Structure learning of probabilistic logic programs by searching the clause space

Bellodi, Elena; Riguzzi, Fabrizio

doi:10.1017/s1471068413000689

Cited by 53 publications

(55 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although there has been significant progress in the field of PILP [25,[49][50][51]26] for learning annotated Prolog programs, PILP under the answer set semantics is still relatively young, and thus, there are few approaches. PrASP [52,53,27] considers the problem of weight learning, and in fact uses a similar example of learning about coins.…”

Section: Relation To Probabilistic Ilpmentioning

confidence: 99%

The complexity and generality of learning answer set programs

Law

Russo

Broda

2018

Artificial Intelligence

View full text Add to dashboard Cite

Traditionally most of the work in the field of Inductive Logic Programming (ILP) has addressed the problem of learning Prolog programs. On the other hand, Answer Set Programming is increasingly being used as a powerful language for knowledge representation and reasoning, and is also gaining increasing attention in industry. Consequently, the research activity in ILP has widened to the area of Answer Set Programming, witnessing the proposal of several new learning frameworks that have extended ILP to learning answer set programs. In this paper, we investigate the theoretical properties of these existing frameworks for learning programs under the answer set semantics. Specifically, we present a detailed analysis of the computational complexity of each of these frameworks with respect to the two decision problems of deciding whether a hypothesis is a solution of a learning task and deciding whether a learning task has any solutions. We introduce a new notion of generality of a learning framework, which enables us to define a framework to be more general than another in terms of being able to distinguish one ASP hypothesis solution from a set of incorrect ASP programs. Based on this notion, we formally prove a generality relation over the set of existing frameworks for learning programs under answer set semantics. In particular, we show that our recently proposed framework, Context-dependent Learning from Ordered Answer Sets, is more general than brave induction, induction of stable models, and cautious induction, and maintains the same complexity as cautious induction, which has the highest complexity of these frameworks.

show abstract

Section: Relation To Probabilistic Ilpmentioning

confidence: 99%

The complexity and generality of learning answer set programs

Law

Russo

Broda

2018

Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…The results are shown in Figure 3: neither FOIL nor Alchemy's MLN method 3 outperform the simple baseline of predicting exactly the facts in the incomplete database. 4 2 Or in FOIL's case, an approach broadly similar to pseudolikelihood. 3 Alchemy's performance is quite sensitive to the precise set of missing facts, so we average over ten runs in the figure.…”

Section: Structure Learning Is Difficult For Kb Completionmentioning

confidence: 99%

“…3 Alchemy's performance is quite sensitive to the precise set of missing facts, so we average over ten runs in the figure. 4 Note that we have also experimented with a more recent "Learning with Structural Motifs (LSM)" variant [15] for learning MLN, but the results were much worse than Alchemy: we only observe a MAP of 10.7 on the missing 5% setting. This is because LSM is designed to learn long…”

Section: Structure Learning Is Difficult For Kb Completionmentioning

confidence: 99%

“…There has been some prior work on learning BDD-based probabilistic programs, however; in particular, they have been used as the substrate for structure learning in Bellodi and Riguzzi's systems SLIPCASE [3] and SLIPCOVER [4]. In these systems, beam search is used to explore a space of probabilistic logic programs, and candidate programs are scored by running a small number of iterations of EM.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Structure Learning via Parameter Learning

Wang

Mazaitis

Cohen

2014

Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management

View full text Add to dashboard Cite

A key challenge in information and knowledge management is to automatically discover the underlying structures and patterns from large collections of extracted information. This paper presents a novel structure-learning method for a new, scalable probabilistic logic called ProPPR. Our approach builds on the recent success of meta-interpretive learning methods in Inductive Logic Programming (ILP), and we further extends it to a framework that enables robust and efficient structure learning of logic programs on graphs: using an abductive second-order probabilistic logic, we show how first-order theories can be automatically generated via parameter learning. To learn better theories, we then propose an iterated structural gradient approach that incrementally refines the hypothesized space of learned firstorder structures. In experiments, we show that the proposed method further improves the results, outperforming competitive baselines such as Markov Logic Networks (MLNs) and FOIL on multiple datasets with various settings; and that the proposed approach can learn structures in a large knowledge base in a tractable fashion.

show abstract

“…Each refinement is scored by learning the parameters with EMBLEM and using the LL of the examples returned by it. SLIPCOVER (Bellodi and Riguzzi 2014) differs from SLIPCASE because the beam search is performed in the space of clauses. In this way, a set of promising clauses is identified and these are added one by one to the empty theory, keeping each clause if the LL improves.…”

Section: Introductionmentioning

confidence: 99%

Bandit-based Monte-Carlo structure learning of probabilistic logic programs

2015

Self Cite

View full text Add to dashboard Cite

Probabilistic logic programming can be used to model domains with complex and uncertain relationships among entities. While the problem of learning the parameters of such programs has been considered by various authors, the problem of learning the structure is yet to be explored in depth. In this work we present an approximate search method based on a one-player game approach, called LEMUR. It sees the problem of learning the structure of a probabilistic logic program as a multi-armed bandit problem, relying on the Monte-Carlo tree search UCT algorithm that combines the precision of tree search with the generality of random sampling. LEMUR works by modifying the UCT algorithm in a fashion similar to FUSE, that considers a finite unknown horizon and deals with the problem of having a huge branching factor. The proposed system has been tested on various real-world datasets and has shown good performance with respect to other state of the art statistical relational learning approaches in terms of classification abilities.

show abstract

Structure learning of probabilistic logic programs by searching the clause space

Cited by 53 publications

References 60 publications

The complexity and generality of learning answer set programs

The complexity and generality of learning answer set programs

Structure Learning via Parameter Learning

Bandit-based Monte-Carlo structure learning of probabilistic logic programs

Contact Info

Product

Resources

About