Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning最新文献_第3页

Pushing Optimal ABox Repair from EL Towards More Expressive Horn-DLs 推动最佳ABox修复从EL到更富有表现力的喇叭- dl

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.25368/2022.131

F. Baader, Francesco Kriegel

{"title":"Pushing Optimal ABox Repair from EL Towards More Expressive Horn-DLs","authors":"F. Baader, Francesco Kriegel","doi":"10.25368/2022.131","DOIUrl":"https://doi.org/10.25368/2022.131","url":null,"abstract":"Ontologies based on Description Logic (DL) represent general background knowledge in a terminology (TBox) and the actual data in an ABox. DL systems can then be used to compute consequences (such as answers to certain queries) from an ontology consisting of a TBox and an ABox. Since both human-made and machine-learned data sets may contain errors, which manifest themselves as unintuitive or obviously incorrect consequences, repairing DL-based ontologies in the sense of removing such unwanted consequences is an important topic in DL research. Most of the repair approaches described in the literature produce repairs that are not optimal, in the sense that they do not guarantee that only a minimal set of consequences is removed. In a series of papers, we have developed an approach for computing optimal repairs, starting with the restricted setting of an EL instance store, extending this to the more general setting of a quantified ABox (where some individuals may be anonymous), and then adding a static EL TBox. Here, we extend the expressivity of the underlying DL considerably, by adding nominals, inverse roles, regular role inclusions and the bottom concept to EL, which yields a fragment of the well-known DL Horn-SROIQ. The ideas underlying our repair approach still apply to this DL, though several non-trivial extensions are needed to deal with the new constructors and axioms. The developed repair approach can also be used to treat unwanted consequences expressed by certain conjunctive queries or regular path queries, and to handle Horn-ALCOI TBoxes with regular role inclusions.","PeriodicalId":351970,"journal":{"name":"Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114153730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 5

Looking Inside the Black-Box: Logic-based Explanations for Neural Networks 黑盒子内部:基于逻辑的神经网络解释

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/45

João Ferreira, Manuel de Sousa Ribeiro, Ricardo Gonçalves, João Leite

{"title":"Looking Inside the Black-Box: Logic-based Explanations for Neural Networks","authors":"João Ferreira, Manuel de Sousa Ribeiro, Ricardo Gonçalves, João Leite","doi":"10.24963/kr.2022/45","DOIUrl":"https://doi.org/10.24963/kr.2022/45","url":null,"abstract":"Deep neural network-based methods have recently enjoyed great popularity due to their effectiveness in solving difficult tasks. Requiring minimal human effort, they have turned into an almost ubiquitous solution in multiple domains. However, due to the size and complexity of typical neural network models' architectures, as well as the sub-symbolical nature of the representations generated by their neuronal activations, neural networks are essentially opaque, making it nearly impossible to explain to humans the reasoning behind their decisions. We address this issue by developing a procedure to induce human-understandable logic-based theories that attempt to represent the classification process of a given neural network model, based on the idea of establishing mappings from the values of the activations produced by the neurons of that model to human-defined concepts to be used in the induced logic-based theory. Exploring the setting of a synthetic image classification task, we provide empirical results to assess the quality of the developed theories for different neural network models, compare them to existing theories on that task, and give evidence that the theories developed through our method are faithful to the representations learned by the neural networks that they are built to describe.","PeriodicalId":351970,"journal":{"name":"Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning","volume":"86 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122686099","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 19

Sum-Product Loop Programming: From Probabilistic Circuits to Loop Programming 和积循环规划:从概率电路到循环规划

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/47

Viktor Pfanschilling, Hikaru Shindo, D. Dhami, K. Kersting

{"title":"Sum-Product Loop Programming: From Probabilistic Circuits to Loop Programming","authors":"Viktor Pfanschilling, Hikaru Shindo, D. Dhami, K. Kersting","doi":"10.24963/kr.2022/47","DOIUrl":"https://doi.org/10.24963/kr.2022/47","url":null,"abstract":"Recently, Probabilistic Circuits such as Sum-Product Networks have received growing attention, as they can represent complex features but still provide tractable inference. Although quite successful, unfortunately, they lack the capability of handling control structures, such as for and while loops. In this work, we introduce Sum-Product Loop Language (SPLL), a novel programming language that is capable of tractable inference on complex probabilistic code that includes loops.\u0000SPLL has dual semantics: every program has generative semantics familiar to most programmers and\u0000probabilistic semantics that assign a probability to each possible result. This way, the programmer can describe how to generate samples almost like in any standard programming language. The language takes care of computing the probability values of all results for free at run time.\u0000We demonstrate that SPLL inherits the beneficial properties of PCs, namely tractability and differentiability, while generalizing to other distributions and programs, and retains substantial computational similarities.","PeriodicalId":351970,"journal":{"name":"Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125341345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2

On the Representation of Darwiche and Pearl's Epistemic States for Iterated Belief Revision 论达尔文和珀尔的认知状态在反复信念修正中的表征

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/32

Nicolas Schwind, S. Konieczny, R. Pérez

引用次数: 5

Hyperintensional Partial Meet Contractions 高强度部分满足收缩

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/34

Marlo Souza, Renata Wassermann

引用次数: 1

Embed2Sym - Scalable Neuro-Symbolic Reasoning via Clustered Embeddings Embed2Sym -基于聚类嵌入的可扩展神经符号推理

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/44

Yaniv Aspis, K. Broda, Jorge Lobo, Alessandra Russo

{"title":"Embed2Sym - Scalable Neuro-Symbolic Reasoning via Clustered Embeddings","authors":"Yaniv Aspis, K. Broda, Jorge Lobo, Alessandra Russo","doi":"10.24963/kr.2022/44","DOIUrl":"https://doi.org/10.24963/kr.2022/44","url":null,"abstract":"Neuro-symbolic reasoning approaches proposed in recent years combine a neural perception component with a symbolic reasoning component to solve a downstream task. By doing so, these approaches can provide neural networks with symbolic reasoning capabilities, improve their interpretability and enable generalization beyond the training task. However, this often comes at the cost of poor training time, with potential scalability issues. In this paper, we propose a scalable neuro-symbolic approach, called Embed2Sym. We complement a two-stage (perception and reasoning) neural network architecture designed to solve a downstream task end-to-end with a symbolic optimisation method for extracting learned latent concepts. Specifically, the trained perception network generates clusters in embedding space that are identified and labelled using symbolic knowledge and a symbolic solver. With the latent concepts identified, a neuro-symbolic model is constructed by combining the perception network with the symbolic knowledge of the downstream task, resulting in a model that is interpretable and transferable. Our evaluation shows that Embed2Sym outperforms state-of-the-art neuro-symbolic systems on benchmark tasks in terms of training time by several orders of magnitude while providing similar if not better accuracy.","PeriodicalId":351970,"journal":{"name":"Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127287070","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 10

Dynamic Deontic Logic for Permitted Announcements 允许通知的动态道义逻辑

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/23

Xu Li, D. Gabbay, Réka Markovich

引用次数: 1

On the Expressive Power of Intermediate and Conditional Effects in Temporal Planning 论时间规划中中介效应和条件效应的表达能力

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/18

N. Gigante, A. Micheli, Enrico Scala

{"title":"On the Expressive Power of Intermediate and Conditional Effects in Temporal Planning","authors":"N. Gigante, A. Micheli, Enrico Scala","doi":"10.24963/kr.2022/18","DOIUrl":"https://doi.org/10.24963/kr.2022/18","url":null,"abstract":"Automated planning is the task of finding a sequence of actions that reach a desired goal, given a description of their applicability and their effects on the world. In temporal planning, actions have a duration and can overlap in time. In modern temporal planning formalisms, two features have been introduced which are very useful from a modeling perspective, but are not yet thoroughly understood: intermediate conditions and effects (ICE) and conditional effects. The expressive power of such constructs is yet not well comprehended, especially when time is dense, and no minimum separation is required between mutex events. This paper reveals that both ICE and conditional effects do not add expressive power with regards to common temporal planning formalisms. In particular, we show how they can be compiled away using a polynomial-size encoding that makes no assumptions on the time domain. This encoding advances our understanding of these features, and allow their use with simple temporal planners that lack their support. Moreover, it provides a constructive proof that temporal planning with ICE and conditional effects remains PSPACE-complete.","PeriodicalId":351970,"journal":{"name":"Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115669532","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Counting Queries over ELHI⊥ Ontologies ELHI⊥本体上的查询计数

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-07-01 DOI: 10.24963/kr.2022/6

Meghyn Bienvenu, Quentin Manière, M. Thomazo

引用次数: 2

Compound Conditionals as Random Quantities and Boolean Algebras 作为随机量和布尔代数的复合条件

Proceedings of the Nineteenth International Conference on Principles of Knowledge Representation and Reasoning Pub Date : 2022-04-29 DOI: 10.24963/kr.2022/15

T. Flaminio, A. Gilio, L. Godo, G. Sanfilippo

引用次数: 5