WiMi Hologram Cloud Inc (WIMI)

[StockpickerCAPC]

WIMI Hologram Cloud Studies A Multi-View Data Fusion Algorithm.

Source
https://www.newstrail.com/wimi-hologram-cloud-studies-a-multi-view-data-fusion-algorithm/

February 13, 2023

With the extensive penetration and deepening of internet technology and applications in our daily life. Thus, people have an increasing demand for better convenience and presentation modes of media. In the Metaverse Era, multi-mode technology, or multi-view data technology can find a suitable way for data integration from multi-data resources and different features and perspectives of objects.

In the Big Data Era, all walks of life have produced large amounts of data. When solving a problem, we usually need to take advantage of multiple different datasets. Different domains produce multiple datasets that are implicitly connected by latent objects. Furthermore, different datasets of the same object and subsets of different features can also be viewed as different views of an object. For example, the information after training from different data sources can identify some of a person’s information, more than fingerprint, face information or signature, etc. A graph can also be represented through different sets of features (such as colors, shapes, etc.), as the datasets describe the same object. Then, these data sets are different from each other, contain unique information, and complement each other. Therefore, integrating different views and extracting different features can describe an object more accurately and comprehensively.

It is understood that the WIMI Hologram Cloud (NASDAQ: WIMI) development team is working on a multi-view-based data fusion algorithm, whose core is a multi-view learning algorithm. The multi-view learning algorithm includes three parts:
- co-training,
- multi-core learning, and
- subspace learning.
The co-training algorithm considers that each sample can be divided into different views, selectively maximizing the likelihood that both parties can agree on the data in different views. Multi-core learning involves a range of machine learning methods, using natural cores that respond to different views, and combining linear and non-linear features to improve learning outcomes. Subspace learning is designed to obtain a subspace that can be shared by different views. With this subspace, it can perform subsequent tasks, such as classification and clustering.

With the continuous improvement of computer level and the continuous progress of data fusion algorithms, the application of data fusion is also constantly promoted. In this era of big data with the information explosion, data has become a strategic asset. Where there is information and data, there will be a demand for data fusion. In the new era of national informatization development strategy, wisdom city, intelligent transportation, and intelligent manufacturing in various fields of digital construction are in steady progress, the construction in the field of digital will have more demand for data fusion algorithms, thus, WIMI based on multi-view data fusion algorithm has extremely broad application prospect.