A Multi-context BDI Recommender System: From Theory to Simulation
Summary (2 min read)
Introduction
- Nowadays, recommender systems must cope with an increasing demand of complexity, for example, an application for recommending routes in a traffic scenario should deal with different contextual information (e.g., information about the user location) and other non-logical components linked to human behavior like desires, beliefs or emotions.
- In particular, one of the most popular agent architectures, the Belief-DesireIntention (BDI) model, seems to be particularly suitable for this task.
- The authors propose a multi-agent simulation to evaluate the overall behavior of the multi-context BDI recommender system they presented in [6], where different strategies are applied.
- For this purpose, the authors give an overview of the different theories used to define contexts, and explain how all those contexts are relied together to define the whole behavior of the system.
- The simulation and the results are discussed in Section III.
II. THE MULTI-AGENT FRAMEWORK
- The specification of their agent model is based on MultiContext Systems (MCS) [7], to allow for a separation of the definitions of the different formal components or units.
- The authors first extended the classical BDI model with others contexts.
- Li which contain axioms from classical propositional logic, and from necessity and possibility measures of possibility theory.
- For specifying and reasoning over plans, i.e., the Planning and Intention contexts, the authors propose to adopt the 5W (Who, What, Why, When, Where) vocabulary1, which is relevant for describing different concepts and constraints related to plans and allows spatial and temporal reasoning over plans and intentions.
- To show the applicability of their multi-agent BDI framework, an experimental model is presented and evaluated in the next section using the NetLogo Platform.
III. THE SYSTEM SIMULATION
- In agent-based systems with spatial reasoning and social behavior, a visual output is needed to display the agents’ moving and interaction in two or three dimensional spaces.
- An agent represents a user with different desires and beliefs that are randomly initialized.
- Plans are a list of activities that consist of moving from one destination to another.
- Each destination contains some rewards that an agent will obtain when it reaches that destination.
- These agents will decide to accept or not this recommendation according to the trust degree in the sender, and whether there is any information in their knowledge base (desire or belief base) that contradicts this one.
A. Experimental Setup
- Table I summarises the parameters that can be varied for different use cases.
- They also have different profiles (desires, trust degrees, positions, etc. . . ).
- Links are also created randomly between agents according to an initial link number defined by the user at the beginning of the simulation .
- The authors used Netlogo 5.3.1 version to implement their simulation.
- The authors developed the rest of the behavior of the agents using the NetLogo language with some extensions.
B. Experimental Results and Discussion
- The model and experimental data were analysed using the RNetLogo extension [12].
- According to these desires and the aforementioned behavior, an agent calculates the recommendation which has a plan as output.
- Only a new belief from an external source that does not contradict the agent initial belief can make it change its intention.
- It shows that utility augments considerably within social agents compared to the utility within individual agents.
- The authors can conclude that agents get more satisfaction collectively from exchanging information than alone.
IV. CONCLUSIONS
- The purpose of the simulation was to evaluate agent behaviors adopting two different strategies (the social and the individual strategy) in order to infer the quality of recommendations.
- Results show that agents achieve a better performance collectively when they are in “communities”, i.e., agents with shared interests (thus similar to each other), then when they are acting as solitary agents.
- The authors believe that the issues of trust and recommendation are tightly related.
- The authors ignore how the framework will behave when errors are introduced in the communicated information.
- The authors believe that approaches like [13] can be used to extend their agent model with further reasoning abilities and, consequently, to deal with information reliability as well.
Did you find this useful? Give us your feedback
Citations
4 citations
Cites background or methods from "A Multi-context BDI Recommender Sys..."
...In this evaluation, different agent’s strategies are considered, following the ideas proposed by (Othmane et al., 2016a):...
[...]
...For more details about this agent model, we refer the reader to (Othmane et al., 2016b)....
[...]
...To answer this question, we introduce CARS, a spatio-temporal Cognitive Agent-based Recom- mender System, extending with spatio-temporal information the system proposed by (Othmane et al., 2016b)....
[...]
...Agent-based recommender systems (Casali et al., 2008a; Chen and Cheng, 2010; Batet et al., 2012; Othmane et al., 2016b) have been proposed in the last years in different scenarios, like tourism, health-care, and traffic, to provide suggestions and support users to achieve their goals....
[...]
...In this evaluation, different agent’s strategies are considered, following the ideas proposed by (Othmane et al., 2016a): • individual agent strategy: agents behave individually without taking into account any information coming from other agents....
[...]
3 citations
Cites methods from "A Multi-context BDI Recommender Sys..."
...This work is an extension of the recommender system based on the BDI model we presented in [12, 11], where a spatio-temporal representation of beliefs and desires is proposed....
[...]
2 citations
Cites methods from "A Multi-context BDI Recommender Sys..."
...This line of work has been continued in the PhD thesis of Amel Ben Othmane [34, 33, 32], I supervised with Andrea Tettamanzi and Nhan Le Than....
[...]
References
85 citations
"A Multi-context BDI Recommender Sys..." refers background in this paper
...If an agent decides to accept the recommendation, then it adds this information to its desire base, and then trigger the recalculation of its intentions according to the updated desire base....
[...]
69 citations
Additional excerpts
...Finally, some conclusions are drawn....
[...]
67 citations
"A Multi-context BDI Recommender Sys..." refers methods in this paper
...The aim of the simulation is to compute a recommendation based on a user initial set of beliefs and desires, and to see how our agent will adapt the recommendation in two cases: ● the agent is part of a social network (social agent), i.e., it has links with other agents similar to it, ● the agent…...
[...]
39 citations
28 citations
"A Multi-context BDI Recommender Sys..." refers methods in this paper
...Among them, we refer in particular to the approach of Casali et al. [5] who propose a BDI recommender agent in the tourism domain....
[...]