Taking decisions in a Hybrid Conversational AI architecture using Influence Diagrams

Abstract

This paper explores the application of the Influence Diagrams model for decision-making in the context of conversational agents. The system consists of a Conversational Recommender System (CoRS), in which the decision-making module is separate from the language generation module. It provides the capability to evolve a belief based on user responses, which in turn influences the decisions made by the conversational agent. The proposed system is based on a pre-existing CoRS that relies on Bayesian Networks informing a separate decision process. The introduction of Influence Diagrams aims to integrate both Bayesian inference and the dialogue move selection phase into a single model, thereby generalising the decision-making process. To test the efectiveness and plausibility of the dialogues generated by the developed CoRS, a dialogue simulator was created and the simulated interactions were evaluated by a pool of human judges.

-

CEUR ceur-ws.org

1. Introduction

erated significant enthusiasm among professionals in the ifeld of artificial intelligence as well as the general public.

In recent years, the success of neural networks has gen- the produced output might have meant. probable continuation of the provided prompt, they leave entirely to the human reader the task of interpreting what

From a linguistics point of view, within the framework

of Austin’s speech act theory [ 3 ] “saying something” Various applications, such as speech recognition, com- equals “doing something”; the act of producing a senputer vision and even interactive conversational mod- tence (locutive act ) is fuelled by an intention (illocutive can have various implications, even within the scien- conversation is a form of intervention in the world: it els like ChatGPT, have increasingly engaged users, inevitably shaping their perception of AI. This perception tific community. Attributing human-level intelligence to the tasks currently accomplished by neural networks is questionable, as these tasks barely rise to the level of abilities possessed by many animals [ 1 ]. Neural-based approaches to artificial intelligence have been criticised act) that produces changes in the world (perlocutive act).

This classic view of the act of speaking highlights that

is put in action to alter in some way the conversational context. This same position is also found in the recent literature about the role of causality in artificial intelligence.

Judea Pearl’s Ladder of Causation [ 4 ] puts intervention capabilities on the second level of the ladder, characterised sociative methods. One notable analysis of the problems because of the limitations that are intrinsic to purely as- by the verb doing, as in Austin’s seminal work. In this work, machine learning capabilities are limited to the that come when considering linguistic material gener- first step of the Ladder, concerned with observational ated without a real understanding of the meaning of what is being said is found in [ 2 ], which highlights that, because of the way it is generated, content produced by

GPT models adheres to at least one formal definition of

bullshit. The fundamental problem with these models is that, while they are trained to capture surface aspects of communication, they are never exposed to the reasons why language is produced. When they output the most Agents (FANTASIA) [ 5 ], an Unreal Engine1 plugin designed to develop embodied conversational agents. Built upon the functionalities ofered by the tool, the FAN

TASIA Interaction Model follows these main principles:

Behaviour Trees (BT) [6] are used to organise and pri

oritise dialogue moves; Graph Databases (i.e., Neo4j grated into the database. A MOVIELENSUSER node is [ 7 ]) are used for knowledge representation and dialogue created for each user in the dataset, and a RATED relastate tracking; Probabilistic Graphical Models (PGM) tionship is established between the MOVIELENSUSER are used for decision making; LLMs are used to verbalise node and a MOVIE node for each reported rating in the the decisions taken by PGMs. dataset. In addition to a number of basic properties such

The latest results obtained using FANTASIA, presented as name, year of birth and ratings, MOVIE and PERSON in [ 8, 9 ], used a decision system based on Bayesian Net- nodes are characterised by authority attributes and hub works to estimate probability distributions over ratings scores calculated by means of the HITS algorithm [ 14 ]. for users of a movie recommender system. The decision As discussed in [ 9 ], these network analysis measures help about the dialogue move was taken by a rule-based sys- model cognitive characteristics that are relevant for the tem taking into account these estimates. In this work, we selection of plausible arguments [ 15 ]. Finally, the graph further develop the approach by generalising the deci- database is used to store a dialogue state graph which sion process using a single model, an Influence Diagram tracks the agent’s relationships with the knowledge do(ID) [ 10 ]. IDs represent an extension of BNs [ 11 ] since, main and other agents, including humans. This graph in addition to probabilistic nodes, they also contain: can be modified by the agent through speech acts in order to evolve it towards graph patterns that the agent • Decision nodes, which represent decision points identifies as goal patterns, i.e. a desired configuration for the agent and which may be multiple within of the dialogue state. In this way, the graph database the model. will be interrogated by the CoRS by extracting a relevant • Utility nodes, which represent utility (or cost) fac- sub-graph taking into account the knowledge base and tors and which will drive the agent’s decisions, belief of the system evolved during the conversation. since the objective will be to maximise the utility In the reference system, the decision-making level inof the model. volves a BN dynamically generated on the basis of the Consequently, in addition to the modelling of proba- extracted relevant sub-graph. In particular, in the case of bilistic inference problems, the use of IDs also enables Movie Recommendation, the actors, films and genres are the modelling and solving of decision-making problems, nodes of the BN, while the (oriented) relations between in accordance with the criterion of maximum expected them represent the causal relations. Initially, each node utility. In this way, the ID encapsulates both the Bayesian is initialised by specifying its own CPT, which can either inference and the decision phase in a single, more flexible be pre-calculated or derived from parent nodes. This and elegant model. network is used to adjust the exploitation/exploration cycle, typical of recommendation dialogue [ 16 ], by taking into account the data extracted from MovieLens (soft ev2. Original system idence) and the feedback gathered through the dialogue with the user (hard evidence). This way, the BN can repThe original system on which the proposed system is resent the probability of each movie and each feature to based was presented in [ 8, 9 ]. This system is a CoRS with be of interest for a user, after applying Bayesian inference. argumentative capabilities based on linguistic and cogni- Based on the information extracted from the Bayesian tive principles. From a design point of view, the original network, a module outside the PGM is responsible for the system followed the FANTASIA Interaction model and decisions taken. Specifically, the system decides whether the PGM of choice were Bayesian Networks (BN), imple- to recommend a candidate item (exploitation move) or, in mented using the aGRuM library [ 12 ]. the case of non-recommendation, to ask the most useful

From the knowledge representation point of view, a question (exploration move), based on the criteria considgraph database is adopted to host information derived ered in [ 8 ]. In the case of exploitation moves, in addifrom Linked Open Data (LOD) sources. For the purposes tion to item recommendation, argumentation is provided of this case study, the movies domain will be consid- based on the three most useful features, whose utility ered. The knowledge base is constructed by collecting is calculated as the harmonic mean of four (normalised) data from diferent sources and enriched using graph parameters related to cognitive properties [ 15 ]. data science techniques, which are employed to capture latent information. The procedure is described in [ 13 ]. The main entities of the knowledge base are repre- 3. Proposed system sented by the labels MOVIE, PERSON and GENRE, which are interconnected by appropriate relationships (such as HAS_GENRE, WORKED_IN, and so on). Additionally, information from the MovieLens 25M2 dataset is inte

The proposed system based on IDs replicates part of the

reference strategy: the aim of this work is to provide a first test of the capabilities of the IDs to handle the problem so we concentrate on the fundamental steps of

2https://grouplens.org/datasets/movielens/

the original strategy. Table 1 shows the characteristics of the reference system, highlighting the ones reproduced by the proposed system. The approach is inspired by the system presented in [ 17 ].

3.1. ID for Movie Recommendation

3.2. Simulation

The current system was tested by simulating a dialogue

between the system and a MovieLens user whose answers are derived from ratings recorded in the dataset.

At the beginning of the conversation, the agent has no information about the user and for this reason the user a movie, then a question must be asked. In particular, immediately specifies the preferred genre. This informathe most useful question must be chosen. In this case tion is derived by searching the database for that genre study, as anticipated, the exploration only involves the for which the average rating of that particular user is entropy of the features, not taking into account other the highest. All following questions are polar questions aspects of the features and other nodes. In particular, and concern PERSON type features. Again, the answer for each feature f extracted from the database, an uncer- is derived by considering the ratings given by the user tainty node H(f) is contextually created. Each node H to the ARGITEMs associated with that feature. Once represents the entropy of the related feature. A decision the genre is known, a positive belief likes is created that node What question is in charge of deciding which ques- associates the user with the preferred genre and at this tion should be asked, and depends on both nodes H and point the database is queried by extracting the best three, the decision node Recommendation, generating a decision the related features and the secondary films. If, from the sequence starting from the latter. The idea is that the ID, the best action is to recommend, the system proposes choice of question must depend both on the entropy of one of the candidate films to the user; otherwise, if the the features that can be chosen and on the decision that best action is not to recommend, the system asks the was made at the time of the recommendation, i.e. the most useful question. If the user’s answer consists of decision to perform an exploitation or an exploration a positive or negative preference, this involves adding move. Among the possible choices of What question, in evidence in the system, adding the user’s stance on that fact, there is also a No question, which only makes sense feature to the database and reconstructing the ID from to choose in the case of an exploitation move. Finally, a a dataframe extracted with the same query used at the Cost utility function represents the utility of the What beginning of the dialogue. The idea is that by keeping question choice. Fig. 2 shows the structure of the explo- track of the user’s stances collected as the system asks ration branch in a generic form. Tab. 2 shows the cost questions, it is possible to extract target movies that are associated with each decision sequence that the system more consistent with the user’s preferences. When a film is capable of undertaking. In particular, the highest cost, is recommended, the system also provides arguments to equal to −100, is applied to those decision sequences that support its choice, consisting of a selection of the most are to be avoided. Conversely, the lowest cost, equal important features related to the recommended film, thus to 0, is applied to the case where the system does not implementing Argumentation-based dialogue [ 18 ]. The ask questions. A variable cost, between −100 and −1, dialogue provided by the simulation is constructed by is calculated in the case where the system decides not using templates causing the generated conversation to to recommend and ask a question about an actor. The sound unnatural. For this reason, these template-based magnitude of this cost will depend on the entropy value dialogues were reformulated by ChatGPT-4 to make the of the relevant uncertainty node. The higher the entropy, conversation more natural, using the following prompt: the lower the cost of the corresponding question. The Rephrase the following dialogue to make it sound more idea is to collect evidence on the uncertainty nodes on natural. Keep the structure and only change the sentences. which the model’s uncertainty is most concentrated, as the system’s objective is to lower the model’s entropy level before making a recommendation.

4. Experimental setup

The experimental phase followed the approach used in [ 8 ]. The approach involves recruiting 20 participants via the Prolific 3 portal who were asked to complete a survey on the Qualtrics4 platform that involves the evaluation of 20 dialogues divided into three types: • Five dialogues taken from INSPIRED Corpus [ 19 ], a dataset of human-human interactions for Movie Recommendation. These dialogues represent the positive subset of the control group. • Five system-generated dialogues where both the extraction of candidate films and the choice of supporting features are random, independent of system belief. These dialogues represent the negative subset of the control group. • Ten dialogues generated by the system using the presented strategy, which represent the target dialogues. (a) Results obtained by the original system (b) Results obtained by the proposed system

3https://www.prolific.com/ 4https://www.qualtrics.com/

Fig. 3 shows the survey task with the four questions asked to the participant for each dialogue, for which the dialogues are higher than those obtained by the negative participant gives a score between 1 and 5. Q1 refers to the dialogues and lower than those obtained by the positive consistency of the questions asked during the exploration dialogues. In particular, the diference between target move, in order to understand whether the features are and negative dialogues is more pronounced on Q4, which selected correctly during the dialogue. Q2 and Q3 refer to is an indicator that the supporting arguments make the the naturalness of the dialogue, with the latter referring recommendation plausible. to the user’s perception of the recommender’s level of As an objective measure, during the generation of the expertise. Finally, Q4 refers to the quality of the features dialogues for each round, the average normalised enchosen to support the recommendation. In conclusion, tropy of the ID was recorded, calculated as the average of the participants were native English speakers living in the normalised entropy among all variable nodes of the the UK or US and they were compensated according to model. In Fig. 5 it can be observed that a) during a tarthe average hourly wage of their home country. get dialogue the average entropy of the model decreases, in contrast to the case where b) the dialogue is random and the average entropy of the model does not tend to 5. Results decrease. The first scenario is compatible with the idea that the system accumulates information as the dialogue Fig. 4 shows the scores obtained by the current system progresses, in accordance with the strategy adopted. In based on ID for each question blue(b), compared with the second scenario, on the other hand, the ID is regenthe scores obtained by the original system based on BN erated at each turn from randomly extracted candidate (a). In both instances, the scores obtained by the target films, making it unlikely that the new extracted features contribute in accumulating coherent information.

To further analyse the data concerning the synthetic dithe probabilistic model. The results achieved in this case were lower than the ones of the original system, but this was expected as only part of the original strategy was replicated. Future work will cover the implementation of the missing functionalities and the deployment of the system in the Unreal Engine, as the technology to implement IDs has been integrated in the FANTASIA plugin.

We will also investigate the possibility of integrating the argument selection process in the ID to fully support Argumentation Based Dialogue.

Acknowledgments

This work is supported by the Supporting Patients with

Embodied Conversational Interfaces and Argumentative alogues, we use a Cumulative Link Mixed Model (CLMM) Language (SPECIAL) project, funded by the University [ 20 ] with Laplace approximation, [ 21 ]. This model ac- of Naples on the ”Fondi per la Ricerca di Ateneo” (FRA) commodates random efects attributable to individual program (CUP: E65F22000050001) participants or specific stimuli, treating them as blocking variables and assesses the likelihood of observing high values on the Likert score in relation to the independent References variable (i.e., dialogue type). The test revealed that the association between the occurrence of high scores, in general, is very strong ( < 0.001 ) for both target and positive dialogues and, as expected, absent for negative values. This result is stronger with respect to the results obtained in [ 8, 9 ], where only a weak association was observed. There are multiple aspects that contribute to this result, in our opinion. First of all, in the original work, the -value was already very close to the strong significance threshold ( = 0.0144 ), so the efect was only technically considered weak even in that case. Also, there is a chance that the simplified situation may have harmed negative dialogues more than the other two categories.

As a final remark, however, the IDs have indeed made the decision process more uniform and flexible, given the introduction of utility functions and a unified framework for decision making. The quality improvement of the decision process management, especially in deciding when to recommend, given the available arguments to support the position has improved the system even in its basic form.

6. Conclusions & future work

The results obtained indicate that the implementation of a knowledge graph exploration strategy based on the ID is more efective than a random strategy. This conclusion is further supported by objective measures, including the system’s entropy, which decreases as the system accumulates information during the dialogue before making a recommendation. It is therefore possible to generalise within an ID a decision-making process that, in the original system, was implemented by a module external to

References

[1] A. Darwiche , Human-level intelligence or animallike abilities? , Communications of the ACM 61 ( 2018 ) 56 - 67 .

[2] M. T. Hicks , J. Humphries , J. Slater , Chatgpt is bullshit, Ethics and Information Technology 26 ( 2024 ) 38 .

[3] J. L. Austin , How to Do Things with Words ., Clarendon Press, 1962 .

[4] J. Pearl , D. Mackenzie , The book of Why, Basic Books, 2018 .

[5] A. Origlia , F. Cutugno , A. Rodà , P. Cosi , C. Zmarich , Fantasia: a framework for advanced natural tools and applications in social, interactive approaches ., Multimedia Tools and Applications 78 ( 2019 ) 13613 - 13648 .

[6] G. Flórez-Puga , M. A. Gomez-Martin , P. P. GomezMartin , B. Diaz-Agudo , P. A. Gonzalez-Calero , Query-enabled behavior trees , IEEE Transactions on Computational Intelligence and AI in Games 1 ( 2009 ) 298 - 308 .

[7] J. Webber , A programmatic introduction to neo4j, in: Proceedings of the 3rd annual conference on Systems, programming, and applications: software for humanity , ACM , 2012 , pp. 217 - 218 .

[8] M. Di Bratto , A. Origlia , M. Di Maro , S. Mennella , Linguistics-based dialogue simulations to evaluate argumentative conversational recommender systems, User Modeling and User-Adapted Interaction ( 2024 ) 1 - 31 .

[9] M. Di Bratto , A. Origlia , M. Di Maro , S. Mennella , On the use of plausible arguments in explainable conversational ai , in: Proceedings of Interspeech , 2024 .

[10] R. A. Howard , J. E. Matheson , Influence diagrams, Decision Analysis 2 ( 2005 ) 127 - 143 .

[11] A. Biedermann , F. Taroni , Bayesian Networks and Influence Diagrams , 2023 , pp. 271 - 280 .

[12] G. Ducamp , C. Gonzales , P. Wuillemin, agrum/pyagrum : a toolbox to build models and algorithms for probabilistic graphical models in python , in: Proceedings of the 10th International Conference on Probabilistic Graphical Models , volume 138 , PMLR , 2020 , pp. 609 - 612 .

[13] A. Origlia , M. Di Bratto , M. Di Maro , S. Mennella , A multi-source graph representation of the movie domain for recommendation dialogues analysis , in: Proceedings of the Thirteenth Language Resources and Evaluation Conference , 2022 , pp. 1297 - 1306 .

[14] J. M. Kleinberg , Authoritative sources in a hyperlinked environment , Journal of the ACM (JACM) 46 ( 1999 ) 604 - 632 .

[15] F. Paglieri , C. Castelfranchi , Revising beliefs through arguments: Bridging the gap between argumentation and belief revision in mas ( 2005 ) 78 - 94 .

[16] C. Gao , W. Lei , X. He , M. de Rijke , T.-S. Chua, Advances and challenges in conversational recommender systems: A survey , arXiv preprint arXiv:2101.09459 ( 2021 ).

[17] P. Lison , C. Kennington , Opendial: A toolkit for developing spoken dialogue systems with probabilistic rules , in: Proceedings of ACL-2016 system demonstrations , 2016 , pp. 67 - 72 .

[18] H. Prakken , Historical overview of formal argumentation , volume 1 , College Publications , 2018 .

[19] S. A. Hayati , D. Kang , Q. Zhu , W. Shi , Z. Yu , Inspired: Toward sociable recommendation dialog systems , in: 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020 , Association for Computational Linguistics (ACL ), 2020 , pp. 8142 - 8152 .

[20] A. Agresti , Categorical data analysis , volume 792 , John Wiley & Sons, 2012 .

[21] Z. Shun , P. McCullagh , Laplace approximation of high dimensional integrals , Journal of the Royal Statistical Society: Series B (Methodological) 57 ( 1995 ) 749 - 760 .