论文修改 (JiangXY 20211020 表述问题)

1. realizes -> implements, achieves
2. between source domains and target domains -> between source and target domains
   Domain adaptation achieves effective transfer between source and target domains with different distributions.
3. How to accurately identify unknown categories of the target domain to achieve domain adaptation in open scenarios is an interesting issue.
   The identification of unknown categories in the target domain for domain adaptation in open scenarios is an interesting issue.
   Identifying unknown categories in the target domain for adaptation in open scenarios is an interesting issue.
   Recognizing unknown categories for domain adaptation in open scenarios is an interesting problem.
   Identifying unknown categories through domain adaptation in open scenarios is an interesting issue. *
4. distributed-driven -> distribution driven
5. to address this interesting issue -> 删除
6. we are better than -> DATL is better than
7. we still have better performance in terms of openness.
   ->
   The accuracy of DATL is good even under very high openness.
   The accuracy of DATL decreases slowly with the increase of openness.
8. Domain adaptation trains a classifier from a domain with rich labels (source domain) to apply to a domain with scarce labels (target domain) (Pan et al. 2011).
   Domain adaptation refers to training a classifier from a domain with rich labels (source domain) to apply to a domain with scarce labels (target domain) (Pan et al. 2011).
9. It successfully solved the challenge
   ->
   This technique successfully handles the challenge
10. tackle these complex target domains. -> tackle these devations.
11. Paul et al.(Busto and Gall 2017) -> citet vs. citep
12. Since then, various improved algorithms have been proposed, such as based on adversarial learning (Saito et al. 2018b), based on self-supervision learning (Bucci, Loghmani, and Tommasi 2020) and based on unsupervised learning (Kundu et al. 2020).
    ->
    Since then, various improved algorithms have been proposed, including those based on adversarial learning (Saito et al. 2018b), self-supervision learning (Bucci, Loghmani, and Tommasi 2020), and unsupervised learning (Kundu et al. 2020).
13. learning algorithm (DATL) -> learning (DATL) algorithm
14. We compare our algorithm -> We compare the DATL algorithm
    -> Our algorithm is compared with …
15. Figure (a) shows a typical -> (a) shows a typical 图内部
16. \usepackage{enumerate}
    \begin{enumerate}[1)]
17. y_i 格式不一致
18. Y = (y_1, y_2, \dots, y_n)
    https://blog.csdn.net/minfanphd/article/details/118859971
19. f(x): \mathcal{R}^m \to Y 常用
    x \mapsto x^2