Collocated with ICLP 2021
September 20th, 2021, Porto
Due to the ongoing pandemic, the workshop will be held online.
The workshop will take place on September 20th
|14:00 - 14:10||Opening|
|14:10 - 14:55||Invited Talk. Nico Potyka: From Probabilistic Programming to Probabilistic Argumentation|
|14:55 - 15:20||Thomas Eiter, Markus Hecher and Rafael Kiesel: aspmc: An Algebraic Answer Set Counter|
|15:20 - 15:35||Break|
|15:35 - 16:20||Invited Talk. Alessandro Antonucci: Machine Learning over Knowledge Bases by Probabilistic Circuits|
|16:20 - 16:45||Felix Weitkämper, Beatrice Sarbu and Kailin Sun: Modelling infectious disease dynamics with probabilistic logic programming|
|16:45 - 17:10||Kilian Rueckschloss and Felix Weitkaemper: Reasoning about Independence in Open Universe Probabilistic Logic Programs|
|17:10 - 17:25||Break|
|17:25 - 18:10||Invited Talk. Manfred Jaeger: Probabilistic relational learning and reasoning: relational Bayesian vs. graph neural networks|
|18:10 - 18:35||Robert Zinkov and Willliam Byrd: probKanren: A Simple Probabilistic extension for microKanren|
|18:35 - 18:45||Closing|
Probabilistic logic programming (PLP) approaches have received much attention in this century. They address the need to reason about relational domains under uncertainty arising in a variety of application domains, such as bioinformatics, the semantic web, robotics, and many more. Developments in PLP include new languages that combine logic programming with probability theory as well as algorithms that operate over programs in these formalisms.
The workshop encompasses all aspects of combining logic, algorithms, programming and probability.
PLP is part of a wider current interest in probabilistic programming. By promoting probabilities as explicit programming constructs, inference, parameter estimation and learning algorithms can be run over programs which represent highly structured probability spaces. Due to logic programming's strong theoretical underpinnings, PLP is one of the more disciplined areas of probabilistic programming. It builds upon and benefits from the large body of existing work in logic programming, both in semantics and implementation, but also presents new challenges to the field. PLP reasoning often requires the evaluation of large number of possible states before any answers can be produced thus breaking the sequential search model of traditional logic programs.
While PLP has already contributed a number of formalisms, systems and well understood and established results in: parameter estimation, tabling, marginal probabilities and Bayesian learning, many questions remain open in this exciting, expanding field in the intersection of AI, machine learning and statistics.
This workshop aims to bring together researchers in all aspects of probabilistic logic programming, including theoretical work, system implementations and applications. Interactions between theoretical and applied minded researchers are encouraged. The presence of this workshop at ICLP is intended to encourage collaboration with researchers from the field of logic programming,
This workshop provides a forum for the exchange of ideas, presentation of results and preliminary work in all areas related to probabilistic logic programming; including, but not limited to:
The registration for PLP2021 is managed through the ICLP registration system. To register, visit the ICLP registration page.
|Notification to authors:|
|Camera ready version due:|
|Workshop date:||September 20th, 2021|
(the deadline for all dates is intended Anywhere on Earth (UTC-12))
Machine Learning over Knowledge Bases by Probabilistic Circuits
Probabilistic Circuits are emerging as an important class of probabilistic graphical models based on logical circuits annotated by probabilities. These are tractable models typically allowing to answer complex inferential queries and perform reasoning in polynomial time. Moreover, the mass functions induced by these model is by construction consistent with the formula induced by the underlying logical circuit.
This makes those circuits a natural choice to address machine learning tasks when the data are known to be consistent with a knowledge base. After having discussed the most relevant features of these models and the existing training and inference algorithms, we present a number of challenging applicative tasks including problems such as preference learning, routing, objects identification and tracking.
Probabilistic relational learning and reasoning: relational Bayesian vs. graph neural networks
Graph neural networks (GNNs) have received in recent years a significant amount of attention in machine learning for graph and network data. Recently, the expressiveness of GNNs for node classification tasks has been investigated, and characterizations in terms of fragments of first-order logic of their discriminative power have been obtained. These characterizations are similar in spirit to results obtained in the past for the relational Bayesian network (RBN) probabilistic-relational modeling framework. In this talk I am going to show how common types of GNNs can be seen as special cases of RBNs in the sense that there is a direct mapping from GNN architectures to an RBN representation, such that the general parameter learning method for RBNs applied to this representation coincides with the training method for GNNs. This gives us the ability to seamlessly integrate graph neural architectures into an RBN model, which, on the other hand, also can incorporate interpretable symbolic representation, and which can support a greater variety of inference tasks than the standard classification tasks supported by GNNs. Thus, RBNs provide a platform both for neuro-symbolic integration, and for the integration of learning and reasoning.
From Probabilistic Programming to Probabilistic Argumentation
One natural approach to propositional probabilistic programming is Nilsson's probabilistic logic. We discuss the basic framework in the propositional setting, some reasoning ideas and extensions that allow handling inconsistent information. We then discuss its relationship to probabilistic epistemic argumentation. Roughly speaking, probabilistic epistemic argumentation is to classical argumentation as Nilsson's probabilistic logic is to classical logic. We discuss similarities and differences between the two and in which scenarios one approach may be better suited than the other.
Submissions will be managed via EasyChair: https://easychair.org/conferences/?conf=plp21.
Contributions should be prepared in the 1-column CEURART style (also available as an overleaf project). A mixture of papers are sought including: new results, work in progress as well as technical summaries of recent substantial contributions. Papers presenting new results should be 6-15 pages in length. Work in progress and technical summaries can be shorter (2-5 pages). The workshop proceedings will clearly indicate the type of each paper.
At least one author of each accepted paper will be required to attend the workshop to present the contribution.