Improving scalability of inductive logic programming via pruning and best-effort optimisation

Inductive Logic Programming: Issues, results and the challenge of Learning Language in Logic

Generative policies have been proposed as a mechanism to learn the constraints and preferences of a system—especially complex systems such as the ones found in coalitions—in a given context so that the system can adapt to unexpected changes seamlessly, thus achieving the system goals with minimal human intervention. Generative policies can help a coalition system to be more effective when working in a distributed, continuously transforming environment with a diverse set of members, resources, and tasks. Learning mechanisms based on logic programming, e.g., Inductive Logic Programming (ILP), have several properties that make them suitable and attractive for the creation and adaptation of generative policies, such as the ability to learn a general model from a small number of examples, and being able to incorporate existing background knowledge. ILP has recently been extended with the introduction of systems for Inductive Learning of Answer Set Programs (ILASP) which are capable of supporting automated acquisition of complex knowledge such as constraints, preferences and rule-based models. Motivated by the capabilities of ILASP, we present AGENP, an Answer Set Grammar-based Generative Policy Framework for Autonomous Managed Systems (AMS) that aims to support the creation and evolution of generative policies by leveraging ILASP. We describe the framework components, i.e., inputs, data structures, mechanisms to support the refinement and instantiation of policies, identification of policy violations, monitoring of policies, and policy adaptation according to changes in the AMS and its context. Additionally, we present the main work-flow for the global and local refinement of policies and their adaptation based on Answer Set Programming (ASP) for policy representation and reasoning using ILASP. We then discuss an application of the AGENP framework and present preliminary results.

The goal of inductive logic programming (ILP) is to search for a hypothesis that generalises training data and background knowledge. The challenge is searching vast hypothesis spaces, which is exacerbated because many logically equivalent hypotheses exist. To address this challenge, we introduce a method to break symmetries in the hypothesis space. We implement our idea in answer set programming. Our experiments on multiple domains, including visual reasoning and game playing, show that our approach can reduce solving times from over an hour to just 17 seconds.

The identification of the correct sense of a word is necessary for many tasks in automatic natural language processing like machine translation, information retrieval, speech and text processing. Automatic Word Sense Disambiguation (WSD) is difficult and accuracies with state-of-the art methods are substantially lower than in other areas of text understanding like part-of-speech tagging. One shortcoming of these methods is that they do not utilize substantial sources of background knowledge, such as semantic taxonomies and dictionaries, which are now available in electronic form (the methods largely use shallow syntactic features). Empirical results from the use of Inductive Logic Programming (ILP) have repeatedly shown the ability of ILP systems to use diverse sources of background knowledge. In this paper we investigate the use of ILP for WSD in two different ways: (a) as a stand-alone constructor of models for WSD; and (b) to build interesting features, which can then be used by standard model-builders such as SVM. In our experiments we examine a monolingual WSD task using the 32 English verbs contained in the SENSEVAL-3 benchmark data; and a bilingual WSD task using 7 highly ambiguous verbs in machine translation from English to Portuguese. Background knowledge available is from eight sources that provide a wide range of syntactic and semantic information. For both WSD tasks, experimental results show that ILP-constructed models and models built using ILP-generated features have higher accuracies than those obtained using a state-of-the art feature-based technique equipped with shallow syntactic features. This suggests that the use of ILP with diverse sources of background knowledge can provide one way for making substantial progress in the field of automatic WSD.

This paper is concerned with problems that arise when submitting large quantities of data to analysis by an Inductive Logic Programming (ILP) system. Complexity arguments usually make it prohibitive to analyse such datasets in their entirety. We examine two schemes that allow an ILP system to construct theories by sampling from this large pool of data. The first, “subsampling”, is a single-sample design in which the utility of a potential rule is evaluated on a randomly selected sub-sample of the data. The second, “logical windowing”, is multiple-sample design that tests and sequentially includes errors made by a partially correct theory. Both schemes are derived from techniques developed to enable propositional learning methods (like decision trees) to cope with large datasets. The ILP system CProgol, equipped with each of these methods, is used to construct theories for two datasets—one artificial (a chess endgame) and the other naturally occurring (a language tagging problem). In each case, we ask the following questions of CProgol equipped with sampling: (1) Is its theory comparable in predictive accuracy to that obtained if all the data were used (that is, no sampling was employed)?s and (2) Is its theory constructed in less time than the one obtained with all the data? For the problems considered, the answers to these questions is “yes”. This suggests that an ILP program equipped with an appropriate sampling method could begin to address problems satisfactorily that have hitherto been inaccessible simply due to data extent.

Learning event models from videos has applications ranging from abnormal event detection to content based video retrieval. Relational learning techniques such as Inductive Logic Programming (ILP) hold promise for building such models, but have not been successfully applied to the very large datasets which result from video data. In this paper we present a novel supervised learning framework to learn event models from large video datasets (~2.5 million frames) using ILP. Efficiency is achieved via the learning from interpretations setting and using a typing system. This allows learning to take place in a reasonable time frame with reduced false positives. The experimental results on video data from an airport apron where events such as Loading, Unloading, Jet-Bridge Parking etc are learned suggests that the techniques are suitable to real world scenarios.

Model transformation by example is a novel trend in model-driven software engineering. The rationale behind this is to utilize existing knowledge represented by source and target models of previously developed systems; such as requirements analysis and software design models, respectively. Such knowledge can be utilized to derive transformation rules to be applied in future system developments. To achieve this goal, machine learning techniques can assist in discovering and formalizing desired transformation rules. Inductive Logic Programming (ILP) represents a highly applicable machine learning technique in this context. Given a set of examples and background knowledge encoded as a set of first-order logic descriptions, an ILP system attempts to derive rules describing different transformation steps in a purely declarative way. The induced rules follow the same logical description as the given examples and background knowledge. The objective of this work is to introduce initial setup of an ILP system that can be utilized to derive analysis-design transformation rules from a set of examples that represent pairs of analysis-design models.

A phenotypic genetic algorithm for inductive logic programming

Recently, computer programs developed within the field of Inductive Logic Programming (ILP) have received some attention for their ability to construct restricted first-order logic solutions using problemspecific background knowledge. Prominent applications of such programs have been concerned with determining “structure-activity” relationships in the areas of molecular biology and chemistry. Typically the task here is to predict the “activity” of a compound, like toxicity, from its chemical structure. Research in the area shows that: (a) ILP programs have been restricted to qualitative predictions of activity (“high”, “low” etc.); (b) When appropriate attributes are available, ILP programs have not been able to better the performance of standard quantitative analysis techniques like linear regression. However ILP programs perform creditably when such attributes are unavailable; and (c) When both are applicable, ILP programs are usually slower than their propositional counterparts. This paper examines the use of ILP programs, not for obtaining theories complete for the sample, but as a method of “discovering” new attributes. These could then be used by methods like linear regression, thus allowing for quantitative predictions and the ability to use structural information as background knowledge. Using structure-activity tasks as a test-bed the utility of ILP programs in constructing new features was evaluated by examining the prediction of chemical activity using linear regression, with and without the aid of ILP learnt logical attributes. In three out of the five datasets examined the addition of ILP attributes produced statistically better results (P 10.01). In addition six important structural features that have escaped the attention of the expert chemists were discovered.KeywordsDihydrofolate ReductaseLogical AttributeInductive Logic ProgrammingBasic TemplateInductive Logic Programming SystemThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Computing similarity in high-dimensional vector spaces is a long-standing problem that has recently seen significant progress with the invention of the word2vec algorithm. Usually, it has been found that using an embedded representation results in much better performance for the task being addressed. It is not known whether embeddings can similarly improve performance with data of the kind considered by Inductive Logic Programming (ILP), in which data apparently dissimilar on the surface, can be similar to each other given domain (background) knowledge. In this paper, using several ILP classification benchmarks, we investigate if embedded representations are similarly helpful for problems where there is sufficient amounts of background knowledge. We use tasks for which we have domain expertise about the relevance of background knowledge available and consider two subsets of background predicates (“sufficient” and “insufficient”). For each subset, we obtain a baseline representation consisting of Boolean-valued relational features. Next, a vector embedding specifically designed for classification is obtained. Finally, we examine the predictive performance of widely-used classification methods with and without the embedded representation. With sufficient background knowledge we find no statistical evidence for an improved performance with an embedded representation. With insufficient background knowledge, our results provide empirical evidence that for the specific case of using deep networks, an embedded representation could be useful.

The predicates used for Inductive Logic Programming (ILP) systems are usually elusive and need to be hand-crafted in advance, which limits the generalization of the system when learning new rules without sufficient background knowledge. Predicate Invention (PI) for ILP is the problem of discovering new concepts that describe hidden relationships in the domain. PI can mitigate the generalization problem for ILP by inferring new concepts, giving the system a better vocabulary to compose logic rules. Although there are several PI approaches for symbolic ILP systems, PI for Neuro-Symbolic-ILP (NeSy-ILP) systems that can handle 3D visual inputs to learn logical rules using differentiable reasoning is still unaddressed. To this end, we propose a neuro-symbolic approach, NeSy- π , to invent predicates from visual scenes for NeSy-ILP systems based on clustering and extension of relational concepts, where π denotes the abbrivation of P redicate I nvention. NeSy- π processes visual scenes as input using deep neural networks for the visual perception and invents new concepts that support the task of classifying complex visual scenes. The invented concepts can be used by any NeSy-ILP system instead of hand-crafted background knowledge. Our experiments show that the NeSy- π is capable of inventing high-level concepts and solving complex visual logic patterns efficiently and accurately in the absence of explicit background knowledge. Moreover, the invented concepts are explainable and interpretable, while also providing competitive results with state-of-the-art NeSy-ILP systems. (github: https://github.com/ml-research/NeSy-PI )

Learning first-order logic programs from relational facts yields intuitive insights into the data. Inductive logic programming (ILP) models are effective in learning first-order logic programs from observed relational data. Symbolic ILP models support rule learning in a data-ecient manner. However, symbolic ILP models are not robust to learn from noisy data. Neuro-symbolic ILP models utilize neural networks to learn logic programs in a differentiable manner which improves the robustness of ILP models. However, most neuro-symbolic methods need a strong language bias to learn logic programs, which reduces the usability and flexibility of ILP models and limits the logic program formats. In addition, most neuro-symbolic ILP methods cannot learn logic programs effectively from both small-size datasets and large-size datasets such as knowledge graphs. In the paper, we introduce a novel differentiable ILP model called differentiable first-order rule learner (DFORL), which is scalable to learn rules from both smaller and larger datasets. Besides, DFORL only needs the number of variables in the learned logic programs as input. Hence, DFORL is easy to use and does not need a strong language bias. We demonstrate that DFORL can perform well on several standard ILP datasets, knowledge graphs, and probabilistic relation facts and outperform several well-known differentiable ILP models. Experimental results indicate that DFORL is a precise, robust, scalable, and computationally cheap differentiable ILP model.

Partially Observable Markov Decision Processes (POMDPs) are a powerful framework for planning under uncertainty. They allow to model state uncertainty as a belief probability distribution. Approximate solvers based on Monte Carlo sampling show great success to relax the computational demand and perform online planning. However, scaling to complex realistic domains with many actions and long planning horizons is still a major challenge, and a key point to achieve good performance is guiding the action-selection process with domain-dependent policy heuristics which are tailored for the specific application domain. We propose to learn high-quality heuristics from POMDP traces of executions generated by any solver. We convert the belief-action pairs to a logical semantics, and exploit data- and time-efficient Inductive Logic Programming (ILP) to generate interpretable belief-based policy specifications, which are then used as online heuristics. We evaluate thoroughly our methodology on two notoriously challenging POMDP problems, involving large action spaces and long planning horizons, namely, rocksample and pocman. Considering different state-of-the-art online POMDP solvers, including POMCP, DESPOT and AdaOPS, we show that learned heuristics expressed in Answer Set Programming (ASP) yield performance superior to neural networks and similar to optimal handcrafted task-specific heuristics within lower computational time. Moreover, they well generalize to more challenging scenarios not experienced in the training phase (e.g., increasing rocks and grid size in rocksample, incrementing the size of the map and the aggressivity of ghosts in pocman).

This paper introduces an automated technique for constructing valid behavior specifications of programs (at the system call level) that are independent of system vulnerabilities and are highly effective in identifying intrusions. The technique employs a machine learning method, inductive logic programming (ILP), for synthesizing first order logic formulas that describe the valid operations of a program from the normal runs of the program. ILP backed by theories and techniques extended from computational logic, allows the use of complex domain-specific background knowledge in the learning process to produce sound and consistent knowledge. A specification induction engine has been developed by extending an existing ILP tool and has been used to construct specifications for several (>10) privileged programs in Unix. Coupling with rich background knowledge in systems and security, the prototype induction engine generates human understandable and analytable specifications that are as good as those specified by a human. Preliminary experiments with existing attacks show that the generated specifications are highly effective in detecting attacks that subvert privileged programs to gain unauthorized accesses to resources.

Despite early interest Predicate Invention has lately been under-explored within ILP. We develop a framework in which predicate invention and recursive generalisations are implemented using abduction with respect to a meta-interpreter. The approach is based on a previously unexplored case of Inverse Entailment for Grammatical Inference of Regular languages. Every abduced grammar H is represented by a conjunction of existentially quantified atomic formulae. Thus ¬H is a universally quantified clause representing a denial. The hypothesis space of solutions for ¬H can be ordered by ?-subsumption. We show that the representation can be mapped to a fragment of Higher-Order Datalog in which atomic formulae in H are projections of first-order definite clause grammar rules and the existentially quantified variables are projections of first-order predicate symbols. This allows predicate invention to be effected by the introduction of first-order variables. Previous work by Inoue and Furukawa used abduction and meta-level reasoning to invent predicates representing propositions. By contrast, the present paper uses abduction with a meta-interpretive framework to invent relations. We describe the implementations of Meta-interpretive Learning (MIL) using two different declarative representations: Prolog and Answer Set Programming (ASP). We compare these implementations against a state-of-the-art ILP system MC-TopLog using the dataset of learning Regular and Context-Free grammars as well learning a simplified natural language grammar and a grammatical description of a staircase. Experiments indicate that on randomly chosen grammars, the two implementations have significantly higher accuracies than MC-TopLog. In terms of running time, Metagol is overall fastest in these tasks. Experiments indicate that the Prolog implementation is competitive with the ASP one due to its ability to encode a strong procedural bias. We demonstrate that MIL can be applied to learning natural grammars. In this case experiments indicate that increasing the available background knowledge, reduces the running time. Additionally ASPM (ASP using a meta-interpreter) is shown to have a speed advantage over Metagol when background knowledge is sparse. We also demonstrate that by combining MetagolR (Metagol with a Regular grammar meta-interpreter) and MetagolCF (Context-Free meta-interpreter) we can formulate a system, MetagolRCF, which can change representation by firstly assuming the target to be Regular, and then failing this, switch to assuming it to be Context-Free. MetagolRCF runs up to 100 times faster than MetagolCF on grammars chosen randomly from Regular and non-Regular Context-Free grammars.