A variational derivation of a class of BFGS-like methods
arXiv:1712.006802 citationsh-index: 30
Originality Incremental advance
AI Analysis
Provides a novel theoretical foundation for quasi-Newton methods, relevant to optimization researchers.
The authors derive a new family of BFGS-like methods using a maximum entropy approach, extending the derivation to block BFGS methods and generalizing a result by Fletcher (1991).
We provide a maximum entropy derivation of a new family of BFGS-like methods. Similar results are then derived for block BFGS methods. This also yields an independent proof of a result of Fletcher 1991 and its generalisation to the block case.