arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2301.00801stat.MLcs.LG
收藏

因果推断(C-inf)--典型相变的非对称情形

Causal Inference (C-inf) -- asymmetric scenario of typical phase transitions

Columbia University(哥伦比亚大学)

Agostino Capponi, Mihailo Stojnic

摘要:

在本文中,我们重新审视并进一步探讨了因果推断(C-inf)和低秩恢复(LRR)之间的数学严格联系。利用随机二元性-自由概率论(RDT-FPT)的联系,我们获得了精确显式的典型C-inf非对称相变(PT)。我们发现了一个加倍的低秩现象,这意味着与[10]中考虑的对称最坏情况相比,在非对称场景中允许的低秩正好是其两倍。因此,最终PT数学表达式与[10]中获得的数学表达式一样简洁,并突出了目标C-inf矩阵低秩与治疗时间之间的直接关系。我们的结果有很强的应用,其中C-inf矩阵不一定是对称的。

英文摘要:

In this paper, we revisit and further explore a mathematically rigorous connection between Causal inference (C-inf) and the Low-rank recovery (LRR) established in [10]. Leveraging the Random duality - Free probability theory (RDT-FPT) connection, we obtain the exact explicit typical C-inf asymmetric phase transitions (PT). We uncover a doubling low-rankness phenomenon, which means that exactly two times larger low rankness is allowed in asymmetric scenarios compared to the symmetric worst case ones considered in [10]. Consequently, the final PT mathematical expressions are as elegant as those obtained in [10], and highlight direct relations between the targeted C-inf matrix low rankness and the time of treatment. Our results have strong implications for applications, where C-inf matrices are not necessarily symmetric.

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。

↑