Negotiating agents that learn about others' preferences Abstract In multiagent systems, an agent does not usually have complete information about the preferences and decision making processes of other agents. This might prevent the agents from making coordinated choices, purely due to their ignorance of what others want. This paper describes the integration of a learning module into a communication-intensive negotiating agent architecture. The learning module gives the agents the ability to learn about other agents' preferences via past interactions. Over time, the agents can incrementally update their models of other agents' preferences and use them to make better coordinated decisions. Combining both communication and learning, as two complement knowledge acquisition methods, helps to reduce the amount of communication needed on average, and is justified in situation where communication is computationaUy costly or simply not desirable (e.g. to preserve the individual privacy).
We introduce the Multiagent Agreement Problem (MAP) to represent a class of multiagent scheduling problems. MAP is based on the Distributed Constraint Reasoning (DCR) paradigm and requires agents to choose values for variables to satisfy not only their own constraints, but also equality constraints with other agents. The goal is to represent problems in which agents must agree on scheduling decisions, for example, to agree on the start time of a meeting. We investigate a challenging class of MAP - private, incremental MAP (piMAP) in which agents do incremental scheduling of activities and there exist privacy restrictions on information exchange. We investigate a range of strategies for piMAP, called bumping strategies. We empirically evaluate these strategies in the domain of calendar management where a personal assistant agent must schedule meetings on behalf of its human user. Our results show that bumping decisions based on scheduling difficulty models of other agents can significantly improve performance over simpler bumping strategies.
Constraint satisfaction/optimization is a powerful paradigm for solving numerous tasks in distributed AI, including planning and scheduling. However, up to now, distributed algorithms for constraint reasoning (especially optimization) have not been applied to large-scale systems due to their prohibitive complexity in terms of number of messages being exchanged. We have developed a series of new techniques for distributed constraint optimization, based on dynamic programming. These approaches require a linear number of messages, whose maximal size depends on a parameter of the problem graph, called the induced width. Thus, these methods are likely to work very well on large but loose problems. We believe that these methods have features that make them particularly interesting for solving planning and scheduling problems in dynamic, distributed environments.
Calendar scheduling is a personal behavior and there are diverse factors on which the user's decision depends. Whether the user is initiating a new meeting or responding to a meeting request she chooses an action with multiple objectives. For instance, when trying to schedule a new meeting at a preferred time and location, the user may also want to minimize change to her existing meetings, and she takes a scheduling action that best compromises the overall objectives. Our goal is to build an agent that can predict the best scheduling action to take, where "best" is defined in terms of the user's true preference. We take a machine learning approach and focus on the problem of learning the user's preference, through observation of the user as she engages in meeting scheduling episodes. We propose a hybrid preference learning framework in which we first learn utility functions of simple individual preferences such as preferred time-of-day, and then qualitatively evaluate complex scheduling options by learning a classifier from pairwise preferences. We summarize proof of principal experiments that illustrate both types of learning.
Microsoft has announced that it's acquired Genee, a productivity app that launched in beta last year and focused on automating the task of scheduling meetings. The terms of the deal were not disclosed. Genee allowed users to set up meetings without having to consult a calendar. Once Genee was CCed on an email asking a contact for a meeting, the app would email them with options based on your availability and preferences and add the appointment to your schedule. It's certainly not the only such service to do this; others like x.ai and Clara offer similar capabilities.