DeepProbLog: Neural Probabilistic Logic Programming
By Robin Manhaeve, Sebastijan Dumancic, Angelika Kimmig, Thomas Demeester, Luc De Raedt. 2018
764 citations recorded by the source corpus.
We introduce DeepProbLog, a probabilistic logic programming language that incorporates deep learning by means of neural predicates. We show how existing inference and learning techniques can be adapted for the new language. Our experiments demonstrate that DeepProbLog supports both symbolic and subsymbolic representations and inference, 1) program induction, 2) probabilistic (logic) programming, and 3) (deep) learning from examples. To the best of our knowledge, this work is the first to propose a framework where general-purpose neural networks and expressive probabilistic-logical modeling and reasoning are integrated in a way that exploits the full expressiveness and strengths of both worlds and can be trained end-to-end based on examples.
Subfields: Artificial Intelligence, Probabilistic & Bayesian Methods, Knowledge Graphs & Representation
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Josh Tenenbaum
- Sebastian Riedel
- Mohit Bansal
- Percy Liang
- Yoshua Bengio
- Luc De Raedt
- Jianfeng Gao
- Luke S. Zettlemoyer
In Probabilistic & Bayesian Methods
- Angelika Kimmig
- Luc De Raedt
- A.S. d'Avila Garcez
- Marco Gori
- Sebastijan Dumancic
- Luciano Serafini
- Thomas Demeester
- Robin Manhaeve
In Knowledge Graphs & Representation
- A.S. d'Avila Garcez
- Luciano Serafini
- Kristian Kersting
- Michael Spranger
- Marco Gori
- L. C. Lamb
- Son Dinh Tran
- Sebastijan Dumancic
Citations in the corpus
- Neural Logic Machines (2019, neutral): “Our paper also differs from existing approaches on using neural networks to augment symbolic rule induction (Lippi & Frasconi, 2009; Manhaeve et al., 2018).”
- Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, neutral): “The interesting aspect is that this blend allows for expressive probabilistic-logical reasoning in an end-to-end fashion [300] .”
- Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning (2019, methodology): “In a related approach, called DeepProbLog, end-to-end learning and reasoning have been studied [28] where outputs of neural networks are used as "neural predicates" for ProbLog [10].”
- Integrating Learning and Reasoning with Deep Logic Models (2019, background): “DeepProbLog [22] extends the popular ProbLog [7] probabilistic programming framework with the integration of deep learners.”
- Learning from Rules Generalizing Labeled Exemplars (2020, background): “There is growing interest in integrating logical rules with labeled examples for training networks, specifically for structured outputs (Manhaeve et al., 2018; Xu et al., 2018; Fischer et al., 2019; Sun et al., 2018; Ren et al., 2018a).”
- SATNet: Bridging deep learning and logical reasoning using a differentiable satisfiability solver (2019, background): “…focused on creating differentiable modules from an existing set of known relationships , so that a deep network can learn the parameters of these relationships (Dai et al., 2018; Manhaeve et al., 2018; Sourek et al., 2018; Xu et al., 2018; Hu et al., 2016; Yang et al., 2017; Selsam et al., 2018).”
- Logical Rule Induction and Theory Learning Using Neural Theorem Proving (2018, background): “A recent work (Manhaeve et al. 2018) starts from a probabilistic logic programming language (problog) and extends it to handle neural predicates which compute probabilities.”
- Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting with their Explanations (2020, background): “Truly rule-based classifiers [38, 29], logic circuits [26], or probabilistic approaches [5, 37, 19, 30], are principally viable alternatives for the Set Transformer, though it remains preferable for this module to handle unordered sets.”
- Conversational Neuro-Symbolic Commonsense Reasoning (2020, neutral): “DeepProbLog [31] developed a probabilistic logic programming language that is suitable for applications containing categorical variables.”
- Exploring End-to-End Differentiable Natural Logic Modeling (2020, background): “Evans and Grefenstette (2018) treat inductive logic programming as a satisfiability problem and Manhaeve et al. (2018) combine high-level symbolic oriented reasoning with low-level neural perception models.”
- Unsupervised Grounding of Plannable First-Order Logic Representation from Images (2019, background): “Neural theorem proving systems (Rockt¨aschel and Riedel 2017; Sourek et al. 2018; Manhaeve et al. 2018) learn to reason about logical relations.”
- Logic Tensor Networks (2020, methodology): “• Neurosymbolic architectures for the integration of inductive learning and deductive reasoning: These architectures seek to enable the integration of inductive and deductive reasoning in a unique fully-differentiable framework [22, 41, 47, 14, 45].”
- Neurosymbolic Programming (2021, background): “A notable exception is DeepProbLog (Manhaeve et al., 2018), which integrates neural components into the ProbLog (De Raedt et al., 2007) probabilistic logic programming system.”
- Reasoning with Transformer-based Models: Deep Learning, but Shallow Reasoning (2021, methodology): “Other approaches develop neuro-symbolic methods, which teach reasoning strategies by gradient-based optimisation [Minervini et al., 2020], or combine probabilistic logic programming with neural networks [Manhaeve et al., 2018].”
- Three Modern Roles for Logic in AI (2020, neutral): “, [35, 47, 50].”
- Neuro-Symbolic = Neural + Logical + Probabilistic (2019, methodology): “We examine the state-of-the-art with regard to these claims and briefly position our own contribution DeepProbLog in this perspective.”