arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2405.01974cs.LGcs.AIq-bio.QM
收藏

几何对齐迁移编码器的多任务扩展

Multitask Extension of Geometrically Aligned Transfer Encoder

LG AI Research(LG人工智能研究院)

Sung Moon Ko, Sumin Lee, Dae-Woong Jeong, Hyunseung Kim, Chanhui Lee, Soorin Yim, Sehui Han

AI总结:

本文提出将几何对齐迁移编码器扩展为多任务框架,通过坐标对齐在不同分子任务间传递互信息,以缓解分子数据稀缺问题并提升目标任务性能。

摘要:

分子数据集常常面临数据不足的问题。众所周知,由于所涉及的实验或模拟十分复杂,收集数据非常困难。在此,我们利用分子数据中不同任务之间的互信息来解决这一问题。我们将一种利用编码空间几何特征的算法——即几何对齐迁移编码器(GATE)——扩展到多任务设置中。因此,我们通过将弯曲坐标对齐到局部平坦坐标上来连接多个分子任务,确保信息从源任务流向目标任务,以支持目标数据上的性能。

英文摘要:

Molecular datasets often suffer from a lack of data. It is well-known that gathering data is difficult due to the complexity of experimentation or simulation involved. Here, we leverage mutual information across different tasks in molecular data to address this issue. We extend an algorithm that utilizes the geometric characteristics of the encoding space, known as the Geometrically Aligned Transfer Encoder (GATE), to a multi-task setup. Thus, we connect multiple molecular tasks by aligning the curved coordinates onto locally flat coordinates, ensuring the flow of information from source tasks to support performance on target data.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑