How `Nonvariational' Are Approximate Coupled Cluster Methods In Practice?
Jan M. L. Martin, Emmanouil Semidalas

TL;DR
This study evaluates the practical nonvariational nature of approximate coupled cluster methods by comparing their energies to full CI results across various molecules, highlighting the accuracy of certain iterative and Lambda approaches.
Contribution
It provides a detailed comparison of limited coupled cluster methods with full CI energies, clarifying their variational properties in practice.
Findings
Iterative CC methods like CCSDT and CCSDTQ closely bound FCI energies.
Quasiperturbative methods like CCSD(T) can over-correlate in static correlation cases.
Lambda approaches such as CCSDT(Q)Λ mitigate over-correlation issues.
Abstract
While limited coupled cluster theory is \textit{formally} nonvariational, it is not broadly appreciated whether this is a major issue \textit{in practice}. We carried out a detailed comparison with \textit{de facto} full CI energies for a relatively large and diverse set of molecules. Fully iterative limited CC methods such as CCSDT, CCSDTQ, CCSDTQ5 do represent practical upper bounds to the FCI energy. While quasiperturbative approaches such as CCSD(T) and especially CCSDT(Q) may significantly over-correlate molecules if there is significant static correlation, this is much less of an issue with Lambda approaches such as CCSDT(Q).
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Clustering Algorithms Research · Bayesian Methods and Mixture Models · Face and Expression Recognition
