skip to main content


Search for: All records

Creators/Authors contains: "Sun, S."

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. null (Ed.)
    Clustering is a machine learning paradigm of dividing sample subjects into a number of groups such that subjects in the same groups are more similar to those in other groups. With advances in information acquisition technologies, samples can frequently be viewed from different angles or in different modalities, generating multi-view data. Multi-view clustering, that clusters subjects into subgroups using multi-view data, has attracted more and more attentions. Although MVC methods have been developed rapidly, there has not been enough survey to summarize and analyze the current progress. Therefore, we propose a novel taxonomy of the MVC approaches. Similar with machine learning methods, we categorize them into generative and discriminative classes. In discriminative class, based on the way to integrate multiple views, we split it further into five groups: Common Eigenvector Matrix, Common Coefficient Matrix, Common Indicator Matrix, Direct Combination and Combination After Projection. Furthermore, we discuss the relationships between MVC and some related topics: multi-view representation, ensemble clustering, multi-task clustering, multi-view supervised and semi-supervised learning. Several representative real-world applications are elaborated for practitioners. Some commonly used multi-view datasets are introduced and several representative MVC algorithms from each group are run to conduct the comparison to analyze how and why they perform on those datasets. To promote future development of MVC approaches, we point out several open problems that may require further investigation and thorough examination. 
    more » « less
  2. null (Ed.)
  3. Abstract

    Sea level rise (SLR) is a long‐lasting consequence of climate change because global anthropogenic warming takes centuries to millennia to equilibrate for the deep ocean and ice sheets. SLR projections based on climate models support policy analysis, risk assessment and adaptation planning today, despite their large uncertainties. The central range of the SLR distribution is estimated by process‐based models. However, risk‐averse practitioners often require information about plausible future conditions that lie in the tails of the SLR distribution, which are poorly defined by existing models. Here, a community effort combining scientists and practitioners builds on a framework of discussing physical evidence to quantify high‐end global SLR for practitioners. The approach is complementary to the IPCC AR6 report and provides further physically plausible high‐end scenarios. High‐end estimates for the different SLR components are developed for two climate scenarios at two timescales. For global warming of +2°C in 2100 (RCP2.6/SSP1‐2.6) relative to pre‐industrial values our high‐end global SLR estimates are up to 0.9 m in 2100 and 2.5 m in 2300. Similarly, for a (RCP8.5/SSP5‐8.5), we estimate up to 1.6 m in 2100 and up to 10.4 m in 2300. The large and growing differences between the scenarios beyond 2100 emphasize the long‐term benefits of mitigation. However, even a modest 2°C warming may cause multi‐meter SLR on centennial time scales with profound consequences for coastal areas. Earlier high‐end assessments focused on instability mechanisms in Antarctica, while here we emphasize the importance of the timing of ice shelf collapse around Antarctica. This is highly uncertain due to low understanding of the driving processes. Hence both process understanding and emission scenario control high‐end SLR.

     
    more » « less
  4. Free, publicly-accessible full text available August 1, 2024
  5. Free, publicly-accessible full text available July 1, 2024