Source author record

Yujia Sun

Yujia Sun appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Catalog footprint

What is connected

2works
2topics
4close collaborators

Actions

Connect this record

Log in to claim

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

2 published item(s)

preprint2021arXiv

s-d coupling enhanced phonon anharmonicity in copper-based compounds

Materials with ultralow thermal conductivity are of great interest for efficient energy conversion and thermal barrier coating. Copper-based semiconductors such as copper chalcogenides and copper halides are known to possess extreme low thermal conductivity, whereas the fundamental origin of the low thermal conductivity observed in the copper-based materials remains elusive. Here, we reveal that s-d coupling induced giant phonon anharmonicity is the fundamental mechanism responsible for the ultralow thermal conductivity of copper compounds. The symmetry controlled strong coupling of high-lying occupied copper 3d orbital with the unoccupied 4s state under thermal vibration remarkably lowers the lattice potential barrier, which enhances anharmonic scattering between phonons. This understanding is confirmed by temperature-dependent Raman spectra measurements. Our study offers an insight at atomic level connecting electronic structures with phonon vibration modes, and thus sheds light on materials properties that rely on electron-phonon coupling, such as thermoelectricity and superconductivity.

preprint2020arXiv

Not only Look, but also Listen: Learning Multimodal Violence Detection under Weak Supervision

Violence detection has been studied in computer vision for years. However, previous work are either superficial, e.g., classification of short-clips, and the single scenario, or undersupplied, e.g., the single modality, and hand-crafted features based multimodality. To address this problem, in this work we first release a large-scale and multi-scene dataset named XD-Violence with a total duration of 217 hours, containing 4754 untrimmed videos with audio signals and weak labels. Then we propose a neural network containing three parallel branches to capture different relations among video snippets and integrate features, where holistic branch captures long-range dependencies using similarity prior, localized branch captures local positional relation using proximity prior, and score branch dynamically captures the closeness of predicted score. Besides, our method also includes an approximator to meet the needs of online detection. Our method outperforms other state-of-the-art methods on our released dataset and other existing benchmark. Moreover, extensive experimental results also show the positive effect of multimodal (audio-visual) input and modeling relationships. The code and dataset will be released in https://roc-ng.github.io/XD-Violence/.