TCL Logic (Typicality-based Composotional Logic) for commonsense conceptual combination and blending by Antonio Lieto and Gian Luca Pozzato
TCL Logic (Typicality-based Compositional Logic) by Antonio Lieto and Gian Luca Pozzato
Below a list of selected, peer reviewed, publications about TCL and its applications (from 2018 to 2021). Full list here.
Back to the home page
TCL (Typicality-based Compositional Logic) is the first ever developed formal (i.e. logic-based) account able to model - with a unique formalism - the problem of both human-like NOUN-NOUN commonsense conceptual combination (i.e. by solving the so-called PET FISH problem, also known as guppy effect) as well as the problem known as conceptual blending (including hierarchical and iterated blending).
This logic integrates a non monotonic description logic of typicality, the probabilistic semantics called DISPONTE and the HEAD-MODIFIER heuristics (coming from cognitive semantics).
TCL has been applied to a number of applications ranging from cognitive modelling (e.g. pet-fish problem, the conjunction fallacy and goal-reasoning heuristics) to computational creativity and multimedia and emotion-oriented recommendations.
The papers introducing the TCL reasoning framework (2018-2020)
Antonio Lieto, Gian Luca Pozzato "A Description Logic of Typicality for Conceptual Combination", in Proceedings of ISMIS 2018, International Symposium on Methodologies for Intelligent Systems, pp.189-199, 2018. This paper introduces TCL and show how it is able to address the problem of NOUN-NOUN commmonsense conceptual combination (PET-FISH problem).
Antonio Lieto, Gian Luca Pozzato "A Description Logic Framework for Commonsense Conceptual Combination Integrating Typicality, Probabilities and Cognitive Heuristics", in Journal of Experimental & Theoretical Artificial Intelligence (JETAI), 32 (5), pp. 769-804, 2020. This paper extends the semantics of TCL and show how it is able to address also other cognitive modelling problems like: the conjunction fallacy, the commonsense combination of multiple concepts (more than 2), the problem of concept invention via conceptual blending as well as the problem of hierarchical and iterated concepts generation in a knowledge base.
Multimedia, Musical and Explainable-Emotion-oriented Creative Recommendations with TCL (2020-2023)
Antonio Lieto, Gian Luca Pozzato, Manuel Striani, Stefano Zoia and Rossana Damiano, DEGARI 2.0: A Diversity-Seeking, Knowledge-Based, Explainalable, and Affective Art Recommender for Social Inclusion, Cognitive Systems Research, 77, 2022, 1-17. Preprint: https://arxiv.org/abs/2101.04017. This paper presents a novel version of DEGARI (Dynamic Emotion Generator And ReclassIfier), an explainable system for emotion attribution & recommendation that uses the TCL logic to generate complex emotions (obtained via the commonsense conceptual combination of basic ones) according to the Plutchik's emotion theory. In this version, the refclassification of items based on the novel generated emotions is ussd to make different types of item recommendations. A short DEMO of DEGARI 2.0 at work in the context of the H2020 SPICE Project is available at https://www.youtube.com/watch?v=uwciHMRLRE8&t=6s.
Antonio Lieto, Gian Luca Pozzato, Stefano Zoia, Viviana Patti and Rossana Damiano, A commonsense reasoning framework for explanatory emotion attribution, generation and re-classification, Knowledge-Based Systems, Volume 227, 107166, 2021. ArXiv preprint: https://arxiv.org/abs/2101.04017. This paper presents DEGARI (Dynamic Emotion Generator And ReclassIfier), an explainable system for emotion attribution & recommendation that uses the TCL logic to generate complex emotions (obtained via the commonsense conceptual combination of basic ones) according to the Plutchik's emotion theory. The novel generated emotions are then used to classify (or reclassify) artistic items.
Eleonora Chiodino, Davide Di Luccio, Antonio Lieto, Alberto Messina, Gian Luca Pozzato, Davide Rubinetti "A Knowledge-based System for the Dynamic Generationand Classification of Novel Contents in Multimedia Broadcasting", in Proceedings of ECAI 2020, 24th European Conference on Artificial Intelligence, 2020. This paper presents DENOTER, a content generator and content suggestion system applied in the context of the RaiPlay platform. It exploits the TCL Commonsense Logical Framework and generates a novel type of "combinatorial" and serendipity-seeking content recommendations. Below the presentation done at ECAI 2020.
Cognitive Modelling Applications (Goal-Oriented Reasoning, Dynamic Knowledge Generation and Creative Problem Solving), (2019-2020)
Antonio Lieto, Federico Perrone, Gian Luca Pozzato and Eleonora Chiodino "Beyond Subgoaling: A Dynamic Knowledge Generation Framework for Creative Problem Solving in Cognitive Architectures", Cognitive Systems Research, 58, 305-316, 2019. This paper shows how TCL allows a cognitive agent to have dynamic knowledge base where new knowledge is generated via commonsense concept combination in a goal-oriented and problem solving perspective. Such perspective allows the agent to generate creative problem-solving solutions, with human-comparable performances, in the context of object invention. A short demo of the system described in the paper for the task of objects invention via dynamic concept combination is available below.
Antonio Lieto and Gian Luca Pozzato "Applying a description logic of typicality as a generative tool for concept combination in computational creativity", Intelligenza Artificiale, vol. 13, no. 1, pp. 93-106, 2019.
Softwares, Reasoners and Tools (from 2018 to 2021)
DEGARI (Dynamic Emotion Generator And ReclassIfier) , 2021, (developers: Antonio Lieto, Gian Luca Pozzato, Stefano Zoia). Available on Github at https://github.com/alieto/DEGARI
DENOTER (Dynamic gEnerator of NOvel contents in mulTimEdia bRroadcasting, 2020, (developers: Davide Di Luccio, Antonio Lieto, Gian Luca Pozzato, Davide Rubinetti). Available at https://di.unito.it/DENOTER.
GOCCIOLA (Generating knOwledge by Concept Combination In descriptiOn Logics of typicAlity), 2019, (developers: Antonio Lieto, Federico Perrone, Gian Luca Pozzato). Available at http://di.unito.it/gocciola.
COCOS (A typicality based COncept COmbination System), 2018, (developers: Antonio Lieto, Gian Luca Pozzato, Alberto Valese). Available at http://di.unito.it/cocos.
The COCOS reasoner is described in Antonio Lieto, Gian Luca Pozzato, Alberto Valese "COCOS: a typicality based COncept COmbination System", in Proc. of CILC 2018, 33rd Italian Conference on Computational Logic. COCOS is used - as a software component - in all the other systems mentioned above.