Learning $\textit{Ex Nihilo}$
It proposes a philosophical challenge for ML researchers to collaborate on integrating logic-based reasoning with statistical methods, but is incremental as it builds on existing formal logic concepts.
The paper introduces the concept of learning ex nihilo, where an agent learns 'from nothing' using deductive and inductive reasoning based on a cognitive calculus, and challenges contemporary machine learning to integrate with such frameworks.
This paper introduces, philosophically and to a degree formally, the novel concept of learning $\textit{ex nihilo}$, intended (obviously) to be analogous to the concept of creation $\textit{ex nihilo}$. Learning $\textit{ex nihilo}$ is an agent's learning "from nothing," by the suitable employment of schemata for deductive and inductive reasoning. This reasoning must be in machine-verifiable accord with a formal proof/argument theory in a $\textit{cognitive calculus}$ (i.e., roughly, an intensional higher-order multi-operator quantified logic), and this reasoning is applied to percepts received by the agent, in the context of both some prior knowledge, and some prior and current interests. Learning $\textit{ex nihilo}$ is a challenge to contemporary forms of ML, indeed a severe one, but the challenge is offered in the spirt of seeking to stimulate attempts, on the part of non-logicist ML researchers and engineers, to collaborate with those in possession of learning-$\textit{ex nihilo}$ frameworks, and eventually attempts to integrate directly with such frameworks at the implementation level. Such integration will require, among other things, the symbiotic interoperation of state-of-the-art automated reasoners and high-expressivity planners, with statistical/connectionist ML technology.