Playgol: Learning Programs Through Play

Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2019/841 ◽

2019 ◽

Author(s):

Andrew Cropper

Keyword(s):

Logic Programming ◽

Real World ◽

Inductive Logic Programming ◽

Inductive Logic ◽

Experimental Results ◽

Learning Performance ◽

Programming System ◽

Sample Complexity ◽

Induction System ◽

Learning Programs

Children learn though play. We introduce the analogous idea of learning programs through play. In this approach, a program induction system (the learner) is given a set of user-supplied build tasks and initial background knowledge (BK). Before solving the build tasks, the learner enters an unsupervised playing stage where it creates its own play tasks to solve, tries to solve them, and saves any solutions (programs) to the BK. After the playing stage is finished, the learner enters the supervised building stage where it tries to solve the build tasks and can reuse solutions learnt whilst playing. The idea is that playing allows the learner to discover reusable general programs on its own which can then help solve the build tasks. We claim that playing can improve learning performance. We show that playing can reduce the textual complexity of target concepts which in turn reduces the sample complexity of a learner. We implement our idea in Playgol, a new inductive logic programming system. We experimentally test our claim on two domains: robot planning and real-world string transformations. Our experimental results suggest that playing can substantially improve learning performance.

Download Full-text

Forgetting to Learn Logic Programs

Proceedings of the AAAI Conference on Artificial Intelligence ◽

10.1609/aaai.v34i04.5776 ◽

2020 ◽

Vol 34 (04) ◽

pp. 3676-3683

Author(s):

Andrew Cropper

Keyword(s):

Logic Programming ◽

Inductive Logic Programming ◽

Inductive Logic ◽

Background Knowledge ◽

Experimental Results ◽

Learning Performance ◽

Sample Complexity ◽

Logic Programs ◽

Hypothesis Space ◽

Alternative Approaches

Most program induction approaches require predefined, often hand-engineered, background knowledge (BK). To overcome this limitation, we explore methods to automatically acquire BK through multi-task learning. In this approach, a learner adds learned programs to its BK so that they can be reused to help learn other programs. To improve learning performance, we explore the idea of forgetting, where a learner can additionally remove programs from its BK. We consider forgetting in an inductive logic programming (ILP) setting. We show that forgetting can significantly reduce both the size of the hypothesis space and the sample complexity of an ILP learner. We introduce Forgetgol, a multi-task ILP learner which supports forgetting. We experimentally compare Forgetgol against approaches that either remember or forget everything. Our experimental results show that Forgetgol outperforms the alternative approaches when learning from over 10,000 tasks.

Download Full-text

Apperception

10.1093/oso/9780198862536.003.0011 ◽

2021 ◽

pp. 218-238

Author(s):

Richard Evans

Keyword(s):

Neural Network ◽

Logic Programming ◽

Inductive Logic Programming ◽

Inductive Logic ◽

Sensory Information ◽

Programming System ◽

Change Over Time ◽

The Neural Network ◽

Binary Neural Network ◽

Over Time

This paper describes a neuro-symbolic system for distilling interpretable logical theories out of streams of raw, unprocessed sensory experience. We combine a binary neural network, that maps raw sensory input to concepts, with an inductive logic programming system, that combines concepts into declarative rules. Both the inductive logic programming system and the binary neural network are encoded as logic programs, so the weights of the neural network and the declarative rules of the theory can be solved jointly as a single SAT problem. This way, we are able to jointly learn how to perceive (mapping raw sensory information to concepts) and apperceive (combining concepts into declarative rules). We apply our system, the Apperception Engine, to the Sokoban domain. Given a sequence of noisy pixel images, the system has to construct objects that persist over time, extract attributes that change over time, and induce rules explaining how the attributes change over time. We compare our system with a neural network baseline, and show that the baseline is significantly outperformed by the Apperception Engine.

Download Full-text