| 6 | 1/1 | 返回列表 |
| 查看: 768 | 回復(fù): 5 | |||
nilhzoul新蟲 (初入文壇)
|
[交流]
南方科技大學(xué)-新加坡國立大學(xué)聯(lián)合研究項目招聘博士后1名
|
|
The project is hiring 1 postdoctoral fellows at SUSTech for Image Processing and Deep Learning). 南方科技大學(xué)課題組現(xiàn)公開招聘博士后1名(圖像處理和深度學(xué)習(xí)方向). Title/項目名稱 Tensor Dimension Reduction for Multimodal Data and its Application in EEG Analysis(多模態(tài)張量降維方法研究和在EEG數(shù)據(jù)分析中的應(yīng)用) Abstract/項目簡介 Multimodal data analysis, driven by the increasing variety of data types in applications, has attracted much interest in statistics, focusing on encoding different modalities into a common representation space to build predictive models or explore relationships between modalities. The main goal of the project is to build models to describe complicated dependencies between modalities. This is particularly challenging when at least one modality is a tensor due to the complexity of tensor structures. Our idea is to introduce a novel Structural Equation Model (SEM) based tensor matrix factorisation model to simultaneously extract latent scores and build relationships between modalities. The context of the proposal is the study of insomnia patients, where EEG data and questionnaire responses are analysed to understand the relationship between these very different types of data. The project outlines four specific problems, including the development of new tensor matrix factorisation models, the investigation of identifiability conditions, the measurement of non-linear dependence in tensor decomposition, and the study of functional connectivity models in the context of tensor matrix factorisation. statistical inference, implementation tools, optimisation algorithms and theory of both statistical and optimisation errors will be studied. Applications to the analysis of other wearable and mobile data will also be explored. The project will include but is not limited to the following four topics (1) SEM-based tensor matrix factorization model; (2) Nonlinear tensor matrix factorisation models and identifiability; (3) Tensor Decomposition with a measure of nonlinear dependence; (4) Functional connectivity analysis based on factorisation model. 由于在實際應(yīng)用中的數(shù)據(jù)類型日益多樣化,多模態(tài)數(shù)據(jù)分析引起了統(tǒng)計學(xué)界的極大興趣。其核心在于將不同模態(tài)的數(shù)據(jù)編碼到一個共同的空間中,以建立預(yù)測模型或探索各模態(tài)之間的關(guān)系。該項目的主要目標(biāo)是建立模型來描述模態(tài)之間潛在的依賴關(guān)系。由于涉及張量結(jié)構(gòu)的復(fù)雜性,當(dāng)至少一種模態(tài)為張量時,這項工作尤其具有挑戰(zhàn)性。我們的思路是引入一種新穎的基于結(jié)構(gòu)方程模型的張量矩陣因式分解模型,以同時提取潛在變量并建立模態(tài)之間的關(guān)系。該項目的背景是針對失眠癥患者的研究,通過分析腦電圖數(shù)據(jù)和問卷答復(fù)來了解這些截然不同數(shù)據(jù)類型之間的關(guān)系。該項目包含了四個具體問題,包括開發(fā)新的張量矩陣因式分解模型、研究可識別性條件、測量張量分解中的非線性依賴性以及在張量矩陣因式分解背景下研究功能連接模型。該項目研究范圍涵蓋了統(tǒng)計推斷、實現(xiàn)工具、優(yōu)化算法以及統(tǒng)計和優(yōu)化誤差理論。此外,還將探討該方法在其他可穿戴和移動數(shù)據(jù)分析方面的應(yīng)用潛力。該項目將包括但不限于以下四個主題:(1)基于結(jié)構(gòu)方程模型的張量矩陣因式分解模型;(2)非線性張量矩陣因式分解模型與可識別性;(3)考慮非線性依賴的張量分解;(4)基于因式分解模型的功能連接分析。 |
新蟲 (初入文壇)
|
PIs at SUSTech Dr. Chao Wang (王超博士) Dr. Wang is an Assistant Professor of the Department of Statistics and Data Science at Southern University of Science and Technology. His research directions are mainly image processing, scientific computing, and interdisciplinary data science. He has published over 30 papers in top-tier journals such as the SIAM series and IEEE Transactions, as well as leading conferences, and has received notable accolades including the Best Paper Award at the 2022 CVPR Workshop, the Shenzhen Pengcheng Peacock Plan Distinguished Professorship (2021), and the Best Paper Award at the 2017 Annual Meeting of China Society for Industrial and Applied Mathematics. He has led research projects such as the National Natural Science Foundation of China Youth Program, Guangdong Basic and Applied Research Foundation Program, and Shenzhen Science and Technology Program, while also contributing as a principal investigator or core member to major initiatives including the National Key Research and Development Program, Hong Kong RGC Research Fund projects, and Shenzhen Fundamental Research Program.王超,南方科技大學(xué)統(tǒng)計與數(shù)據(jù)科學(xué)系副研究員,博導(dǎo),其研究方向主要為圖像處理、科學(xué)計算與交叉學(xué)科的數(shù)據(jù)科學(xué)。在本領(lǐng)域期刊SIAM系列、IEEE匯刊等雜志及學(xué)術(shù)會議發(fā)表學(xué)術(shù)論文三十余篇。在2022年CVPR研討會獲得最佳論文,在2021年獲深圳市鵬城孔雀計劃特聘崗位,在2017年獲得中國工業(yè)與應(yīng)用數(shù)學(xué)學(xué)會年會最佳論文。主持國自然青年基金、廣東省面上基金以及深圳市穩(wěn)定支持面上項目,以課題負(fù)責(zé)人或核心成員參與國家重點研發(fā)項目、香港研資局科研基金項目以及深圳重點項目。 王超副研究員個人網(wǎng)頁https://wangcmath.github.io/ |
新蟲 (初入文壇)
新蟲 (初入文壇)
新蟲 (初入文壇)
新蟲 (初入文壇)
| 6 | 1/1 | 返回列表 |
| 最具人氣熱帖推薦 [查看全部] | 作者 | 回/看 | 最后發(fā)表 | |
|---|---|---|---|---|
|
[考研] 環(huán)境調(diào)劑 +8 | chenhanheng 2026-03-02 | 8/400 |
|
|---|---|---|---|---|
|
[考研] 266求調(diào)劑 +4 | 哇塞王帥 2026-03-03 | 4/200 |
|
|
[考研] 264求調(diào)劑 +6 | 26調(diào)劑 2026-03-03 | 6/300 |
|
|
[考研] 290求調(diào)劑 +9 | ErMiao1020 2026-03-02 | 9/450 |
|
|
[考研] 085600材料與化工 298 調(diào)劑 +3 | 小西笑嘻嘻 2026-03-03 | 3/150 |
|
|
[基金申請] 沒有青基直接申請面上,感覺自己瘋了 +5 | kevin63t 2026-03-02 | 6/300 |
|
|
[考研] 289求調(diào)劑 +7 | BrightLL 2026-03-02 | 9/450 |
|
|
[考研]
材料270求調(diào)劑
6+6
|
Eiiiio 2026-03-01 | 11/550 |
|
|
[考研] 材料學(xué)碩318求調(diào)劑 +11 | February_Feb 2026-03-01 | 11/550 |
|
|
[考研]
材料工程專碩283求調(diào)劑
5+8
|
,!? 2026-03-02 | 10/500 |
|
|
[考研] 課題組接收材料類調(diào)劑研究生 +6 | gaoxiaoniuma 2026-02-28 | 9/450 |
|
|
[考研] 0856材料與化工,270求調(diào)劑 +11 | YXCT 2026-03-01 | 13/650 |
|
|
[考研] 0856材料調(diào)劑 +5 | 沿岸有貝殼OUC 2026-03-02 | 5/250 |
|
|
[考研] 化學(xué),材料,環(huán)境類求調(diào)劑 +7 | 考研版棒棒 2026-03-02 | 7/350 |
|
|
[考研] 材料085601調(diào)劑 +5 | 多多子. 2026-03-02 | 5/250 |
|
|
[考研] 一志愿華南理工大學(xué)材料與化工326分,求調(diào)劑 +3 | wujinrui1 2026-02-28 | 3/150 |
|
|
[考研] 292求調(diào)劑 +7 | yhk_819 2026-02-28 | 7/350 |
|
|
[考研] 0856材料求調(diào)劑 +4 | 麻辣魷魚 2026-02-28 | 4/200 |
|
|
[考研] 313求調(diào)劑 +3 | 水流年lc 2026-02-28 | 3/150 |
|
|
[考研] 311求調(diào)劑 +6 | 亭亭亭01 2026-03-01 | 6/300 |
|