Here we present a network model in which self-sustaining recurrent activation emerges from simple cascades of activation. It is demonstrated that the ability to support such self-sustaining activation in our model depends on network connectivity as well as the ability to grow new links over time. Additionally, we explore how the probability of emergence of self-sustaining activity can be modulated by changing various network parameters and suggest potential applications of our findings.

This paper summarizes our ongoing project on developing an architecture for a robot that can acquire new words and their meanings while engaging in multi-domain dialogues. These two functions are crucial in making conversational service robots work in real tasks in the real world. Household robots and office robots need to be able to work in multiple task domains and they also need to engage in dialogues in multiple domains corresponding to those task domains. Lexical acquisition is necessary because speech understanding cannot be done without enough knowledge on words that are possibly spoken in the task domain. Our architecture is based on a multi-expert model in which multiple domain experts are employed and one of them is selected based on the user utterance and the situation to engage in the control of the dialogue and physical behaviors. We incorporate experts that have an ability to acquire new lexical entries and their meanings grounded on the physical world through spoken interactions. By appropriately selecting those experts, lexical acquisition in multi-domain dialogues becomes possible. An example robotic system based on this architecture that can acquire object names and location names demonstrates the viability of the architecture.

Robots are being considered for applications where they serve as proxies for humans interacting with another human,such as emergency response, hostage negotiation, and healthcare. In these domains, the human (“dependent”) is connected to multiple other humans (“controllers”) via the robot proxy for long periods of time. The dependent may want to interact with humans but also to engage the robot as a medium to the World Wide Web. In the future, medical personnel may use the robot for victim assistance and comfort while the rescue team plans and monitors extrication. Other applications include healthcare, where the robot is the link between a patient and a medical provider for intermittent,routine interactions, and hostage negotiation, where police may use a bomb squad robot to talk with and build rapport with the suspect while the SWAT team uses the robot’s sensors to build and maintain situation awareness.Under funding from the National Science Foundation, we are finishing the first year of investigating verbal and nonverbal communication strategies for robots who are serving as proxies for multiple humans interact with the humans who are dependent on them. Our work posits that such a robot would occupy a novel social medium position according to the Computers as Social Actors (CASA) model [Nass,Steuer, and Tauber1994] [Reeves and Nass1996]. Given that teleoperated robots are treated socially, it is unlikely that a rescue robot would be treated as a pure medium even if playing music or videos. Likewise, the limitations of autonomy and the interactions of specialists with the dependent prevent the robot from being a true social actor. Instead, social actor and pure medium are two extremes on the agent identity spectrum, with a social medium occupying a middle position.A social medium would be perceived as a loyal, helpful “go between” who is an advocate for the dependent, rather than a device for accomplishing the goals of multiple controllers(medical specialist, structural engineer, rescue operations official, etc.). To explore the social medium identity,we have built a physical prototype of a Survivor Buddy and are creating autonomous affective behaviors and a social medium toolkit to explore human-robot interaction.

In human robot dialogue, identifying intended referents from human partners’ spatial language is challenging. This is partly due to automated inference of potentially ambiguous underlying reference system (i.e., frame of reference ). To improve spatial language understanding, we conducted an empirical study to investigate the prevalence of ambiguities of frame of reference. Our findings indicate that ambiguities do arise frequently during human robot dialogues. Although situational factors from the spatial arrangement are less indicative for the underlying reference system, linguistic cues and individual preferences may allow reliable disambiguation.

The development of hand-crafted action and dialog generation models for a social robot is a time consuming process that yields a solution only for the relatively narrow range of interactions envisioned by the programmers. In this paper, we propose a data-driven solution for interactive behavior generation that leverages online games as a means of collecting large-scale data corpora for human-robot interaction research. We present a system in which action and dialog models for a collaborative human-robot task are learned based on a reproduction of the task in a two-player online game called Mars Escape.

As robots become incorporated into our environments, they must be equipped with the ability to communicate effectively with us. In particular, robots that perform longer tasks for a small set of people (e.g., a companion robot to escort visitors to meetings all day) need to be able to start and maintain interesting and relevant dialog with any and all humans involved.In this work, we present our ongoing work on our robot, CoBot, which is assigned an all-day task to escort a visitor around our building and perform tasks for her. We first describe CoBot's dialog manager which is responsible for the task-oriented dialog, including dialog to meet the visitor's needs, CoBot's notifications of interesting locations around the building, and the robot's own requests for help. We, then, focus two aspects of the dialog manager: 1) how CoBot can invoke more accurate answers to its requests for help from the visitor and 2) how to reduce repetitive dialog which can happen during all-day interactions. We provide an example dialog between CoBot and a visitor to illustrate the dialog manager's capabilities.

An important difficulty in developing spoken dialogue systems for robots is the open-ended nature of most interactions. Robotic agents must typically operate in complex, continuously changing environments which are difficult to model and do not provide any clear, predefined goal. Directly capturing this complexity in a single, large dialogue policy is thus inadequate. This paper presents a new approach which tackles the complexity of open-ended interactions by breaking it into a set of small, independent policies, which can be activated and deactivated at runtime by a dedicated mechanism. The approach is currently being implemented in a spoken dialogue system for autonomous robots.

Hutter (2007) recently introduced the loss rank principle (LoRP) as a generalpurpose principle for model selection. The LoRP enjoys many attractive properties and deserves further investigations. The LoRP has been well-studied for regression framework in Hutter and Tran (2010). In this paper, we study the LoRP for classification framework, and develop it further for model selection problems in unsupervised learning where the main interest is to describe the associations between input measurements, like cluster analysis or graphical modelling. Theoretical properties and simulation studies are presented.

Lasso and other regularization procedures are attractive methods for variable selection, subject to a proper choice of shrinkage parameter. Given a set of potential subsets produced by a regularization algorithm, a consistent model selection criterion is proposed to select the best one among this preselected set. The approach leads to a fast and efficient procedure for variable selection, especially in high-dimensional settings. Model selection consistency of the suggested criterion is proven when the number of covariates d is fixed. Simulation studies suggest that the criterion still enjoys model selection consistency when d is much larger than the sample size. The simulations also show that our approach for variable selection works surprisingly well in comparison with existing competitors. The method is also applied to a real data set.

The performance of the Lasso is well understood under the assumptions of the standard linear model with homoscedastic noise. However, in several applications, the standard model does not describe the important features of the data. This paper examines how the Lasso performs on a non-standard model that is motivated by medical imaging applications. In these applications, the variance of the noise scales linearly with the expectation of the observation. Like all heteroscedastic models, the noise terms in this Poisson-like model are \textit{not} independent of the design matrix. More specifically, this paper studies the sign consistency of the Lasso under a sparse Poisson-like model. In addition to studying sufficient conditions for the sign consistency of the Lasso estimate, this paper also gives necessary conditions for sign consistency. Both sets of conditions are comparable to results for the homoscedastic model, showing that when a measure of the signal to noise ratio is large, the Lasso performs well on both Poisson-like data and homoscedastic data. Simulations reveal that the Lasso performs equally well in terms of model selection performance on both Poisson-like data and homoscedastic data (with properly scaled noise variance), across a range of parameterizations. Taken as a whole, these results suggest that the Lasso is robust to the Poisson-like heteroscedastic noise.