NeSy 2024

The NeSy event series celebrates neurosymbolic AI since 2005 when the first NeSy workshop took place in Edinburgh, Scotland, as part of IJCAI2005. NeSy is about the integration of neural and symbolic AI, combining statistical machine learning based on neural networks with knowledge representation and reasoning from symbolic AI. 

Neural networks and statistical Machine Learning have achieved industrial relevance in a number of areas from healthcare to finance and business, obtaining state-of-the-art performance at language modelling, speech and image recognition, sensor data and graph analytics. Symbolic AI is challenged by such unstructured large data, but offers sound and well-understood formal reasoning and explanation via knowledge representation that can be inspected to interpret how decisions follow from data. Neural and symbolic AI approaches also contrast in the problems that they excel at: deep learning excels at scene recognition, but fails at planning and rich deductive reasoning.

Neurosymbolic AI aims to build rich computational AI models, systems and applications by combining neural and symbolic learning and reasoning. It hopes to create synergies among the strengths of neural and symbolic AI while overcoming their complementary weaknesses. The NeSy workshop series is the premier venue for the presentation and discussion of the theory and practice of neurosymbolic computing. Since NeSy 2005 the workshop provided an atmosphere for the free exchange of ideas bringing together the community of scientists and practitioners that straddle the line between connectionism and symbolic AI.

Keynote talks (in person)

• Catia Pesquita (University of Lisbon, Portugal)

• Kristian Kersting (TU Darmstadt, Germany)  

• Willem (Jelle) Zuidema (University of Amsterdam, The Netherlands)

Format and Presentation

NeSy2024 will include keynote talks, special sessions with invited talks and poster presentations, recently-published conference and journal paper presentations, specialised NeSy paper discussions and presentations, and audience discussion and social events to allow the group to meet, discuss and obtain a better understanding of the issues, challenges and ideas being presented. Authors of accepted papers may be assigned either an oral or poster presentation slot in the final programme. Accepted papers will be published by Springer. Registration will be open to anybody willing to participate and is mandatory for one author of each accepted paper.

Venue and Registration

The conference took place in person at the Residència d'Investigadors in the historic center of Barcelona (Catalonia, Spain). As in previous editions, NeSy2024 put emphasis on making attendance fees affordable for early career researchers and community members from less financially affluent institutions. 

Call for NeSy2024 Paper Submissions (closed)

NeSy invites submissions of the latest and ongoing research work on neurosymbolic AI for presentation at the conference. Research papers in any of the areas of neurosymbolic computing listed below are welcome.

Publication

All accepted papers will be published by Springer LNCS and are expected to be presented at the conference. Accepted papers will be chosen for spotlight oral and poster presentations. Revised and extended versions of the best papers will be invited for submission to the Neurosymbolic AI journal.

Topics of Interest

NeSy invites theoretical and applied paper submissions combining neural networks and symbolic AI. We further invite paper submissions containing experimental and in-the-wild systems having neurosymbolic computing as a strong use case. Topics of interest include, but are not limited to: 

• Knowledge representation and reasoning using deep neural networks;

• Symbolic knowledge extraction from neural and statistical learning systems;

• Explainable AI methods, systems and techniques integrating connectionist and symbolic AI;

• Explainability of Large Language Models and multimodality;

• Enhancing deep learning systems through structured background knowledge.

• Neurosymbolic cognitive agents and biologically-inspired neurosymbolic integration;

• Integration of logics and probabilities in neural networks;

• Neurosymbolic methods for structure learning, transfer learning, meta, multi-task and continual learning, relational learning, graph neural networks;

• Novel connectionist systems able to perform traditionally symbolic AI tasks (e.g. abduction, deduction, out-of-distribution learning);

• Novel symbolic systems able to perform traditionally connectionist tasks (e.g. learning from unstructured data, distributed learning);

• Embedding methods for structured information, such as knowledge graphs, mathematical expressions, grammars, knowledge bases, logical theories;

• Applications of neurosymbolic and hybrid systems, including in simulation, finance, healthcare, robotics, semantic web, software engineering, systems engineering, bioinformatics and visual intelligence.

Special Tracks

NeSy2024 hosts a special track on Explainable AI (XAI) and Generative AI (GenAI). Contributions targeting either of the two topics should be submitted to the corresponding special track on EasyChair. See the respective section below for details on the focus of the tracks.

Full Paper Submission

Researchers and practitioners are invited to submit original papers that are not being reviewed or published elsewhere. Submitted papers need not be anonymous, must be written in English, should be formatted using single column and 11pt font, and should not exceed 12 pages in the case of research and experience papers, or 6 pages in the case of position papers, including all figures, but excluding references and appendices. Reviewers will not be asked to read or evaluate the appendix. You should use the LNCS Latex template or Word template.

All submitted papers will be judged based on their relevance, originality, significance, technical quality and organisation. 

Please submit your papers via EasyChair. All questions about submissions should be addressed to Ernesto Jimenez Ruiz at ernesto.jimenez-ruiz@city.ac.uk

Important Dates for Full Paper Submission (Anywhere on Earth)

• Abstract submission: 21 April 2024

• Paper submission deadline: 26 April 2024 

• Author Notification: 7 June 2024

• Camera-ready paper due: 21 June 2024

Late/Short Paper Submission

Researchers and practitioners are invited to submit original late/short papers that are not being reviewed or published elsewhere. Submitted late/short papers need not be anonymous, must be written in English, should be formatted using single column and 11pt font, and should not exceed 6 pages, including all figures, but excluding references and appendices. Reviewers will not be asked to read or evaluate the appendix. You should use the LNCS Latex template or Word template.

All submitted papers will be judged based on their relevance, originality, significance, technical quality and organization. Please submit your papers via EasyChair. All questions about submissions should be addressed to Ernesto Jimenez Ruiz at ernesto.jimenez-ruiz@city.ac.uk

Important Dates for Late/Short Paper Submission (Anywhere on Earth)

• Paper submission deadline: 15 May 2024 

• Author Notification: 14 June 2024

• Camera-ready paper due: 21 June 2024

Recently-published paper presentations (closed) 

Authors of highly relevant, recently published papers in neurosymbolic AI are invited to submit proposals to present their papers at NeSy2024. This includes papers published recently at top conferences and journals such as JAIR, MLJ, AIJ, IEEE TNNLS, AAAI, IJCAI, NeurIPS and ICML. 

You are asked to submit a 2-page (including references, acknowledgements, etc.) extended abstract including a reference to the published paper using the CEUR Latex template or the CEUR Word template. Please submit your extended abstract in PDF format via EasyChair. At least one author is expected to register and attend NeSy to present the paper. Extended abstracts will be published on the NeSy2024 website but will not be included in the official LNCS proceedings.  

Important Dates (Anywhere on Earth):

• Extended abstract submission deadline: 7 July 2024

• Author Notification: 15 July 2024

Questions about recently-published paper submissions should be addressed to Tarek R. Besold at tarek.besold@sony.com.

NeSy2024 Special Track on Explainable AI

Current approaches to Explainable AI primarily concentrate on the mechanistic aspects of generating explanations, often lacking well-founded justifications and logical reasoning for the decisions made. This frequently results in explanations that are not easily understandable or useful for users. Neurosymbolic AI strives to provide a principled approach to the integration of learning and reasoning, seeking to establish connections between neural models and logical representations. Through this neurosymbolic integration, novel forms of explanations can be generated. Learning symbolic explanations can simulate reasoning scenarios that incorporate formal approaches to human and common-sense reasoning, e.g. analogical reasoning, concept refinement, and computational argumentation. These techniques aim to produce explanations that align more closely with human thought processes, thereby enhancing the perceived interpretability of explanations for humans. 

The XAI special track at NeSy2024 focuses on essential aspects of neurosymbolic Explainable AI, inviting contributions that explore new approaches to learning and reasoning about explanations, along with novel theories and methods for representing, manipulating, and evaluating explanations. Beyond advancing algorithmic and methodological aspects, contributions are encouraged to delve into the ethical implications and societal impacts of deploying AI in real-world applications, with a particular emphasis on methods that promote fairness and accountability.

NeSy2024 Special Track on Generative AI

Generative AI has witnessed remarkable progress in generating diverse and realistic content across domains such as text, images, music, and video. The emergence of large language models (LLMs) and diffusion models trained on massive datasets is reshaping AI landscape. However, ensuring desired features such as controllability and reasoning capabilities in generative models remains a significant challenge. LLMs especially have shown promising results in the broad area of Natural Language Processing, but their reasoning capabilities and “understanding” of symbolic meaning remain uncertain. Symbol manipulation plays a vital role in assessing LLMs' reasoning abilities and their integration with classical computational systems in real-world applications. Additionally, Neuro Symbolic approaches address biases in training data, and reasoning about causality and temporal dependencies. The integration of symbolic reasoning and cognitive processes in AI systems is crucial for understanding complex scenarios. 

Community Slack and Mailing List

The NeSy association has a community slack for Neurosymbolic Artificial Intelligence with more than 1000 participants. Please contact Pascal Hitzler at hitzler@ksu.edu to receive an invitation if you wish to join. 

Please also consider joining the (low traffic) mailing list of the NeSy association here. 

Sponsorship