Wangd-wiki-article-2022-memorization

来自cslt Wiki
2023年12月28日 (四) 05:43Cslt讨论 | 贡献的版本

(差异) ←上一版本 | 最后版本 (差异) | 下一版本→ (差异)
跳转至: 导航搜索


Nobody wants to experience 2022 one time again, I suppose. Although too many things we need to memorize, most of them can be fairly judged by the history, at least after 10 years.



We surely will memorize the pandemic of Omicrown, the expected spreading and unexpected way of ending, the painful lockdown and inevitable infection. Fortunately, we all live; but unfortunately, some people lose their relatives and friends.



We surely will memorize the war in Ukraine; although it is far away in another continent, we still feel sorrow for the people, just because we are all human. This is nothing to do with race, religion and politics.



We witnessed the continuous progress in both science and technology, and people are benefiting more from it. Especially in the field of AI, breakthrough has emerged by combining AI with diverse scientific areas, such as medicine, astronomy, material and math.



We are pleased to get two interesting ideas in our research, one is a cycle consistency loss which shows a new way of information factorization, and another is homomorphy processing for superpositional signals, which we are working on with passion.



We also completed a new book named 《AI explained by graph》, which contributes to AI education for students in primary and middle schools. Many colleagues were involved in the planning, writing, editing and proof reading.



We are approaching to a new year. It is hard to predict how it will look like, hopefully better than 2022 at least.