Differences

This shows you the differences between two versions of the page.

--- bibtex [2022/01/07 18:37]
anton
+++ bibtex [2022/01/07 18:44]
anton
@@ Line 1: / Line 1: @@
+@CONFERENCE{Sychev2021728,
+author={Sychev, O.},
+title={Combining neural networks and symbolic inference in a hybrid cognitive architecture},
+journal={Procedia Computer Science},
+year={2021},
+volume={190},
+pages={728-734},
+doi={10.1016/j.procs.2021.06.085},
+url={https://www.sciencedirect.com/science/article/pii/S1877050921013405?via%3Dihub},
+affiliation={Volgograd State Technical University, Lenin Ave, 28, Volgograd, 400005, Russian Federation},
+abstract={Recently, there has been a big progress in developing artificial deep-learning neural networks and large-scale knowledge graphs. However, the results in these two research fields have serious drawbacks. The solutions offered by neural networks remain unstable and prone to adversarial attacks: while the percentage of correct answers increases, the incorrect answers often contain glaring errors. Large knowledge graphs contain a lot of facts but little knowledge; they are mostly used for information search and retrieval. The ability to reason conclusions on them is limited, and the majority of modern research turns to approximate methods like neural networks and graph embeddings to draw conclusions and use the accumulated knowledge. In this work, I propose a hybrid cognitive architecture inspired by the observable features of human thinking. Pruning obviously wrong solutions seems to be more natural for human symbolic reasoning than making far-fetched strict logical conclusions, while generating new ideas is often intuitive. So a hybrid cognitive architecture can employ generative neural networks as a sort of "intuition" (generating possible solutions) and symbolic inference as a control contour to verify and filter the generated solutions, weeding out dangerous and wrong ideas. This requires creating knowledge graphs containing negative information: the information of what cannot happen or should not be done and why. The problems of creating negative knowledge graphs are discussed. Hybrid cognitive systems using the proposed architecture will be a lot more trustworthy as they will have a system of human-verifiable rules that ensures avoiding the worst errors which can be used in many fields from decision making to natural-language parsing. © 2020 Elsevier B.V.. All rights reserved.},
+correspondence_address1={Sychev, O.; Volgograd State Technical University, Lenin Ave, 28, Russian Federation; эл. почта: oasychev@gmail.com},
+publisher={Elsevier B.V.},
+issn={18770509},
+language={English},
+abbrev_source_title={Procedia Comput. Sci.},
+thanks = {rfbr-20-07-00764}
+}
+@ARTICLE{Anikin2021_3,
+author={Anikin, A. and Sychev, O. and Denisov, M.},
+title={Ontology reasoning for explanatory feedback generation to teach how algorithms work},
+journal={Frontiers in Artificial Intelligence and Applications},
+year={2021},
+volume={338},
+pages={V-VI},
+doi={10.3233/FAIA210100},
+art_number={239-244},
+url={https://ebooks.iospress.nl/doi/10.3233/FAIA210100},
+affiliation={Volgograd State Technical University, Russian Federation; Software Engineering School, Volgograd, Russian Federation},
+abstract={Developing algorithms using control structures and understanding their building blocks are essential skills in mastering programming. Ontologies and software reasoning is a promising method for developing intelligent tutoring systems in well-defined domains (like programming languages and algorithms); it can be used for many kinds of teaching tasks. In this work, we used a formal model consisting of production rules for Apache Jena reasoner as a basis for developing a constraint-based tutor for introductory programming domain. The tutor can determine fault reasons for any incorrect answer that a student can enter. The problem the student should solve is building an execution trace for the given algorithm. The problem is a closed-ended question that requires arranging given actions in the (unique) correct order; some actions can be used several times, while others can be omitted. Using formal reasoning to check domain constraints allowed us to provide explanatory feedback for all kinds of errors students can make. © 2021 The authors and IOS Press.},
+editor={Frasson C., Kabassi K., Voulodimos A.},
+publisher={IOS Press BV},
+issn={09226389},
+isbn={9781643682044},
+language={English},
+abbrev_source_title={Front. Artif. Intell. Appl.},
+thanks = {rfbr-20-07-00764}
+}
+@CONFERENCE{Sychev2021_2,
+author={Sychev, O. and Anikin, A. and Terekhov, G.},
+title={Demonstrating Concepts through Visual Simulators: Two Cases in the Programming Domain},
+journal={Proceedings of IEEE Symposium on Visual Languages and Human-Centric Computing, VL/HCC},
+year={2021},
+volume={2010-October},
+doi={10.1109/VL/HCC51201.2021.9576398},
+url={https://ieeexplore.ieee.org/document/9576398},
+affiliation={Volgograd State Technical University, Software Engineering Department, Volgograd, Russian Federation},
+abstract={One of the ways to demonstrate subject-domain concepts is to let the learner play in the domain-related sandbox, while receiving the explanations of the domain laws that were broken when the user makes a wrong move. This can help the learners who understand definitions of the concepts poorly. We present two visual simulators for demonstrating concept properties and domain laws, implemented as web applications. They can be used to study programming, enabling trial-and-error learning, supported by the error messages, explaining why the user's action was wrong. © 2021 IEEE.},
+editor={Harms K., Cunha J., Oney S., Kelleher C.},
+publisher={IEEE Computer Society},
+issn={19436092},
+isbn={9781665445924},
+language={English},
+abbrev_source_title={Proc. of IEEE Symp. Vis. Lang. Hum.-Cent. Comput., VL/HCC},
+thanks = {rfbr-20-07-00764}
+}
 @ARTICLE{Sychev2021mdpi,
 author={Oleg Sychev and Nikita Penskoy and Anton Anikin and Mikhail Denisov and Artem Prokudin},