Author name cluster

Xin Sui

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

RLDM Conference 2019 Conference Abstract

The Dynamics of Frustration

Bowen J Fung
Xin Sui
Colin F. Camerer
Dean Mobbs

Frustration is a widely experienced emotional state that has been linked to a wide range of societal and individual issues. Early research characterized a “frustration effect” whereby behavior is invigorated im- mediately subsequent to the non-delivery of an expected reward. Here we present an experimental approach that aimed to measure and model the effect of frustrative non-reward on motor vigor within a reinforcement learning framework. Subjects were instructed to earn rewards by squeezing a dynamometer handgrip at a specific force, while we surreptitiously recorded non-instrumental motor responses in between trials. We found that the non-instrumental motor responses were significantly predicted by a simple, parameter-free associative learning model that represented primary frustration. This trial-by-trial analysis allowed us to precisely quantify the conditions under which this classic frustration effect arises, thereby situating this sub- jective state within a mathematical framework. Unlike earlier work that employed one-shot extinction trials, our data point to a parametric effect of frustration on generalized motor output. This adds to the growing body of literature that relates reinforcement learning mechanisms to domains outside of choice, and provides a quantitative link between reward, emotion, and behavior. The dependence of frustration on reward history and its apparent Pavlovian effect on motor output also strongly suggest that frustration serves an adaptive role in behavior.

PDF Details

IJCAI Conference 2013 Conference Paper

Analysis and Optimization of Multi-Dimensional Percentile Mechanisms

Xin Sui
Craig Boutilier
Tuomas Sandholm

We consider the mechanism design problem for agents with single-peaked preferences over multi-dimensional domains when multiple alternatives can be chosen. Facility location and committee selection are classic embodiments of this problem. We propose a class of percentile mechanisms, a form of generalized median mechanisms, that are strategy-proof, and derive worst-case approximation ratios for social cost and maximum load for L1 and L2 cost models. More importantly, we propose a samplebased framework for optimizing the choice of percentiles relative to any prior distribution over preferences, while maintaining strategy-proofness. Our empirical investigations, using social cost and maximum load as objectives, demonstrate the viability of this approach and the value of such optimized mechanisms vis-à-vis mechanisms derived through worst-case analysis.

PDF Details DOI

IJCAI Conference 2013 Conference Paper

Multi-Dimensional Single-Peaked Consistency and Its Approximations

Xin Sui
Alex Francois-Nienaber
Craig Boutilier

Single-peakedness is one of the most commonly used domain restrictions in social choice. However, the extent to which agent preferences are single-peaked in practice, and the extent to which recent proposals for approximate single-peakedness can further help explain voter preferences, is unclear. In this article, we assess the ability of both single-dimensional and multi-dimensional approximations to explain preference proﬁles drawn from several real-world elections. We develop a simple branch-andbound algorithm that ﬁnds multi-dimensional, singlepeaked axes that best ﬁt a given proﬁle, and which works with several forms of approximation. Empirical results on two election data sets show that preferences in these elections are far from single-peaked in any onedimensional space, but are nearly single-peaked in two dimensions. Our algorithms are reasonably efﬁcient in practice, and also show excellent anytime performance.

PDF Details DOI

AAAI Conference 2011 Conference Paper

Efficiency and Privacy Tradeoffs in Mechanism Design

Xin Sui
Craig Boutilier

A key problem in mechanism design is the construction of protocols that reach socially efﬁcient decisions with minimal information revelation. This can reduce agent communication, and further, potentially increase privacy in the sense that agents reveal no more private information than is needed to determine an optimal outcome. This is not always possible: previous work has explored the tradeoff between communication cost and efﬁciency, and more recently, communication and privacy. We explore a third dimension: the tradeoff between privacy and efﬁciency. By sacriﬁcing efﬁciency, we can improve the privacy of a variety of existing mechanisms. We analyze these tradeoffs in both second-price auctions and facility location problems (introducing new incremental mechanisms for facility location along the way). Our results show that sacriﬁces in efﬁciency can provide gains in privacy (and communication), in both the average and worst case.

PDF Details

AAMAS Conference 2008 Conference Paper

Identifying Beneficial Teammates using Multi-Dimensional Trust

Jaesuk Ahn
Xin Sui
David DeAngelis
Suzanne Barber

Multi-agent teams must be capable of selecting the most beneficial teammates for different situations. Multi-dimensional trustworthiness assessments have been shown significantly beneficial to agents when selecting appropriate teammates to achieve a given goal. Reliability, quality, availability, timeliness and compatibility define the behavioral constraints of the multidimensional trust (MDT) model. Given the MDT model in this research, an agent learns to identify the most beneficial teammates by prioritizing each dimension differently. An agent’s attitudes towards rewards, risks and urgency are used to drive an agent’s prioritization of dimensions in a MDT model. Each agent is equipped with a Temporal-Difference (TD) learning mechanism with tile coding to identify its optimal set of attitudes and change its attitudes when the environment changes. Experimental results show that changing attitudes to give preferences for respective dimensions in the MDT offers a superior means to finding the best teammates for goal achievement.

PDF