Federico Sabbatini, Giovanni Ciatto, Roberta Calegari, Andrea Omicini
The XAI community is currently studying and developing symbolic knowledge-extraction (SKE) algorithms as a means to produce human-intelligible explanations for black-box machine learning predictors, so as to achieve believability in human-machine interaction. However, many extraction procedures exist in the literature, and choosing the most adequate one is increasingly cumbersome, as novel methods keep on emerging. Challenges arise from the fact that SKE algorithms are commonly defined based on theoretical assumptions that typically hinder practical applicability. This paper focuses on hypercube-based SKE methods, a quite general class of extraction techniques mostly devoted to regression-specific tasks. We first show that hypercube-based methods are flexible enough to support classification problems as well, then we propose a general model for them, and discuss how they support SKE on datasets, predictors, or learning tasks of any sort. Empirical examples are reported as well –based upon the PSyKE framework –, showing the applicability of hypercube-based methods to actual classification tasks.
{"title":"Towards a unified model for symbolic knowledge extraction with hypercube-based methods","authors":"Federico Sabbatini, Giovanni Ciatto, Roberta Calegari, Andrea Omicini","doi":"10.3233/IA-230001","DOIUrl":"https://doi.org/10.3233/IA-230001","url":null,"abstract":"The XAI community is currently studying and developing symbolic knowledge-extraction (SKE) algorithms as a means to produce human-intelligible explanations for black-box machine learning predictors, so as to achieve believability in human-machine interaction. However, many extraction procedures exist in the literature, and choosing the most adequate one is increasingly cumbersome, as novel methods keep on emerging. Challenges arise from the fact that SKE algorithms are commonly defined based on theoretical assumptions that typically hinder practical applicability. This paper focuses on hypercube-based SKE methods, a quite general class of extraction techniques mostly devoted to regression-specific tasks. We first show that hypercube-based methods are flexible enough to support classification problems as well, then we propose a general model for them, and discuss how they support SKE on datasets, predictors, or learning tasks of any sort. Empirical examples are reported as well –based upon the PSyKE framework –, showing the applicability of hypercube-based methods to actual classification tasks.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"17 1","pages":"63-75"},"PeriodicalIF":1.5,"publicationDate":"2023-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48828622","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}
The first Workshop “From Objects to Agents” (WOA) was held in Parma in May 2000. The workshop started as a joint initiative of the Agents and Multi-Agent Systems Working Group of the Italian Association for Artificial Intelligence (MAS-AIxIA) together with the Italian Association for Advanced Technologies based on Object-Oriented Concepts (TABOO). The workshop was meant to provide a forum for researchers and practitioners interested in understanding the possibilities that the intricate connection between agent technologies and object-oriented technologies could open. The first WOA counted more than fifty registered participants from both the academia and the software indus-try. In the years, MAS-AIxIA took full charge of the workshop, which shifted its focus towards all topics related to agents and multi-agent systems, and became a stand-alone initiative with an international perspective organised by an independent community of researchers and practitioners based in Italy. As such, the workshop has always been located in Italy, with the workshop Steering Committee constantly committed to involve every major Italian research group working on agents and multi-agent systems. The workshop was hosted in the following venues (in alphabetical order): Bologna
{"title":"Special issue for the 23rd workshop \"from objects to agents\" (WOA 2022)","authors":"Angelo Ferrando, V. Mascardi","doi":"10.3233/ia-230015","DOIUrl":"https://doi.org/10.3233/ia-230015","url":null,"abstract":"The first Workshop “From Objects to Agents” (WOA) was held in Parma in May 2000. The workshop started as a joint initiative of the Agents and Multi-Agent Systems Working Group of the Italian Association for Artificial Intelligence (MAS-AIxIA) together with the Italian Association for Advanced Technologies based on Object-Oriented Concepts (TABOO). The workshop was meant to provide a forum for researchers and practitioners interested in understanding the possibilities that the intricate connection between agent technologies and object-oriented technologies could open. The first WOA counted more than fifty registered participants from both the academia and the software indus-try. In the years, MAS-AIxIA took full charge of the workshop, which shifted its focus towards all topics related to agents and multi-agent systems, and became a stand-alone initiative with an international perspective organised by an independent community of researchers and practitioners based in Italy. As such, the workshop has always been located in Italy, with the workshop Steering Committee constantly committed to involve every major Italian research group working on agents and multi-agent systems. The workshop was hosted in the following venues (in alphabetical order): Bologna","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"17 1","pages":"3-5"},"PeriodicalIF":1.5,"publicationDate":"2023-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46652714","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}
To the aim of constructing effective human-AI teams, that can be useful for improving caregiving in medicine and enhancing human performance also in other sectors (i.e., teaching), agents which interact with humans should be endowed with an emotion recognition and management module, capable of empathy, and of modeling aspects of the Theory of Mind, in the sense of being able to reconstruct what someone is thinking or feeling. In this paper, we propose an architecture for such a module, based upon an enhanced notion of Behavior Trees. We illustrate the effectiveness of the proposed architecture on a significant example and on a wider case study.
{"title":"Empathetic human-agent interaction via emotional behavior trees","authors":"Pierangelo Dell'Acqua, S. Costantini","doi":"10.3233/IA-230014","DOIUrl":"https://doi.org/10.3233/IA-230014","url":null,"abstract":"To the aim of constructing effective human-AI teams, that can be useful for improving caregiving in medicine and enhancing human performance also in other sectors (i.e., teaching), agents which interact with humans should be endowed with an emotion recognition and management module, capable of empathy, and of modeling aspects of the Theory of Mind, in the sense of being able to reconstruct what someone is thinking or feeling. In this paper, we propose an architecture for such a module, based upon an enhanced notion of Behavior Trees. We illustrate the effectiveness of the proposed architecture on a significant example and on a wider case study.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"17 1","pages":"89-100"},"PeriodicalIF":1.5,"publicationDate":"2023-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44751629","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}
G. Bella, Domenico Cantone, Carmelo Fabio Longo, Marianna Nicolosi Asmundo, Daniele Francesco Santamaria
Semantic representation is a key enabler for several application domains, and the multi-agent systems realm makes no exception. Among the methods for semantically representing agents, one has been essentially achieved by taking a behaviouristic vision, through which one can describe how they operate and engage with their peers. The approach essentially aims at defining the operational capabilities of agents through the mental states related with the achievement of tasks. The OASIS ontology — An Ontology for Agent, Systems, and Integration of Services, presented in 2019 — pursues the behaviouristic approach to deliver a semantic representation system and a communication protocol for agents and their commitments. This paper reports on the main modelling choices concerning the representation of agents in OASIS 2, the latest major upgrade of OASIS, and the achievement reached by the ontology since it was first introduced, in particular in the context of ontologies for blockchains.
{"title":"The Ontology for Agents, Systems and Integration of Services: OASIS version 2","authors":"G. Bella, Domenico Cantone, Carmelo Fabio Longo, Marianna Nicolosi Asmundo, Daniele Francesco Santamaria","doi":"10.3233/IA-230002","DOIUrl":"https://doi.org/10.3233/IA-230002","url":null,"abstract":" Semantic representation is a key enabler for several application domains, and the multi-agent systems realm makes no exception. Among the methods for semantically representing agents, one has been essentially achieved by taking a behaviouristic vision, through which one can describe how they operate and engage with their peers. The approach essentially aims at defining the operational capabilities of agents through the mental states related with the achievement of tasks. The OASIS ontology — An Ontology for Agent, Systems, and Integration of Services, presented in 2019 — pursues the behaviouristic approach to deliver a semantic representation system and a communication protocol for agents and their commitments. This paper reports on the main modelling choices concerning the representation of agents in OASIS 2, the latest major upgrade of OASIS, and the achievement reached by the ontology since it was first introduced, in particular in the context of ontologies for blockchains.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"17 1","pages":"51-62"},"PeriodicalIF":1.5,"publicationDate":"2023-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45155287","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}
We consider an extended version of sabotage games played over Attack Graphs. Such games are two-player zero-sum reachability games between an Attacker and a Defender. This latter player can erase particular subsets of edges of the Attack Graph. To reason about such games we introduce a variant of Sabotage Modal Logic (that we call Subset Sabotage Modal Logic) in which one modality quantifies over non-empty subset of edges. We show that we can characterize the existence of winning Attacker strategies by formulas of Subset Sabotage Modal Logic.
{"title":"Attack Graphs & Subset Sabotage Games","authors":"Davide Catta, J. Leneutre, Vadim Malvone","doi":"10.3233/IA-221080","DOIUrl":"https://doi.org/10.3233/IA-221080","url":null,"abstract":" We consider an extended version of sabotage games played over Attack Graphs. Such games are two-player zero-sum reachability games between an Attacker and a Defender. This latter player can erase particular subsets of edges of the Attack Graph. To reason about such games we introduce a variant of Sabotage Modal Logic (that we call Subset Sabotage Modal Logic) in which one modality quantifies over non-empty subset of edges. We show that we can characterize the existence of winning Attacker strategies by formulas of Subset Sabotage Modal Logic.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"17 1","pages":"77-88"},"PeriodicalIF":1.5,"publicationDate":"2023-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43443338","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}
In this paper, we provide a unified presentation of the Configurable Markov Decision Process (Conf-MDP) framework. A Conf-MDP is an extension of the traditional Markov Decision Process (MDP) that models the possibility to configure some environmental parameters. This configuration activity can be carried out by the learning agent itself or by an external configurator. We introduce a general definition of Conf-MDP, then we particularize it for the cooperative setting, where the configuration is fully functional to the agent’s goals, and non-cooperative setting, in which agent and configurator might have different interests. For both settings, we propose suitable solution concepts. Furthermore, we illustrate how to extend the traditional value functions for MDPs and Bellman operators to this new framework.
{"title":"A unified view of configurable Markov Decision Processes: Solution concepts, value functions, and operators","authors":"A. Metelli","doi":"10.3233/ia-220140","DOIUrl":"https://doi.org/10.3233/ia-220140","url":null,"abstract":"In this paper, we provide a unified presentation of the Configurable Markov Decision Process (Conf-MDP) framework. A Conf-MDP is an extension of the traditional Markov Decision Process (MDP) that models the possibility to configure some environmental parameters. This configuration activity can be carried out by the learning agent itself or by an external configurator. We introduce a general definition of Conf-MDP, then we particularize it for the cooperative setting, where the configuration is fully functional to the agent’s goals, and non-cooperative setting, in which agent and configurator might have different interests. For both settings, we propose suitable solution concepts. Furthermore, we illustrate how to extend the traditional value functions for MDPs and Bellman operators to this new framework.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"16 1","pages":"165-184"},"PeriodicalIF":1.5,"publicationDate":"2022-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"48339477","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}
Andrea Loreggia, Roberta Calegari, E. Lorini, Francesca Rossi, G. Sartor
Preferences are ubiquitous in our everyday life. They are essential in the decision making process of individuals. Recently, they have also been employed to represent ethical principles, normative systems or guidelines. In this work we focus on a ceteris paribus semantics for deontic logic: a state of affairs where a larger set of respected prescriptions is preferable to a state of affairs where some are violated. Conditional preference networks (CP-nets) are a compact formalism to express and analyse ceteris paribus preferences, with some desirable computational properties. In this paper, we show how deontic concepts (such as contrary-to-duty obligations) can be modeled with generalized CP-nets (GCP-nets) and how to capture the distinction between strong and weak permission in this formalism. To do that, we leverage on an existing restricted deontic logic that will be mapped into conditional preference nets.
{"title":"How to model contrary-to-duty with GCP-nets","authors":"Andrea Loreggia, Roberta Calegari, E. Lorini, Francesca Rossi, G. Sartor","doi":"10.3233/ia-221057","DOIUrl":"https://doi.org/10.3233/ia-221057","url":null,"abstract":"Preferences are ubiquitous in our everyday life. They are essential in the decision making process of individuals. Recently, they have also been employed to represent ethical principles, normative systems or guidelines. In this work we focus on a ceteris paribus semantics for deontic logic: a state of affairs where a larger set of respected prescriptions is preferable to a state of affairs where some are violated. Conditional preference networks (CP-nets) are a compact formalism to express and analyse ceteris paribus preferences, with some desirable computational properties. In this paper, we show how deontic concepts (such as contrary-to-duty obligations) can be modeled with generalized CP-nets (GCP-nets) and how to capture the distinction between strong and weak permission in this formalism. To do that, we leverage on an existing restricted deontic logic that will be mapped into conditional preference nets.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"16 1","pages":"185-198"},"PeriodicalIF":1.5,"publicationDate":"2022-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47572997","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}
Roberta Calegari, Giovanni Ciatto, Andrea Omicini, Giuseppe Vizzari
Roberta Calegaria,∗, Giovanni Ciattob, Andrea Omicinib and Giuseppe Vizzaric aAlma Mater Research Institute for Human-Centered Artificial Intelligence (AlmaAI), ALMA MATER STUDIORUM—Università di Bologna, Italy bDipartimento di Informatica — Scienza e Ingegneria (DISI), ALMA MATER STUDIORUM—Università di Bologna, Italy cDipartimento di Informatica, Sistemistica e Comunicazione, Università degli Studi di Milano–Bicocca, Milano, Italy
罗伯塔Calegaria∗、约翰·Ciattob,安德里亚·朱塞佩·Vizzaric Omicinib和脱线aAlma Research Institute for人为中心Artificial Intelligence (AlmaAI),母校STUDIORUM计算机—博洛尼亚大学,意大利bDipartimento—科学和工程(旅),母校STUDIORUM计算机—博洛尼亚大学,意大利cDipartimento Sistemistica和通讯、universita degli Studi di米兰Bicocca—米兰、意大利
{"title":"Special Issue for the 22nd Workshop \"From Objects to Agents\" (WOA 2021)","authors":"Roberta Calegari, Giovanni Ciatto, Andrea Omicini, Giuseppe Vizzari","doi":"10.3233/ia-220141","DOIUrl":"https://doi.org/10.3233/ia-220141","url":null,"abstract":"Roberta Calegaria,∗, Giovanni Ciattob, Andrea Omicinib and Giuseppe Vizzaric aAlma Mater Research Institute for Human-Centered Artificial Intelligence (AlmaAI), ALMA MATER STUDIORUM—Università di Bologna, Italy bDipartimento di Informatica — Scienza e Ingegneria (DISI), ALMA MATER STUDIORUM—Università di Bologna, Italy cDipartimento di Informatica, Sistemistica e Comunicazione, Università degli Studi di Milano–Bicocca, Milano, Italy","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"16 1","pages":"3-5"},"PeriodicalIF":1.5,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42797227","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}
Federico Sabbatini, Giovanni Ciatto, Roberta Calegari, Andrea Omicini
A common practice in modern explainable AI is to post-hoc explain black-box machine learning (ML) predictors – such as neural networks – by extracting symbolic knowledge out of them, in the form of either rule lists or decision trees. By acting as a surrogate model, the extracted knowledge aims at revealing the inner working of the black box, thus enabling its inspection, representation, and explanation. Various knowledge-extraction algorithms have been presented in the literature so far. Unfortunately, running implementations of most of them are currently either proofs of concept or unavailable. In any case, a unified, coherent software framework supporting them all – as well as their interchange, comparison, and exploitation in arbitrary ML workflows – is currently missing. Accordingly, in this paper we discuss the design of PSyKE, a platform providing general-purpose support to symbolic knowledge extraction from different sorts of black-box predictors via many extraction algorithms. Notably, PSyKE targets symbolic knowledge in logic form, allowing the extraction of first-order logic clauses. The extracted knowledge is thus both machine- and human-interpretable, and can be used as a starting point for further symbolic processing—e.g. automated reasoning.
{"title":"Symbolic knowledge extraction from opaque ML predictors in PSyKE: Platform design & experiments","authors":"Federico Sabbatini, Giovanni Ciatto, Roberta Calegari, Andrea Omicini","doi":"10.3233/ia-210120","DOIUrl":"https://doi.org/10.3233/ia-210120","url":null,"abstract":"A common practice in modern explainable AI is to post-hoc explain black-box machine learning (ML) predictors – such as neural networks – by extracting symbolic knowledge out of them, in the form of either rule lists or decision trees. By acting as a surrogate model, the extracted knowledge aims at revealing the inner working of the black box, thus enabling its inspection, representation, and explanation. Various knowledge-extraction algorithms have been presented in the literature so far. Unfortunately, running implementations of most of them are currently either proofs of concept or unavailable. In any case, a unified, coherent software framework supporting them all – as well as their interchange, comparison, and exploitation in arbitrary ML workflows – is currently missing. Accordingly, in this paper we discuss the design of PSyKE, a platform providing general-purpose support to symbolic knowledge extraction from different sorts of black-box predictors via many extraction algorithms. Notably, PSyKE targets symbolic knowledge in logic form, allowing the extraction of first-order logic clauses. The extracted knowledge is thus both machine- and human-interpretable, and can be used as a starting point for further symbolic processing—e.g. automated reasoning.","PeriodicalId":42055,"journal":{"name":"Intelligenza Artificiale","volume":"16 1","pages":"27-48"},"PeriodicalIF":1.5,"publicationDate":"2022-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45027749","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：