Search

Reward machines allow the definition of rewards for temporally extended tasks and behaviors. Specifying "informative" reward machines can be challenging. One way to address this is to generate reward machines from a high-level abstract description of the learning environment, using techniques such as AI planning. However, previous planning-based approaches generate a reward machine based on a single (sequential or partial-order) plan, and do not allow maximum flexibility to the learning agent. In this paper we propose a new approach to synthesising reward machines which is based on the set of partial order plans for a goal. We prove that learning using such "maximally permissive" reward machines results in higher rewards than learning using RMs based on a single plan. We present experimental results which support our theoretical claims by showing that our approach obtains higher rewards than the single-plan approach in practice., Comment: Paper accepted for publication at the European Conference on Artificial Intelligence (ECAI) 2024

Automated testing of computer games is a challenging problem, especially when lengthy scenarios have to be tested. Automating such a scenario boils down to finding the right sequence of interactions given an abstract description of the scenario. Recent works have shown that an agent-based approach works well for the purpose, e.g. due to agents' reactivity, hence enabling a test agent to immediately react to game events and changing state. Many games nowadays are multi-player. This opens up an interesting possibility to deploy multiple cooperative test agents to test such a game, for example to speed up the execution of multiple testing tasks. This paper offers a cooperative multi-agent testing approach and a study of its performance based on a case study on a 3D game called Lab Recruits.

Recent developments enable the quantification of causal control given a structural causal model (SCM). This has been accomplished by introducing quantities which encode changes in the entropy of one variable when intervening on another. These measures, named causal entropy and causal information gain, aim to address limitations in existing information theoretical approaches for machine learning tasks where causality plays a crucial role. They have not yet been properly mathematically studied. Our research contributes to the formal understanding of the notions of causal entropy and causal information gain by establishing and analyzing fundamental properties of these concepts, including bounds and chain rules. Furthermore, we elucidate the relationship between causal entropy and stochastic interventions. We also propose definitions for causal conditional entropy and causal conditional information gain. Overall, this exploration paves the way for enhancing causal machine learning tasks through the study of recently-proposed information theoretic quantities grounded in considerations about causality., Comment: In Proceedings of the conference CLeaR (Causal Learning and Reasoning) 2024

Improving sample efficiency is central to Reinforcement Learning (RL), especially in environments where the rewards are sparse. Some recent approaches have proposed to specify reward functions as manually designed or learned reward structures whose integrations in the RL algorithms are claimed to significantly improve the learning efficiency. Manually designed reward structures can suffer from inaccuracy and existing automatically learning methods are often computationally intractable for complex tasks. The integration of inaccurate or partial reward structures in RL algorithms fail to learn optimal policies. In this work, we propose an RL algorithm that can automatically structure the reward function for sample efficiency, given a set of labels that signify subtasks. Given such minimal knowledge about the task, we train a high-level policy that selects optimal sub-tasks in each state together with a low-level policy that efficiently learns to complete each sub-task. We evaluate our algorithm in a variety of sparse-reward environments. The experiment results show that our approach significantly outperforms the state-of-art baselines as the difficulty of the task increases.

Artificial intelligence models and methods commonly lack causal interpretability. Despite the advancements in interpretable machine learning (IML) methods, they frequently assign importance to features which lack causal influence on the outcome variable. Selecting causally relevant features among those identified as relevant by these methods, or even before model training, would offer a solution. Feature selection methods utilizing information theoretical quantities have been successful in identifying statistically relevant features. However, the information theoretical quantities they are based on do not incorporate causality, rendering them unsuitable for such scenarios. To address this challenge, this article proposes information theoretical quantities that incorporate the causal structure of the system, which can be used to evaluate causal importance of features for some given outcome variable. Specifically, we introduce causal versions of entropy and mutual information, termed causal entropy and causal information gain, which are designed to assess how much control a feature provides over the outcome variable. These newly defined quantities capture changes in the entropy of a variable resulting from interventions on other variables. Fundamental results connecting these quantities to the existence of causal effects are derived. The use of causal information gain in feature selection is demonstrated, highlighting its superiority over standard mutual information in revealing which features provide control over a chosen outcome variable. Our investigation paves the way for the development of methods with improved interpretability in domains involving causation., Comment: In Proceedings of the European Conference on Artificial Intelligence (ECAI) 2023. Work presented at the third XI-ML workshop of ECAI 2023

Observers can glean information from others' emotional expressions through the act of drawing inferences from another individual's emotional expressions. It is important for socially aware artificial systems to be capable of doing that as it can facilitate social interaction among agents, and is particularly important in human-robot interaction for supporting a more personalized treatment of users. In this short paper, we propose a methodology for developing a formal model that allows agents to infer another agent's values from her emotion expressions.

The increasing applications of AI systems require personalized explanations for their behaviors to various stakeholders since the stakeholders may have various knowledge and backgrounds. In general, a conversation between explainers and explainees not only allows explainers to obtain the explainees' background, but also allows explainees to better understand the explanations. In this paper, we propose an approach for an explainer to communicate personalized explanations to an explainee through having consecutive conversations with the explainee. We prove that the conversation terminates due to the explainee's justification of the initial claim as long as there exists an explanation for the initial claim that the explainee understands and the explainer is aware of.

Training a dialogue policy using deep reinforcement learning requires a lot of exploration of the environment. The amount of wasted invalid exploration makes their learning inefficient. In this paper, we find and define an important reason for the invalid exploration: dead-ends. When a conversation enters a dead-end state, regardless of the actions taken afterward, it will continue in a dead-end trajectory until the agent reaches a termination state or maximum turn. We propose a dead-end resurrection (DDR) algorithm that detects the initial dead-end state in a timely and efficient manner and provides a rescue action to guide and correct the exploration direction. To prevent dialogue policies from repeatedly making the same mistake, DDR also performs dialogue data augmentation by adding relevant experiences containing dead-end states. We first validate the dead-end detection reliability and then demonstrate the effectiveness and generality of the method by reporting experimental results on several dialogue datasets from different domains.

The complexity of computer games is ever increasing. In this setup, guiding an automated test algorithm to find a solution to solve a testing task in a game's huge interaction space is very challenging. Having a model of a system to automatically generate test cases would have a strong impact on the effectiveness and efficiency of the algorithm. However, manually constructing a model turns out to be expensive and time-consuming. In this study, we propose an online agent-based search approach to solve common testing tasks when testing computer games that also constructs a model of the system on-the-fly based on the given task, which is then exploited to solve the task. To demonstrate the efficiency of our approach, a case study is conducted using a game called Lab Recruits.

Communication is an effective mechanism for coordinating the behavior of multiple agents. In the field of multi-agent reinforcement learning, agents can improve the overall learning performance and achieve their objectives by communication. Moreover, agents can communicate various types of messages, either to all agents or to specific agent groups, and through specific channels. With the growing body of research work in MARL with communication (Comm-MARL), there is lack of a systematic and structural approach to distinguish and classify existing Comm-MARL systems. In this paper, we survey recent works in the Comm-MARL field and consider various aspects of communication that can play a role in the design and development of multi-agent reinforcement learning systems. With these aspects in mind, we propose several dimensions along which Comm-MARL systems can be analyzed, developed, and compared., Comment: 10 pages, 4 figures, 10 tables

Norms have been widely proposed as a way of coordinating and controlling the activities of agents in a multi-agent system (MAS). A norm specifies the behaviour an agent should follow in order to achieve the objective of the MAS. However, designing norms to achieve a particular system objective can be difficult, particularly when there is no direct link between the language in which the system objective is stated and the language in which the norms can be expressed. In this paper, we consider the problem of synthesising a norm from traces of agent behaviour, where each trace is labelled with whether the behaviour satisfies the system objective. We show that the norm synthesis problem is NP-complete.

Player experience (PX) evaluation has become a field of interest in the game industry. Several manual PX techniques have been introduced to assist developers to understand and evaluate the experience of players in computer games. However, automated testing of player experience still needs to be addressed. An automated player experience testing framework would allow designers to evaluate the PX requirements in the early development stages without the necessity of participating human players. In this paper, we propose an automated player experience testing approach by suggesting a formal model of event-based emotions. In particular, we discuss an event-based transition system to formalize relevant emotions using Ortony, Clore, & Collins (OCC) theory of emotions. A working prototype of the model is integrated on top of Aplib, a tactical agent programming library, to create intelligent PX test agents, capable of appraising emotions in a 3D game case study. The results are graphically shown e.g. as heat maps. Emotion visualization of the test agent would ultimately help game designers in creating content that evokes a certain experience in players., Comment: This is a preprint of an article with the same title, accepted in 9th International Workshop on Engineering Multi-Agent Systems (EMAS 2021) which was held as a part of 20th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2021)

The goal of entity matching in knowledge graphs is to identify entities that refer to the same real-world objects using some similarity metric. The result of entity matching can be seen as a set of entity pairs interpreted as the same-as relation. However, the identified set of pairs may fail to satisfy some structural properties, in particular transitivity, that are expected from the same-as relation. In this work, we show that an ad-hoc enforcement of transitivity, i.e. taking the transitive closure, on the identified set of entity pairs may decrease precision dramatically. We therefore propose a methodology that starts with a given similarity measure, generates a set of entity pairs that are identified as referring to the same real-world objects, and applies the cluster editing algorithm to enforce transitivity without adding many spurious links, leading to overall improved performance.

Reinforcement learning (RL) agents in human-computer interactions applications require repeated user interactions before they can perform well. To address this "cold start" problem, we propose a novel approach of using cognitive models to pre-train RL agents before they are applied to real users. After briefly reviewing relevant cognitive models, we present our general methodological approach, followed by two case studies from our previous and ongoing projects. We hope this position paper stimulates conversations between RL, HCI, and cognitive science researchers in order to explore the full potential of the approach.

Public sentiment (the opinions, attitudes or feelings expressed by the public) is a factor of interest for government, as it directly influences the implementation of policies. Given the unprecedented nature of the COVID-19 crisis, having an up-to-date representation of public sentiment on governmental measures and announcements is crucial. While the 'staying-at-home' policy makes face-to-face interactions and interviews challenging, analysing real-time Twitter data that reflects public opinion toward policy measures is a cost-effective way to access public sentiment. In this context, we collect streaming data using the Twitter API starting from the COVID-19 outbreak in the Netherlands in February 2020, and track Dutch general public reactions on governmental measures and announcements. We provide temporal analysis of tweet frequency and public sentiment over the past seven months. We also identify public attitudes towards two Dutch policies in case studies: one regarding social distancing and one regarding wearing face masks. By presenting those preliminary results, we aim to provide visibility into the social media discussions around COVID-19 to the general public, scientists and policy makers. The data collection and analysis will be updated and expanded over time., Comment: 25 pages, 8 figures

In this paper we address the interplay among intention, time, and belief in dynamic environments. The first contribution is a logic for reasoning about intention, time and belief, in which assumptions of intentions are represented by preconditions of intended actions. Intentions and beliefs are coherent as long as these assumptions are not violated, i.e. as long as intended actions can be performed such that their preconditions hold as well. The second contribution is the formalization of what-if scenarios: what happens with intentions and beliefs if a new (possibly conflicting) intention is adopted, or a new fact is learned? An agent is committed to its intended actions as long as its belief-intention database is coherent. We conceptualize intention as commitment toward time and we develop AGM-based postulates for the iterated revision of belief-intention databases, and we prove a Katsuno-Mendelzon-style representation theorem., Comment: 83 pages, 4 figures, Artificial Intelligence journal pre-print

Multiagent Systems (MAS) research reached a maturity to be confidently applied to real-life complex problems. Successful application of MAS methods for behavior modeling, strategic reasoning, and decentralized governance, encouraged us to focus on applicability of MAS techniques in a class of industrial systems and to elaborate on potentials and challenges for method integration/contextualization. We direct attention towards a form of industrial practices called Industrial Symbiosis Systems (ISS) as a highly dynamic domain of application for MAS techniques. In ISS, firms aim to reduce their material and energy footprint by circulating reusable resources among the members. To enable systematic reasoning about ISS behavior and support firms' (as well as ISS designers') decisions, we see the opportunity for marrying industrial engineering with engineering multiagent systems. This enables introducing (1) representation frameworks to reason about dynamics of ISS, (2) operational semantics to develop computational models for ISS, and (3) coordination mechanisms to enforce desirable ISS behaviors. We argue for applicability and expressiveness of resource-bounded formalisms and norm-aware mechanisms for the design and deployment of ISS practices. In this proposal, we elaborate on different dimensions of ISS, present a methodological foundation for ISS development, and finally discuss open problems., Comment: 10 pages; In Pre-proceedings of the 7th International Workshop on Engineering Multi-Agent Systems (EMAS 2019), 13th-14th May 2019, Montreal, Canada; Editors: Rafael H. Bordini, Louise A. Dennis, Yves Lesp\'erance

Norms with sanctions have been widely employed as a mechanism for controlling and coordinating the behavior of agents without limiting their autonomy. The norms enforced in a multi-agent system can be revised in order to increase the likelihood that desirable system properties are fulfilled or that system performance is sufficiently high. In this paper, we provide a preliminary analysis of some types of norm revision: relaxation and strengthening. Furthermore, with the help of some illustrative scenarios, we show the usefulness of norm revision for better satisfying the overall system objectives., Comment: Appears in the preproceedings of the 29th Benelux Conference on Artificial Intelligence (BNAIC 2017), Groningen, The Netherlands, pp. 281 to 290

This paper builds on an existing notion of group responsibility and proposes two ways to define the degree of group responsibility: structural and functional degrees of responsibility. These notions measure the potential responsibilities of (agent) groups for avoiding a state of affairs. According to these notions, a degree of responsibility for a state of affairs can be assigned to a group of agents if, and to the extent that, the group has the potential to preclude the state of affairs., Comment: Presented in the 27th Belgian-Netherlands Conference on Artificial Intelligence (BNAIC 2015), Hasselt, Belgium

In this short note we address the issue of expressing norms (such as obligations and prohibitions) in temporal logic. In particular, we address the argument from [Governatori 2015] that norms cannot be expressed in Linear Time Temporal Logic (LTL)., Comment: 3 pages