Abstract
Symbolic knowledge representation and reasoning and deep learning are fundamentally different approaches to artificial intelligence with complementary capabilities. The former are transparent and data-efficient, but they are sensitive to noise and cannot be applied to non-symbolic domains where the data is ambiguous. The latter can learn complex tasks from examples, are robust to noise, but are black boxes; require large amounts of –not necessarily easily obtained– data, and are slow to learn and prone to adversarial examples. Either paradigm excels at certain types of problems where the other paradigm performs poorly. In order to develop stronger AI systems, integrated neuro-symbolic systems that combine artificial neural networks and symbolic reasoning are being sought. In this context, one of the fundamental open problems is how to perform logic-based deductive reasoning over knowledge bases by means of trainable artificial neural networks. This paper provides a brief summary of the authors’ recent efforts to bridge the neural and symbolic divide in the context of deep deductive reasoners. Throughout the paper we will discuss strengths and limitations of models in term of accuracy, scalability, transferability, generalizabiliy, speed, and interpretability, and finally, will talk about possible modifications to enhance desirable capabilities. More specifically, in terms of architectures, we are looking at Memory-augmented networks, Logic Tensor Networks, and compositions of LSTM models to explore their capabilities and limitations in conducting deductive reasoning. We are applying these models on Resource Description Framework (RDF), first-order logic, and the description logic \(\mathcal {E}{\mathscr{L}}^{+}\) respectively.
Similar content being viewed by others
Notes
The topic is being investigated, of course, with some recent progress being made. For example, [1] report on an application of deep learning to planning, and explicitly frame it as work towards bridging the “subsymbolic-symbolic boundary.”
Evrahimi et al [22] is under review at AAAI-MAKE 2021 symposium at the time of submitting this journal paper.
Induction like in Inductive Logic Programming or Relational Learning has statistical aspects and is much closer in nature to a machine learning task, and thus arguably easier to tackle using machine learning approaches.
Any may be too grandiose a statement, but these are the ones we are looking at.
Attempting to find finite representations for infinite sets – in the cases where this would even be reasonably possible – would add another layer of complication which we are currently not considering.
In fact, it comes with three different ones, but we have only considered the most comprehensive one, the RDFS Semantics.
Datalog is equivalent to function-free definite logic programming [34].
https://github.com/thunlp/KRLPapers has an extensive listing of existing work on knowledge embeddings.
Some deductive entailment algorithms can even be understood as simply a type of syntax rewriting systems.
We will use the prefix dbr: to refer to DBpedia entities.
DNN cannot be used because the training consists of just positive instantiated atoms, the network would eventually just learn to output 1 for every input.
References
Asai M, Fukunaga A (2018) Classical planning in deep latent space: Bridging the subsymbolic-symbolic boundary. In: McIlraith SA, Weinberger KQ (eds) Proceedings of the Thirty-Second AAAI Conference on artificial intelligence, New Orleans, Louisiana, USA, February 2-7, 2018, AAAI Press
Bach SH, Broecheler M, Huang B, Getoor L (2017) Hinge-loss Markov random fields and probabilistic soft logic. J Mach Learn Res 18:1–67
Bader S, Hitzler P, Hölldobler S (2008) Connectionist model generation: a first-order approach. Neurocomputing 71(13-15):2420–2432
Bader S, Hitzler P, Hölldobler S, Witzel A (2007) A fully connectionist model generator for covered first-order logic programs. In: Veloso MM (ed) IJCAI 2007, Proceedings of the 20th international joint conference on artificial intelligence, Hyderabad, India, January 6-12, 2007, pp 666–671
Bahdanau D, Bosc T, Jastrzebski S, Grefenstette E, Vincent P, Bengio Y (2017) Learning to compute word embeddings on the fly. arXiv:1706.00286
Besold TR, d’Avila Garcez A, Bader S, Bowman H, Domingos P, Hitzler P, Kühnberger K, Lamb L, Lowd D, Lima PMV, de Penning L, Pinkas G, Poon H, Zaverucha G (2017) Neural-symbolic learning and reasoning: A survey and interpretation. arXiv:1711.03902
Bianchi F, Hitzler P (2019) On the capabilities of logic tensor networks for deductive reasoning. In: AAAI Spring symposium: combining machine learning with knowledge engineering
Bianchi F, Palmonari M, Hitzler P, Serafini L (2019) Complementing logical reasoning with sub-symbolic commonsense. In: International joint conference on rules and reasoning, Springer, pp 161–170
Bianchi F, Palmonari M, Nozza D (2018) Towards encoding time in text-based entity embeddings. In: International semantic web conference, Springer, pp 56–71
Bianchi F, Rossiello G, Costabello L, Palmonari M, Minervini P (2020) Knowledge graph embeddings and explainable AI arXiv:2004.14843
Bojanowski P, Grave E, Joulin A, Mikolov T (2017) Enriching word vectors with subword information. Trans Assoc Comput Linguist 5:135–146
Bordes A, Usunier N, García-Durán A, Weston J, Yakhnenko O (2013) Translating embeddings for modeling multi-relational data. In: Burges CJC, Bottou L, Ghahramani Z, Weinberger KQ (eds) Advances in neural information processing systems 26: 27th Annual conference on neural information processing systems 2013. Proceedings of a meeting held December 5-8, 2013, Lake Tahoe, Nevada, United States, pp 2787–2795
Cai H, Zheng VW, Chang K (2018) A comprehensive survey of graph embedding: problems, techniques and applications. IEEE Trans Knowl Data Eng
Confalonieri R, Besold TR, Weyde T, Creel K, Lombrozo T, Mueller ST, Shafto P (2019) What makes a good explanation? cognitive dimensions of explaining intelligent machines. In: Goel AK, Seifert CM, Freksa C. (eds) Proceedings of the 41th Annual meeting of the cognitive science society, CogSci 2019: Creativity + Cognition + Computation, Montreal, Canada, July 24-27, 2019, pp 25–26. cognitivesciencesociety.org. https://mindmodeling.org/cogsci2019/papers/0013/index.html
Cyganiak R., Wood D, Lanthaler M (eds.) RDF 1.1 Concepts and Abstract Syntax. W3C Recommendation 25 February 2014 (2014). Available from http://www.w3.org/TR/rdf11-concepts/
Das R, Dhuliawala S, Zaheer M, Vilnis L, Durugkar I, Krishnamurthy A, Smola A, McCallum A (2018) Go for a walk and arrive at the answer: Reasoning over paths in knowledge bases using reinforcement learning. In: International conference on learning representations
De Raedt L, Kimmig A (2015) Probabilistic (logic) programming concepts. Mach Learn 100 (1):5–47
De Silva TS, MacDonald D, Paterson G, Sikdar KC, Cochrane B (2011) Systematized nomenclature of medicine clinical terms (SNOMED CT) to represent computed tomography procedures. Comput Methods Prog Biomed 101(3):324–329. https://doi.org/10.1016/j.cmpb.2011.01.002
Donadello I, Serafini L, d’Avila Garcez A (2017) Logic tensor networks for semantic image interpretation. In: Sierra C (ed) Proceedings of the twenty-sixth international joint conference on artificial intelligence, IJCAI 2017, Melbourne, Australia, August 19-25, 2017, pp 1596–1602. ijcai.org
Dong H, Mao J, Lin T, Wang C, Li L, Zhou D (2018) Neural logic machines. In: International conference on learning representations
Eberhart A, Ebrahimi M, Zhou L, Shimizu C, Hitzler P (2020) Completion reasoning emulation for the description logic EL+. In: Martin A, Hinkelmann K, Fill H, Gerber A, Lenat D, Stolle R, van Harmelen F (eds) Proceedings of the AAAI 2020 Spring symposium on combining machine learning and knowledge engineering in practice, AAAI-MAKE 2020, Palo Alto, CA, USA, March 23-25, 2020, Volume I, CEUR Workshop Proceedings. CEUR-WS.org. http://ceur-ws.org/Vol-2600/paper5.pdf, vol 2600
Ebrahimi M, Sarker MK, Bianchi F, Xie N, Doran D, Hitzler P (2018) Reasoning over RDF knowledge bases using deep learning arXiv:1811.04132
Fung P, Wu C, Madotto A (2018) Mem2seq: Effectively incorporating knowledge bases into end-to-end task-oriented dialog systems. In: Gurevych I, Miyao Y (eds) Proceedings of the 56th annual meeting of the association for computational linguistics, ACL 2018, Melbourne, Australia, July 15-20, 2018, Volume 1: Long Papers, pp 1468–1478. Association for Computational Linguistics
d’Avila Garcez A, Lamb L, Gabbay DM (2009) Neural-symbolic cognitive reasoning. Springer, Heidelberg
d’Avila Garcez A, Besold TR, Raedt LD, Földiák P, Hitzler P, Icard T, Kühnberger K, Lamb L, Miikkulainen R, Silver DL (2015) Neural-symbolic learning and reasoning: Contributions and challenges. In: 2015 AAAI spring symposia, Stanford University, Palo Alto, California, USA, March 22-25, 2015. http://www.aaai.org/ocs/index.php/SSS/SSS15/paper/view/10281. AAAI Press
Grefenstette E (2013) Towards a formal distributional semantics: Simulating logical calculi with tensors. In: Second joint conference on lexical and computational semantics (* SEM), Volume 1: Proceedings of the main conference and the shared task: semantic textual similarity, pp 1–10
Grefenstette E, Hermann KM, Suleyman M, Blunsom P (2015) Learning to transduce with unbounded memory. In: Cortes C, Lawrence ND, Lee DD, Sugiyama M, Garnett R (eds) Advances in neural information processing systems 28: Annual conference on neural information processing systems 2015, December 7-12, 2015, Montreal, Quebec, Canada, pp 1828–1836
Gust H, Kühnberger K, Geibel P (2007) Learning models of predicate logical theories with neural networks based on topos theory. In: Hammer B, Hitzler P (eds) Perspectives of neural-symbolic integration, studies in computational intelligence, vol 77. Springer, pp 233–264
Hammer B, Hitzler P (eds) (2007) Perspectives of neural-symbolic integration, studies in computational intelligence, vol 77. Springer, Berlin
Hitzler P (2021) Semantic Web: A review of the field. Communications of the ACM. To appear
Hitzler P, Hölldobler S, Seda AK (2004) Logic programs and connectionist networks. J Appl Logic 2(3):245–272
Hitzler P, Krötzsch M, Parsia B, Patel-Schneider PF, Rudolph S OWL 2 Web Ontology Language: primer (Second Edition). W3C Recommendation 11 December 2012 (2012). Available from http://www.w3.org/TR/owl2-primer/
Hitzler P, Krötzsch M, Rudolph S (2010) Foundations of semantic web technologies. Chapman & Hall/CRC, Boca Raton
Hitzler P, Seda AK (2011) Mathematical aspects of logic programming semantics. Chapman and Hall / CRC studies in informatics series. CRC Press, Boca Raton
Hohenecker P, Lukasiewicz T (2020) Ontology reasoning with deep neural networks. J Artif Intell Res 68:503–540
Hölldobler S, Kalinke Y (1994) Ein massiv paralleles Modell für die Logikprogrammierung. In: WLP, pp 89–92
Kazakov Y, Krötzsch M, Simančík F (2012) Elk: a reasoner for OWL EL ontologies. System Description
Kiros R, Zhu Y, Salakhutdinov RR, Zemel R, Urtasun R, Torralba A, Fidler S (2015) Skip-thought vectors. In: Cortes C, Lawrence ND, Lee DD, Sugiyama M, Garnett R (eds) Advances in neural information processing systems 28. Curran Associates, Inc., pp 3294–3302
Klir G, Yuan B (1995) Fuzzy sets and fuzzy logic, vol 4, Prentice hall, New Jersey
Kulmanov M, Liu-Wei W, Yan Y, Hoehndorf R (2019) EL embeddings: geometric construction of models for the description logic EL++. In: Proceedings of the 28th International joint conference on artificial intelligence, AAAI Press, pp 6103–6109
Lao N, Mitchell T, Cohen WW (2011) Random walk inference and learning in a large scale knowledge base. In: Proceedings of the conference on empirical methods in natural language processing. Association for Computational Linguistics, pp 529–539
Le Q, Mikolov T (2014) Distributed representations of sentences and documents. In: International conference on machine learning, pp 1188–1196
Lin Y, Liu Z, Sun M, Liu Y, Zhu X (2015) Learning entity and relation embeddings for knowledge graph completion. In: Bonet B, Koenig S (eds) Proceedings of the Twenty-Ninth AAAI conference on artificial intelligence, January 25-30, 2015, Austin, Texas, USA, AAAI Press, pp 2181–2187
Ling W, Dyer C, Black AW, Trancoso I, Fermandez R, Amir S, Marujo L, Luís T (2015) Finding function in form: Compositional character models for open vocabulary word representation. In: Proceedings of the 2015 conference on empirical methods in natural language processing, pp 1520–1530
Makni B, Hendler J (2019) Deep learning for noise-tolerant RDFS reasoning. Semantic Web 10(5):823–862
McCarthy J (1988) Epistemological challenges for connectionism. Behav Brain Sci 44
McCulloch WS, Pitts W (1943) A logical calculus of the ideas immanent in nervous activity. Bull Math Biophys 5(4):115–133
Meza-Ruiz I, Riedel S (2009) Jointly identifying predicates, arguments and senses using Markov logic. In: Proceedings of human language technologies: The 2009 annual conference of the north american chapter of the association for computational linguistics, pp 155–163
Mikolov T, Sutskever I, Chen K, Corrado GS, Dean J (2013) Distributed representations of words and phrases and their compositionality. In: Advances in neural information processing systems, pp 3111–3119
Minervini P, Bošnjak M, Rocktäschel T, Riedel S, Grefenstette E (2020) Differentiable reasoning on large knowledge bases and natural language. In: Proceedings of the AAAI conference on artificial intelligence. https://doi.org/10.1609/aaai.v34i04.5962, vol 34, pp 5182–5190
Minervini P, Riedel S, Stenetorp P, Grefenstette E, Rocktäschel T (2020) Learning reasoning strategies in end-to-end differentiable proving. In: ICML
Neelakantan A, Roth B, McCallum A (2015) Compositional vector space models for knowledge base completion. In: Proceedings of the 53rd annual meeting of the association for computational linguistics and the 7th international joint conference on natural language processing (Volume 1: Long Papers), pp 156–166
Nguyen DQ, Nguyen DQ, Nguyen TD, Phung D (2019) A convolutional neural network-based model for knowledge base completion and its application to search personalization. Semantic Web 10(5):947–960
Pennington J, Socher R, Manning C (2014) Glove: Global vectors for word representation. In: Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pp 1532–1543
Raghu D, Gupta N (2018) Mausam: Hierarchical pointer memory network for task oriented dialogue. arXiv:1805.01216
Richardson M, Domingos P (2006) Markov logic networks. Mach Learn 62(1-2):107–136
Ristoski P, Paulheim H (2016) RDF2Vec: RDF graph embeddings for data mining. In: International semantic web conference. Springer, pp 498–514
Rocktȧschel T, Riedel S (2017) End-to-end differentiable proving. In: Guyon I, von Luxburg U, Bengio S, Wallach HM, Fergus R, Vishwanathan SVN, Garnett R (eds) Advances in neural information processing systems 30: Annual conference on neural information processing systems 2017, 4-9 December 2017, Long Beach, CA, USA, pp 3791–3803
Rocktäschel T, Riedel S (2017) End-to-end differentiable proving. In: Advances in neural information processing systems, pp 3788–3800
Rocktäschel T, Singh S, Riedel S (2015) Injecting logical background knowledge into embeddings for relation extraction. In: Proceedings of the 2015 conference of the north american chapter of the association for computational linguistics: human language technologies, pp 1119–1129
Serafini L, d’Avila Garcez A (2016) Logic tensor networks: Deep learning and logical reasoning from data and knowledge. arXiv:1606
Serafini L, d’Avila Garcez A (2016) Learning and reasoning with logic tensor networks. In: Adorni G, Cagnoni S, Gori M, Maratea M (eds) AI⋆A 2016: Advances in artificial intelligence – XVth international conference of the italian association for artificial intelligence, Genova, Italy, November 29 – December 1, 2016, Proceedings, Lecture Notes in Computer Science, vol 10037. Springer, pp 334–348
Serafini L, d’Avila Garcez A (2016) Logic tensor networks: Deep learning and logical reasoning from data and knowledge. In: Besold TR, Lamb LC, Serafini L, Tabor W (eds) Proceedings of the 11th international workshop on neural-symbolic learning and reasoning (NeSy’16) co-located with the Joint Multi-Conference on Human-Level Artificial Intelligence (HLAI 2016), New York City, NY, USA, July 16-17, 2016., CEUR Workshop Proceedings. CEUR-WS.org, vol 1768
Shastri L (1999) Advances in SHRUTI-A neurally motivated model of relational knowledge representation and rapid inference using temporal synchrony. Appl Intell 11(1):79–108
Shastri L (2007) SHRUTI: A neurally motivated architecture for rapid, scalable inference. In: Hammer B, Hitzler P (eds) Perspectives of neural-symbolic integration, studies in computational intelligence, vol 77. Springer, pp 183–203
Socher R, Chen D, Manning CD, Ng AY (2013) Reasoning with neural tensor networks for knowledge base completion. In: Burges CJC, Bottou L, Ghahramani Z, Weinberger KQ (eds) Advances in neural information processing systems 26: 27th annual conference on neural information processing systems 2013. Proceedings of a meeting held December 5-8, 2013, Lake Tahoe, Nevada, United States., pp 926–934
Sukhbaatar S, Szlam A, Weston J, Fergus R (2015) End-to-end memory networks. In: Cortes C, Lawrence ND, Lee DD, Sugiyama M, Garnett R (eds) Advances in neural information processing systems 28: annual conference on neural information processing systems 2015, December 7-12, 2015, Montreal, Quebec, Canada, pp 2440–2448
Towell GG, Shavlik JW (1994) Knowledge-based artificial neural networks. Artif Intell 70 (1-2):119–165
Trouillon T, Welbl J, Riedel S, Gaussier E, Bouchard G (2016) Complex embeddings for simple link prediction. In: Balcan M, Weinberger KQ (eds) Proceedings of the 33nd International conference on machine learning, ICML 2016, New York City, NY, USA, June 19-24, 2016, JMLR Workshop and Conference Proceedings. JMLR.org, vol 48, pp 2071–2080
Wang Q, Mao Z, Wang B, Guo L (2017) Knowledge graph embedding: A survey of approaches and applications. IEEE Trans Knowl Data Eng 29(12):2724–2743
Wang WY, Cohen WW (2016) Learning first-order logic embeddings via matrix factorization. In: IJCAI, pp 2132–2138
Wang Z, Zhang J, Feng J, Chen Z (2014) Knowledge graph embedding by translating on hyperplanes. In: Brodley CE, Stone P (eds) Proceedings of the Twenty-Eighth AAAI Conference on artificial intelligence, July 27 -31, 2014, Quėbec City, Quėbec, Canada., AAAI Press, pp 1112–1119
Weston J, Chopra S, Bordes A (2014) Memory networks. arXiv:1410.3916. Published as ICLR 2015 conference paper
Wikibooks contributors: Algorithm implementation/strings/levensh- tein distance (2019 (accessed November 19, 2019)). https://en.wikibooks.org/wiki/Algorithm_Implementation/Strings/Levenshtein_distance#Python
Yang B, Yih W, He X, Gao J, Deng L (2014) Embedding entities and relations for learning and inference in knowledge bases. arXiv:1412.6575
Yang F, Yang Z, Cohen WW (2017) Differentiable learning of logical rules for knowledge base reasoning. In: Advances in neural information processing systems, pp 2319–2328
Acknowledgements
This work was supported by the Air Force Office of Scientific Research under award number FA9550-18-1-0386 and by the National Science Foundation (NSF) under award OIA-2033521 “KnowWhereGraph: Enriching and Linking Cross-Domain Knowledge Graphs using Spatially-Explicit AI Technologies.”
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This article belongs to the Topical Collection: 30th Anniversary Special Issue
Rights and permissions
About this article
Cite this article
Ebrahimi, M., Eberhart, A., Bianchi, F. et al. Towards bridging the neuro-symbolic gap: deep deductive reasoners. Appl Intell 51, 6326–6348 (2021). https://doi.org/10.1007/s10489-020-02165-6
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10489-020-02165-6