-wise and channel-wise dropout for DG, which randomly selects one layer and then randomly selects its channels to conduct dropout. Particularly, the proposed method can generate a variety of data variants to better deal with the overfitting issue. We also provide theoretical analysis for our dropout method and prove that it can effectively reduce the generalization error bound. Besides, we leverage the progressive scheme to increase the dropout ratio with the training progress, which can gradually boost the difficulty of training the model to enhance its robustness. Extensive experiments on three standard benchmark datasets have demonstrated that our method outperforms several state-of-the-art DG methods. Our code is available at https://github.com/lingeringlight/PLACEdropout.

cs.LG↗
arXiv2023

Exploring Flat Minima for Domain Generalization with Large Learning Rates

Domain Generalization (DG) aims to generalize to arbitrary unseen domains. A promising approach to improve model generalization in DG is the identification of flat minima. One typical method for this task is SWAD, which involves averaging weights along the training trajectory. However, the success of weight averaging depends on the diversity of weights, which is limited when training with a small learning rate. Instead, we observe that leveraging a large learning rate can simultaneously promote weight diversity and facilitate the identification of flat regions in the loss landscape. However, employing a large learning rate suffers from the convergence problem, which cannot be resolved by simply averaging the training weights. To address this issue, we introduce a training strategy called Lookahead which involves the weight interpolation, instead of average, between fast and slow weights. The fast weight explores the weight space with a large learning rate, which is not converged while the slow weight interpolates with it to ensure the convergence. Besides, weight interpolation also helps identify flat minima by implicitly optimizing the local entropy loss that measures flatness. To further prevent overfitting during training, we propose two variants to regularize the training weight with weighted averaged weight or with accumulated history weight. Taking advantage of this new perspective, our methods achieve state-of-the-art performance on both classification and semantic segmentation domain generalization benchmarks. The code is available at https://github.com/koncle/DG-with-Large-LR.

cs.CV↗
arXiv2023

Measurements of the absolute branching fractions of $Ω^-$ decays and test of the $ΔI = 1/2$ rule

Based on a data set of $(27.12\pm0.10)\times 10^8$ $ψ(3686)$ events collected at the BESIII experiment, the absolute branching fractions of the three dominant $Ω^-$ decays are measured to be $\mathcal{B}_{Ω^- \to Ξ^0 π^-} = (25.03\pm0.44\pm0.53)\%$, $\mathcal{B}_{Ω^- \to Ξ^- π^0} = (8.43\pm0.52\pm0.28)\%$, and $\mathcal{B}_{Ω^- \to ΛK^-} = (66.3\pm0.8\pm2.0)\%$, where the first and second uncertainties are statistical and systematic, respectively. The ratio between $\mathcal{B}_{Ω^- \to Ξ^0 π^-}$ and $\mathcal{B}_{Ω^- \to Ξ^- π^0}$ is determined to be $2.97\pm0.19\pm0.11$, which is in good agreement with the PDG value of $2.74\pm0.15$, but greater by more than four standard deviations than the theoretical prediction of 2 obtained from the $ΔI = 1/2$ rule.

hep-ex↗
arXiv2023

Measurement of $e^{+}e^{-}\toϕη'$ cross sections at center-of-mass energies from 3.508 to 4.951 GeV and search for the decay $ψ(3770)\toϕη'$

The cross sections of the $e^{+}e^{-}\toϕη'$ process at center-of-mass energies from 3.508 to 4.951 GeV are measured with high precision using 26.1 fb$^{-1}$ data collected with the BESIII detector operating at the BEPCII storage ring. The cross sections are of the order of a few picobarn, and decrease as the center-of-mass energy increases as $s^{-n/2}$ with $n=4.35\pm 0.14$. This result is in agreement with the Nambu-Jona-Lasinio model prediction of $n=3.5\pm 0.9$. In addition, the charmless decay $ψ(3770)\toϕη'$ is searched for by fitting the measured cross sections, yet no significant signal is observed. The upper limit of ${\cal B}(ψ(3770)\toϕη')$ at the 90\% confidence level is determined to be $2.3\times 10^{-5}$.

hep-ex↗
arXiv2023

Search for the semileptonic decays $D^+_s \to K_1(1270)^0 e^+ν_e$ and $D^+_s \to b_1(1235)^0 e^+ν_e$

By analyzing 7.33\,fb$^{-1}$ of $e^+e^-$ collision data collected at center-of-mass energies between 4.128 and 4.226 GeV with the BESIII detector, we search for the semileptonic decays $D^+_s \to K_1(1270)^0 e^+ν_e$ and $D^+_s \to b_1(1235)^0 e^+ν_e$ for the first time. No significant signals are observed for either decay mode. The upper limits on the (product) branching fractions are determined to be ${\mathcal B}[D^+_s \to K_1(1270)^0 e^+ν_e] < 4.1\times 10^{-4}$ and ${\mathcal B}[D^+_s \to b_1(1235)^0 e^+ν_e]\cdot {\mathcal B}[b_1(1235)^0\to ωπ^0] < 6.4\times 10^{-4}$ at 90\% confidence level.

hep-ex↗