We develop and test computational methods for guiding collaboration that demonstrate how shared plans can be created in real-world settings, where agents can be expected to have diverse and varying goals, preferences, and availabilities. The methods are motivated and evaluated in the realm of ridesharing, using GPS logs of commuting data. We consider challenges with coordination among self-interested people aimed at minimizing the cost of transportation and the impact of travel on the environment. We present planning, optimization, and payment mechanisms that provide fair and efficient solutions to the rideshare collaboration challenge. We evaluate different VCG-based payment schemes in terms of their computational efficiency, budget balance, incentive compatibility, and strategy proofness. We present the behavior and analyses provided by the ABC ridesharing prototype system. The system learns about destinations and preferences from GPS traces and calendars, and considers time, fuel, environmental, and cognitive costs. We review how ABC generates rideshare plans from hundreds of real-life GPS traces collected from a community of commuters and reflect about the promise of employing the ABC methods to reduce the number of vehicles on the road, thus reducing CO2 emissions and fuel expenditures.

To harness modern multi-core processors, it is imperative to develop parallel versions of fundamental algorithms. In this paper, we present a general approach to best-first heuristic search in a shared-memory setting. Each thread attempts to expand the most promising open nodes. By using abstraction to partition the state space, we detect duplicate states without requiring frequent locking. We allow speculative expansions when necessary to keep threads busy. We identify and fix potential livelock conditions in our approach, verifying its correctness using temporal logic. In an empirical comparison on STRIPS planning, grid pathfinding, and sliding tile puzzle problems using an 8-core machine, we show that A* implemented in our framework yields faster search than improved versions of previous parallel search proposals. Our approach extends easily to other best-first searches, such as Anytime weighted A*.

In many applications, the data, such as web pages and research papers, contain relation (link) structure among entities in addition to textual content information. Matrix factorization (MF) methods, such as latent semantic indexing (LSI), have been successfully used to map either content information or relation information into a lower-dimensional latent space for subsequent processing. However, how to simultaneously model both the relation information and the content information effectively with an MF framework is still an open research problem. In this paper, we propose a novel MF method called relation regularized matrix factorization (RRMF) for relational data analysis. By using relation information to regularize the content MF procedure, RRMF seamlessly integrates both the relation information and the content information into a principled framework. We propose a linear-time learning algorithm with convergence guarantee to learn the parameters of RRMF. Extensive experiments on real data sets show that RRMF can achieve state-of-the-art performance.

An important problem in computational social choice concerns whether it is possible to prevent manipulation of voting rules by making it computationally intractable. To answer this, a key question is how frequently voting rules are manipulable. We [Xia and Conitzer, 2008] recently defined the class of generalized scoring rules (GSRs) and characterized the frequency of manipulability for such rules. We showed, by examples, that most common rules seem to fall into this class. However, no natural axiomatic characterization of the class was given, leaving the possibility that there are natural rules to which these results do not apply. In this paper, we characterize the class of GSRs based on two natural properties: it is equal to the class of rules that are anonymous and finitely locally consistent. Generalized scoring rules also have other uses in computational social choice. For these uses, the order of the GSR (the dimension of its score vector) is important. Our characterization result implies that the order of a GSR is related to the minimum number of locally consistent components of the rule. We proceed to bound the minimum number of locally consistent components for some common rules.

Calendar management tools assist users with coordinating their daily life. Different tasks have to be scheduled according to the user preferences. In many cases, tasks are at different locations and travel times have to be considered. Therefore, these kinds of calendar management problems can be regarded as spatio-temporal optimisation problems and are often variants of traveling salesman problems (TSP) or vehicle routing problems. While standard TSPs require a solution to include all tasks, prize-collecting TSPs are more suited for calendar management problems as they require a solution that optimises the total sum of prizes we assigned to tasks at different locations. If we now add time windows that limit when tasks can occur, these prize-collecting TSPs with time windows (TW-TSP) are excellent abstractions of spatio-temporal optimisation problems such as calendar management. Due to the inherent complexity of TW-TSPs, the existing literature considers mainly approximation algorithms or special cases. We present a novel algorithm for TW-TSPs that enables us to find the optimal solution to TW-TSP problems occurring in real-world calendar management applications efficiently. Our algorithm is a fixed-parameter tractable algorithm that depends on the maximal number of tasks that can be revisited from some other task, a parameter which is small in the application scenario we consider.

CP-networks have been proposed as a simple and intuitive graphical tool for representing conditional ceteris paribus preference statements over the values of a set of variables. While the problem of reasoning with CP-networks has been receiving some attention, there are very few works that address the problem of learning CP-networks. In this work we investigate the task of learning CP-networks, given access to a set of pairwise comparisons. We first prove that the learning problem is intractable, even under several simplifying assumptions. We then present an algorithm that, under certain assumptions about the observed pairwise comparisons, identifies a CP-network that entails these comparisons. We finally show that the proposed algorithm is a PAC-learner, and, thus, that the CP-networks it induces accurately predict the user's preferences on previously unseen situations.

Exploration for robotic mapping is typically handled using greedy entropy reduction. Here we show how to apply information lookahead planning to a challenging instance of this problem in which an Autonomous Underwater Vehicle (AUV) maps hydrothermal vents. Given a simulation of vent behaviour we derive an observation function to turn the planning for mapping problem into a POMDP. We test a variety of information state MDP algorithms against greedy, systematic and reactive search strategies. We show that directly rewarding the AUV for visiting vents induces effective mapping strategies. We evaluate the algorithms in simulation and show that our information lookahead method outperforms the others.

Typical graph-theoretic approaches for semi-supervised classification infer labels of unlabeled instances with the help of graph Laplacians. Founded on the spectral decomposition of the graph Laplacian, this paper learns a kernel matrix via minimizing the leave-one-out classification error on the labeled instances.  To this end, an efficient algorithm is presented based on linear programming, resulting in a transductive spectral kernel. The idea of our algorithm stems from regularization methodology and also has a nice interpretation in terms of spectral clustering. A simple classifier can be readily built upon the learned kernel, which suffices to give prediction for any data point aside from those in the available dataset. Besides this usage, the spectral kernel can be effectively used in tandem with conventional kernel machines such as SVMs. We demonstrate the efficacy of the proposed algorithm through experiments carried out on challenging classification tasks.

Graph-based manifold-ranking methods have been successfully applied to topic-focused multi-document summarization. This paper further proposes to use the multi-modality manifold-ranking algorithm for extracting topic-focused summary from multiple documents by considering the within-document sentence relationships and the cross-document sentence relationships as two separate modalities (graphs). Three different fusion schemes, namely linear form, sequential form and score combination form, are exploited in the algorithm. Experimental results on the DUC benchmark datasets demonstrate the effectiveness of the proposed multi-modality learning algorithms with all the three fusion schemes.

In many real world planning domains, some observation information is optional and useless to the execution of a plan; on the other hand, information acquisition may require some kind of cost. The problem of observation reduction for strong plans has been addressed in the literature. However, observation reduction for plans with contexts (which are more general and useful than strong plans in robotics) is still a open problem. In this paper, we present an attempt to solve the problem. Our first contribution is the definition of structured plans, which can encode sequential, conditional and iterative behaviors, and is expressive enough for dealing with incomplete observation information and internal states of the agent. A second contribution is an observation reduction algorithm for plans with contexts, which can transform a plan with contexts into a structured plan that only branches on necessary observation information.