Pith.

What the papers on Pith cite · a cited work

Implicitly defined layers in neural networks

2020 · cs.CV

2 papers on Pith cite it

The work's abstract

In conventional formulations of multilayer feedforward neural networks, the individual layers are customarily defined by explicit functions. In this paper we demonstrate that defining individual layers in a neural network \emph{implicitly} provide much richer representations over the standard explicit one, consequently enabling a vastly broader class of end-to-end trainable architectures. We present a general framework of implicitly defined layers, where much of the theoretical analysis of such layers can be addressed through the implicit function theorem. We also show how implicitly defined layers can be seamlessly incorporated into existing machine learning libraries. In particular with respect to current automatic differentiation techniques for use in backpropagation based training. Finally, we demonstrate the versatility and relevance of our proposed approach on a number of diverse example problems with promising results.

How it is cited

As: method 1

With: use method 1

Counted by a program from the sentences in which the citing papers name this work.

Papers on Pith that cite it

Narrow the list

Field

Year submitted

How it is cited

2 of 2 citing papers. Ordered by the review's novelty score weighted by whether the review supports the claim, highest first; papers with no score come last. Among equal scores, newer papers first. The order is not a quality score. How a review's support is weighed

  1. The paper:

    JFB: Jacobian-Free Backpropagation for Implicit Networks

    Computer Science · arXiv:2103.12803

    Samy Wu Fung, Howard Heaton, Qiuwei Li and 3 more

    Pith says

    Machine review · deepseek-v4.1-flash · 13 Nov 2026 · what was read was not recorded

    Holds up, with conditions. A practical, honestly-scoped training trick for implicit networks whose descent theorem probably doesn't certify its own experiments.

    4 major objections, the first at §3, Assumption 3.3 and Theorem 3.1 · 9 minor

    Cites this work as reference 68 of its list.

  2. Cites this work as reference 21 of its list.