Searcharxiv⌕ Search

arXiv subjects

Yang Sun

Publications and source records attributed to Yang Sun.

At least 127 records · Page 7Linked to original sources
arXiv2024

Elasticity and acoustic velocities of $δ$-AlOOH at extreme conditions: a methodology assessment

Hydrous phases play a fundamental role in the deep-water cycle on Earth. Understanding their stability and thermoelastic properties is essential for constraining their abundance using seismic tomography. However, determining their elastic properties at extreme conditions is notoriously challenging. The challenges stem from the complex behavior of hydrogen bonds under high pressures and temperatures (P,Ts). In this study, we evaluate how advanced molecular dynamics simulation techniques can address these challenges by investigating the adiabatic elasticity and acoustic velocities of $δ$-AlOOH, a critical and prototypical high-pressure hydrous phase. We compared the performances of three methods to assess their viability and accuracy. The thermoelastic tensor was computed up to 140 GPa and temperatures up to 2,700 K using molecular dynamics with a DeePMD machine-learning interatomic potential based on the SCAN meta-GGA functional. The excellent agreement with ambient condition single-crystal ultrasound measurements and the correct description of velocity changes induced by H-bond disorder-symmetrization transition observed at 10 GPa in Brillouin scattering measurements underscores the accuracy and efficacy of our approach.

physics.comp-ph↗
arXiv2024

Unveiling a Family of Dimerized Quantum Magnets in Ternary Metal Borides

Dimerized quantum magnets are exotic crystalline materials where Bose-Einstein condensation of magnetic excitations can happen. However, known dimerized quantum magnets are limited to only a few oxides and halides. Here, we unveil 9 dimerized quantum magnets and 11 conventional antiferromagnets in ternary metal borides MTB$_4$ (M = Sc, Y, La, Ce, Lu, Mg, Ca, Al; T = V, Cr, Mn, Fe, Co, Ni). In this type of structure, 3d transition-metal atoms T are arranged in dimers. Quantum magnetism in these compounds is dominated by strong antiferromagnetic interactions between Cr (both Cr and Mn for M = Mg and Ca) atoms within the structural dimers, with much weaker interactions between the dimers. These systems are proposed to be close to a quantum critical point between a disordered singlet spin-dimer phase, with a spin gap, and the ordered conventional Néel antiferromagnetic phase. This new family of dimerized quantum magnets greatly enriches the materials inventory that allows investigations of the spin-gap phase. All the quantum-, conventionally-, and non-magnetic systems identified, together with experimental synthesis methods of a phase suitable for characterization, provide a platform with abundant possibilities to tune the magnetic exchange coupling by doping and study this unconventional type of quantum phase transition. This work opens up new avenues for studying the quantum magnetism of spin dimers in borides and establishes a theoretical workflow for future searches for dimerized quantum magnets in other families or types of materials.

cond-mat.mtrl-sci↗
arXiv2024

SMILES Initial Data Release: Unveiling the Obscured Universe with MIRI Multi-band Imaging

The James Webb Space Telescope (JWST) is revolutionizing our view of the Universe through unprecedented sensitivity and resolution in the infrared, with some of the largest gains realized at its longest wavelengths. We present the Systematic Mid-infrared Instrument (MIRI) Legacy Extragalactic Survey (SMILES), an eight-band MIRI survey with Near-Infrared Spectrograph (NIRSpec) spectroscopic follow-up in the GOODS-S/HUDF region. SMILES takes full advantage of MIRI's continuous coverage from $5.6-25.5\,μ$m over a $\sim34$ arcmin$^2$ area to greatly expand our understanding of the obscured Universe up to cosmic noon and beyond. This work, together with a companion paper by Rieke et al., covers the SMILES science drivers and technical design, early results with SMILES, data reduction, photometric catalog creation, and the first data release. As part of the discussion on early results, we additionally present a high-level science demonstration on how MIRI's wavelength coverage and resolution will advance our understanding of cosmic dust using the full range of polycyclic aromatic hydrocarbon (PAH) emission features from $3.3-18\,μ$m. Using custom background subtraction, we produce robust reductions of the MIRI imaging that maximize the depths reached with our modest exposure times ($\sim0.6 - 2.2$ ks per filter). Included in our initial data release are (1) eight MIRI imaging mosaics reaching depths of $0.2-18\,μ$Jy ($5σ$) and (2) a $5-25.5\,μ$m photometric catalog with over 3,000 sources. Building upon the rich legacy of extensive photometric and spectroscopy coverage of GOODS-S/HUDF from the X-ray to the radio, SMILES greatly expands our investigative power in understanding the obscured Universe.

astro-ph.GA↗
arXiv2024

High-throughput screening for boride superconductors

A high-throughput screening using density functional calculations is performed to search for stable boride superconductors from the existing materials database. The workflow employs the fast frozen phonon method as the descriptor to evaluate the superconducting properties quickly. 23 stable candidates are identified from the screening. For almost all found binary compounds, the superconductivity was obtained earlier experimentally or computationally. For ternary borides, previous studies are very limited. Our extensive search among ternary systems confirmed superconductivity in known systems and found several new compounds. Among these discovered superconducting ternary borides, Ta(MoB)$_2$ shows the highest superconducting temperature of ~12K. Most predicted compounds were synthesized previously; therefore, our predictions can be examined experimentally. Our work also demonstrates that the boride systems can have diverse structural motifs that lead to superconductivity.

cond-mat.mtrl-sci↗
arXiv2024

Melting Temperature of Iron Under the Earth's Inner Core Condition from Deep Machine Learning

Constraining the melting temperature of iron under Earth's inner core conditions is crucial for understanding core dynamics and planetary evolution. Here, we develop a deep potential (DP) model for iron that explicitly incorporates electronic entropy contributions governing thermodynamics under Earth's core conditions. Extensive benchmarking demonstrates the DP's high fidelity across relevant iron phases and extreme pressure and temperature conditions. Through thermodynamic integration and direct solid-liquid coexistence simulations, the DP predicts melting temperatures for iron at the inner core boundary, consistent with previous \textit{ab initio} results. This resolves the previous discrepancy of iron's melting temperature at ICB between the DP model and \textit{ab initio} calculation and suggests the crucial contribution of electronic entropy. Our work provides insights into machine learning melting behavior of iron under core conditions and provides the basis for future development of binary or ternary DP models for iron and other elements in the core.

physics.geo-ph↗
arXiv2024

Task-Customized Mixture of Adapters for General Image Fusion

General image fusion aims at integrating important information from multi-source images. However, due to the significant cross-task gap, the respective fusion mechanism varies considerably in practice, resulting in limited performance across subtasks. To handle this problem, we propose a novel task-customized mixture of adapters (TC-MoA) for general image fusion, adaptively prompting various fusion tasks in a unified model. We borrow the insight from the mixture of experts (MoE), taking the experts as efficient tuning adapters to prompt a pre-trained foundation model. These adapters are shared across different tasks and constrained by mutual information regularization, ensuring compatibility with different tasks while complementarity for multi-source images. The task-specific routing networks customize these adapters to extract task-specific information from different sources with dynamic dominant intensity, performing adaptive visual feature prompt fusion. Notably, our TC-MoA controls the dominant intensity bias for different fusion tasks, successfully unifying multiple fusion tasks in a single model. Extensive experiments show that TC-MoA outperforms the competing approaches in learning commonalities while retaining compatibility for general image fusion (multi-modal, multi-exposure, and multi-focus), and also demonstrating striking controllability on more generalization experiments. The code is available at https://github.com/YangSun22/TC-MoA .

cs.CV↗
arXiv2024

Probing the state of hydrogen in $δ$-AlOOH at mantle conditions with machine learning potential

Hydrous and nominally anhydrous minerals (NAMs) are a fundamental class of solids of enormous significance to geophysics. They are the water carriers in the deep geological water cycle and impact structural, elastic, plastic, and thermodynamic properties and phase relations in Earth's forming aggregates (rocks). They play a critical role in the geochemical and geophysical processes that shape the planet. Their complexity has prevented predictive calculations of their properties, but progress in materials simulations ushered by machine learning potentials is transforming this state of affairs. Here, we adopt a hybrid approach that combines deep learning potentials (DP) with the SCAN meta-GGA functional to simulate a prototypical hydrous system. We illustrate the success of this approach to simulate $δ$-AlOOH ($δ$), a phase capable of transporting water down to near the core-mantle boundary of the Earth (~2,900 km depth and ~135 GPa) in subducting slabs. A high-throughput sampling of phase space using molecular dynamics simulations with DP-potentials sheds light on the hydrogen-bond behavior and proton diffusion at geophysical conditions. These simulations provide a pathway for a deeper understanding of these crucial components that shape Earth's internal state.

physics.comp-ph↗
arXiv2024

Prediction of ambient pressure superconductivity in cubic ternary hydrides with MH$_6$ octahedra

Exploring high-temperature superconducting (high-$T_c$) material at ambient pressure holds immense significance for physics, chemistry, and materials science. In this study, we perform a high-throughput screening of strong electron-phonon interactions in X$_2$MH$_6$ compounds (X = Li, Na, Mg, Al, K, Ca, Ga, Rb, Sr, and In; M are $3d$, $4d$, and $5d$ transition metals). These compounds have a cubic structure featuring an MH$_6$ octahedron motif. Our screening calculations suggest that 26 compounds exhibit dynamic stability and strong electron-phonon coupling. Among these 26 compounds, Mg$_2$RhH$_6$, Mg$_2$IrH$_6$, Al$_2$MnH$_6$, and Li$_2$CuH$_6$ show promising energetic stability and $T_c$ of more than 50 K at ambient pressure. This study underscores promising high-$T_c$ compounds at ambient pressure with distinctive MH$_6$ motifs.

cond-mat.mtrl-sci↗
arXiv2024

Evolution of Gas Flows along the Starburst to Post-Starburst to Quiescent Galaxy Sequence

We measure velocity offsets in the NaI $λ\lambda5890, 5896$ (Na D) interstellar medium absorption lines to track how neutral galactic winds change as their host galaxies evolve. Our sample of $\sim$80,000 SDSS spectra at $0.010 < z < 0.325$ includes starburst, post-starburst, and quiescent galaxies, forming an evolutionary sequence of declining star formation rate (SFR). We detect bulk flows across this sequence, mostly at higher host stellar masses($log(M_{\star}/M_{\odot})>10$). Along this sequence, the fraction of outflows decreases ($76\pm2\%$ to $65\pm4\%$ to a 3$σ$ upper limit of $34\%$), and the mean velocity offset changes from outflowing to inflowing ($-84.6\pm5.9$ to $-71.6\pm11.4$ to $76.6\pm2.3\,km\,s^{-1}$). Even within the post-starburst sample, wind speed decreases with time elapsed since the starburst ended. These results reveal that outflows diminish as galaxies age. For post-starbursts, there is evidence for an AGN contribution, especially to the speediest outflows: 1) SFR declines faster in time than outflow velocity, a decoupling arguing against massive stellar feedback; 2) of the few outflows strong enough to escape the interstellar medium (9/105), three of the four hosts with measured emission lines are Seyfert galaxies. For disky starburst galaxies, however, the trends suggest flows out of the stellar disk plane (with outflow 1/2-opening angle $> 45$ degree) instead of from the nucleus: the wind velocity decreases as the disk becomes more edge-on, and the outflow fraction, constant at $\sim$90$\%$ for disk inclinations $i<45$ degree, steadily decreases from $\sim$90$%$ to 20$\%$ for $i>45$ degree.

astro-ph.GA↗
arXiv2024

Deep-learning-based prediction of the tetragonal$\rightarrow$cubic transition in davemaoite

Davemaoite, i.e., $CaSiO_3$ perovskite (CaPv), is the third most abundant phase in the lower mantle and exhibits a tetragonal-cubic phase transition at high pressures and temperatures. The phase boundary in CaPv has recently been proposed to be close to the cold slab adiabatic and cause mid-mantle seismic wave speed anomalies (Thomson et al., Nature 572, 643, 2019). In this study, we utilized accurate deep-learning-based simulations and thermodynamic integration techniques to compute free energies at temperatures ranging from 300 to 3,000 K and pressures up to 130 GPa. Our results indicate that CaPv exhibits a single cubic phase throughout lower-mantle conditions. This suggests that the phase diagram proposed by Thomson et al. requires revision, and mid-mantle seismic anomalies are likely attributable to other mechanisms.

physics.geo-ph↗
arXiv2024

ACAV: A Framework for Automatic Causality Analysis in Autonomous Vehicle Accident Recordings

The rapid progress of autonomous vehicles~(AVs) has brought the prospect of a driverless future closer than ever. Recent fatalities, however, have emphasized the importance of safety validation through large-scale testing. Multiple approaches achieve this fully automatically using high-fidelity simulators, i.e., by generating diverse driving scenarios and evaluating autonomous driving systems~(ADSs) against different test oracles. While effective at finding violations, these approaches do not identify the decisions and actions that \emph{caused} them -- information that is critical for improving the safety of ADSs. To address this challenge, we propose ACAV, an automated framework designed to conduct causality analysis for AV accident recordings in two stages. First, we apply feature extraction schemas based on the messages exchanged between ADS modules, and use a weighted voting method to discard frames of the recording unrelated to the accident. Second, we use safety specifications to identify safety-critical frames and deduce causal events by applying CAT -- our causal analysis tool -- to a station-time graph. We evaluate ACAV on the Apollo ADS, finding that it can identify five distinct types of causal events in 93.64% of 110 accident recordings generated by an AV testing engine. We further evaluated ACAV on 1206 accident recordings collected from versions of Apollo injected with specific faults, finding that it can correctly identify causal events in 96.44% of the accidents triggered by prediction errors, and 85.73% of the accidents triggered by planning errors.

cs.SE↗
arXiv2024

REDriver: Runtime Enforcement for Autonomous Vehicles

Autonomous driving systems (ADSs) integrate sensing, perception, drive control, and several other critical tasks in autonomous vehicles, motivating research into techniques for assessing their safety. While there are several approaches for testing and analysing them in high-fidelity simulators, ADSs may still encounter additional critical scenarios beyond those covered once they are deployed on real roads. An additional level of confidence can be established by monitoring and enforcing critical properties when the ADS is running. Existing work, however, is only able to monitor simple safety properties (e.g., avoidance of collisions) and is limited to blunt enforcement mechanisms such as hitting the emergency brakes. In this work, we propose REDriver, a general and modular approach to runtime enforcement, in which users can specify a broad range of properties (e.g., national traffic laws) in a specification language based on signal temporal logic (STL). REDriver monitors the planned trajectory of the ADS based on a quantitative semantics of STL, and uses a gradient-driven algorithm to repair the trajectory when a violation of the specification is likely. We implemented REDriver for two versions of Apollo (i.e., a popular ADS), and subjected it to a benchmark of violations of Chinese traffic laws. The results show that REDriver significantly improves Apollo's conformance to the specification with minimal overhead.

cs.SE↗
arXiv2023

OVD-Explorer: Optimism Should Not Be the Sole Pursuit of Exploration in Noisy Environments

In reinforcement learning, the optimism in the face of uncertainty (OFU) is a mainstream principle for directing exploration towards less explored areas, characterized by higher uncertainty. However, in the presence of environmental stochasticity (noise), purely optimistic exploration may lead to excessive probing of high-noise areas, consequently impeding exploration efficiency. Hence, in exploring noisy environments, while optimism-driven exploration serves as a foundation, prudent attention to alleviating unnecessary over-exploration in high-noise areas becomes beneficial. In this work, we propose Optimistic Value Distribution Explorer (OVD-Explorer) to achieve a noise-aware optimistic exploration for continuous control. OVD-Explorer proposes a new measurement of the policy's exploration ability considering noise in optimistic perspectives, and leverages gradient ascent to drive exploration. Practically, OVD-Explorer can be easily integrated with continuous control RL algorithms. Extensive evaluations on the MuJoCo and GridChaos tasks demonstrate the superiority of OVD-Explorer in achieving noise-aware optimistic exploration.

cs.LG↗
arXiv2023

Learning with Noisy Labels for Human Fall Events Classification: Joint Cooperative Training with Trinity Networks

With the increasing ageing population, fall events classification has drawn much research attention. In the development of deep learning, the quality of data labels is crucial. Most of the datasets are labelled automatically or semi-automatically, and the samples may be mislabeled, which constrains the performance of Deep Neural Networks (DNNs). Recent research on noisy label learning confirms that neural networks first focus on the clean and simple instances and then follow the noisy and hard instances in the training stage. To address the learning with noisy label problem and protect the human subjects' privacy, we propose a simple but effective approach named Joint Cooperative training with Trinity Networks (JoCoT). To mitigate the privacy issue, human skeleton data are used. The robustness and performance of the noisy label learning framework is improved by using the two teacher modules and one student module in the proposed JoCoT. To mitigate the incorrect selections, the predictions from the teacher modules are applied with the consensus-based method to guide the student module training. The performance evaluation on the widely used UP-Fall dataset and comparison with the state-of-the-art, confirms the effectiveness of the proposed JoCoT in high noise rates. Precisely, JoCoT outperforms the state-of-the-art by 5.17% and 3.35% with the averaged pairflip and symmetric noises, respectively.

cs.CV↗
arXiv2023

MP-MVS: Multi-Scale Windows PatchMatch and Planar Prior Multi-View Stereo

Significant strides have been made in enhancing the accuracy of Multi-View Stereo (MVS)-based 3D reconstruction. However, untextured areas with unstable photometric consistency often remain incompletely reconstructed. In this paper, we propose a resilient and effective multi-view stereo approach (MP-MVS). We design a multi-scale windows PatchMatch (mPM) to obtain reliable depth of untextured areas. In contrast with other multi-scale approaches, which is faster and can be easily extended to PatchMatch-based MVS approaches. Subsequently, we improve the existing checkerboard sampling schemes by limiting our sampling to distant regions, which can effectively improve the efficiency of spatial propagation while mitigating outlier generation. Finally, we introduce and improve planar prior assisted PatchMatch of ACMP. Instead of relying on photometric consistency, we utilize geometric consistency information between multi-views to select reliable triangulated vertices. This strategy can obtain a more accurate planar prior model to rectify photometric consistency measurements. Our approach has been tested on the ETH3D High-res multi-view benchmark with several state-of-the-art approaches. The results demonstrate that our approach can reach the state-of-the-art. The associated codes will be accessible at https://github.com/RongxuanTan/MP-MVS.

cs.CV↗
arXiv2023

Unveiling the effect of Ni on the formation and structure of Earth's inner core

Ni is the second most abundant element in the Earth's core. Yet, its effects on the inner core's structure and formation process are usually disregarded because of its electronic and size similarity with Fe. Using ab initio molecular dynamics simulations, we find that the bcc phase can spontaneously crystallize in liquid Ni at temperatures above Fe's melting point at inner core pressures. The melting temperature of Ni is shown to be 700-800 K higher than that of Fe at 323-360 GPa. hcp, bcc, and liquid phase relation differ for Fe and Ni. Ni can be a bcc stabilizer for Fe at high temperatures and inner core pressures. A small amount of Ni can accelerate Fe's crystallization at core pressures. These results suggest Ni may substantially impact the structure and formation process of the solid inner core.

cond-mat.mtrl-sci↗
arXiv2023

Multi-Task Cooperative Learning via Searching for Flat Minima

Multi-task learning (MTL) has shown great potential in medical image analysis, improving the generalizability of the learned features and the performance in individual tasks. However, most of the work on MTL focuses on either architecture design or gradient manipulation, while in both scenarios, features are learned in a competitive manner. In this work, we propose to formulate MTL as a multi/bi-level optimization problem, and therefore force features to learn from each task in a cooperative approach. Specifically, we update the sub-model for each task alternatively taking advantage of the learned sub-models of the other tasks. To alleviate the negative transfer problem during the optimization, we search for flat minima for the current objective function with regard to features from other tasks. To demonstrate the effectiveness of the proposed approach, we validate our method on three publicly available datasets. The proposed method shows the advantage of cooperative learning, and yields promising results when compared with the state-of-the-art MTL approaches. The code will be available online.

cs.CV↗