-
A Principled Approach to Failure Analysis and Model Repairment: Demonstration in Medical Imaging
Authors:
Thomas Henn,
Yasukazu Sakamoto,
Clément Jacquet,
Shunsuke Yoshizawa,
Masamichi Andou,
Stephen Tchen,
Ryosuke Saga,
Hiroyuki Ishihara,
Katsuhiko Shimizu,
Yingzhen Li,
Ryutaro Tanno
Abstract:
Machine learning models commonly exhibit unexpected failures post-deployment due to either data shifts or uncommon situations in the training environment. Domain experts typically go through the tedious process of inspecting the failure cases manually, identifying failure modes and then attempting to fix the model. In this work, we aim to standardise and bring principles to this process through an…
▽ More
Machine learning models commonly exhibit unexpected failures post-deployment due to either data shifts or uncommon situations in the training environment. Domain experts typically go through the tedious process of inspecting the failure cases manually, identifying failure modes and then attempting to fix the model. In this work, we aim to standardise and bring principles to this process through answering two critical questions: (i) how do we know that we have identified meaningful and distinct failure types?; (ii) how can we validate that a model has, indeed, been repaired? We suggest that the quality of the identified failure types can be validated through measuring the intra- and inter-type generalisation after fine-tuning and introduce metrics to compare different subtyping methods. Furthermore, we argue that a model can be considered repaired if it achieves high accuracy on the failure types while retaining performance on the previously correct data. We combine these two ideas into a principled framework for evaluating the quality of both the identified failure subtypes and model repairment. We evaluate its utility on a classification and an object detection tasks. Our code is available at https://github.com/Rokken-lab6/Failure-Analysis-and-Model-Repairment
△ Less
Submitted 25 September, 2021;
originally announced September 2021.
-
Reflexive combinatory algebras
Authors:
Marlou M. Gijzen,
Hajime Ishihara,
Tatsuji Kawai
Abstract:
We introduce the notion of reflexivity for combinatory algebras. Reflexivity can be thought of as an equational counterpart of the Meyer-Scott axiom of combinatory models, which indeed allows us to characterise an equationally definable counterpart of combinatory models. This new structure, called strongly reflexive combinatory algebra, admits a finite axiomatisation with seven closed equations, a…
▽ More
We introduce the notion of reflexivity for combinatory algebras. Reflexivity can be thought of as an equational counterpart of the Meyer-Scott axiom of combinatory models, which indeed allows us to characterise an equationally definable counterpart of combinatory models. This new structure, called strongly reflexive combinatory algebra, admits a finite axiomatisation with seven closed equations, and the structure is shown to be exactly the retract of combinatory models. Lambda algebras can be characterised as strongly reflexive combinatory algebras which are stable. Moreover, there is a canonical construction of a lambda algebra from a strongly reflexive combinatory algebra. The resulting axiomatisation of lambda algebras by the seven axioms for strong reflexivity together with those for stability is shown to correspond to the axiomatisation of lambda algebras due to Selinger [J.Funct.Programming, 12(6), 549--566, 2002].
△ Less
Submitted 30 June, 2022; v1 submitted 23 April, 2021;
originally announced April 2021.
-
Classical propositional logic and decidability of variables in intuitionistic propositional logic
Authors:
Hajime Ishihara
Abstract:
We improve the answer to the question: what set of excluded middles for propositional variables in a formula suffices to prove the formula in intuitionistic propositional logic whenever it is provable in classical propositional logic.
We improve the answer to the question: what set of excluded middles for propositional variables in a formula suffices to prove the formula in intuitionistic propositional logic whenever it is provable in classical propositional logic.
△ Less
Submitted 6 August, 2014; v1 submitted 1 May, 2014;
originally announced May 2014.
