Spurious samples in deep generative models: bug or feature?
Authors:
Balázs Kégl,
Mehdi Cherti,
Akın Kazakçı
Abstract:
Traditional wisdom in generative modeling literature is that spurious samples that a model can generate are errors and they should be avoided. Recent research, however, has shown interest in studying or even exploiting such samples instead of eliminating them. In this paper, we ask the question whether such samples can be eliminated all together without sacrificing coverage of the generating distr…
▽ More
Traditional wisdom in generative modeling literature is that spurious samples that a model can generate are errors and they should be avoided. Recent research, however, has shown interest in studying or even exploiting such samples instead of eliminating them. In this paper, we ask the question whether such samples can be eliminated all together without sacrificing coverage of the generating distribution. For the class of models we consider, we experimentally demonstrate that this is not possible without losing the ability to model some of the test samples. While our results need to be confirmed on a broader set of model families, these initial findings provide partial evidence that spurious samples share structural properties with the learned dataset, which, in turn, suggests they are not simply errors but a feature of deep generative nets.
△ Less
Submitted 3 October, 2018;
originally announced October 2018.
Conceptive Artificial Intelligence: Insights from design theory
Authors:
Akin Osman Kazakci
Abstract:
The current paper offers a perspective on what we term conceptive intelligence - the capacity of an agent to continuously think of new object definitions (tasks, problems, physical systems, etc.) and to look for methods to realize them. The framework, called a Brouwer machine, is inspired by previous research in design theory and modeling, with its roots in the constructivist mathematics of intuit…
▽ More
The current paper offers a perspective on what we term conceptive intelligence - the capacity of an agent to continuously think of new object definitions (tasks, problems, physical systems, etc.) and to look for methods to realize them. The framework, called a Brouwer machine, is inspired by previous research in design theory and modeling, with its roots in the constructivist mathematics of intuitionism. The dual constructivist perspective we describe offers the possibility to create novelty both in terms of the types of objects and the methods for constructing objects. More generally, the theoretical work on which Brouwer machines are based is called imaginative constructivism. Based on the framework and the theory, we discuss many paradigms and techniques omnipresent in AI research and their merits and shortcomings for modeling aspects of design, as described by imaginative constructivism. To demonstrate and explain the type of creative process expressed by the notion of a Brouwer machine, we compare this concept with a system using genetic algorithms for scientific law discovery.
△ Less
Submitted 2 April, 2014;
originally announced April 2014.
