A usual strategy to select the final answer in factoid Question-Answering (QA) relies on redundancy. A score is given to each candidate answer as a function of its frequency of occurrence, and the final answer is selected from the set of candidates sorted in decreasing order of score. For that purpose, systems often try to group together semantically equivalent answers. However, they hold several other semantic relations, such as inclusion, which are not considered, and candidates are mostly seen independently, as competitors. Our hypothesis is that not just equivalence, but other relations between candidate answers have impact on the performance of a redundancy-based QA system. In this paper, we describe experimental studies to back up this hypothesis. Our findings show that, with relatively simple techniques to recognize relations, systems' accuracy can be improved for answers of categories Number, Date and Entity.

Topic models have been used successfully for a variety of problems, often in the form of application-specific extensions of the basic Latent Dirichlet Allocation (LDA) model. Because deriving these new models in order to encode domain knowledge can be difficult and time-consuming, we propose the Fold·all model, which allows the user to specify general domain knowledge in First-Order Logic (FOL). However, combining topic modeling with FOL can result in inference problems beyond the capabilities of existing techniques. We have therefore developed a scalable inference technique using stochastic gradient descent which may also be useful to the Markov Logic Network (MLN) research community. Experiments demonstrate the expresive power of Fold·all, as well as the scalability of our proposed inference method.

Multi-agent path planning is a challenging problem with numerous real-life applications, including robotics, logistics, military operations planning, disaster rescue, and computer games. We look at navigating large numbers of mobile units to their targets on navigation graphs such as grid maps. The size of problems examined is significantly larger than can be handled using optimal multi-agent pathfinding algorithms in practice. We introduced MAPP, a tractable algorithm for multi-agent path planning on undirected graphs. MAPP and its extended versions are complete on well specified and tractably testable classes of problems. They have low-polynomial worst-case upper bounds for the running time, the memory requirements, and the length of solutions. Experiments on realistic game grid maps, with uniformly randomly generated start and target locations for each unit, show MAPP as a state-of-the-art multi-agent pathfinding algorithm in terms of scalability and success ratio (i.e., percentage of solved units). Even on challenging scenarios with 2000 units, MAPP solves 92% to 99.7% of units. FAR and WHCA*, two fast but incomplete algorithms that were previously state-of-the-art in terms of scalability, solve as few as 17.5% and 12.3% of these problems. The quality of MAPP's solutions is empirically analyzed using multiple quality criteria: total travel distance, makespan, and sum of actions (including move and wait actions). MAPP is competitive in terms of solution quality and speed with FAR and WHCA*. MAPP further provides the formal characterizations that FAR and WHCA* lack, on problems it can solve as well as low-polynomial upper bounds on the resources required. As optimal algorithms have limited scalability, we evaluated the solution quality of suboptimal algorithms using lower bounds of optimal values. We showed that MAPP's solutions have a reasonable quality. For example, MAPP's total travel distance is on average 19% longer than a lower bound on the optimal value.

My work focuses on building computational agents that assist people in managing their activities in environments in which tempo and complexity outstrip people’s cognitive capacity,such as in coordinating rescue teams in the aftermath of a disaster, or in helping people with dementia manage their everyday lives. A critical challenge faced in such environments is not only that individuals must factor complicated constraints into deciding how and when to act on their own goals, but also that their decisions are further constrained by choices made by others with whom they interact, such as between cooperating teams in disaster relief or between patients and caregivers in an assisted-living facility. An additional challenge in such situations is that the interests of individuals, such as privacy and autonomy, along with slow, costly, uncertain,or otherwise problematic communication may further limitindividuals’ abilities to work together. My work assumes that a computational agent is associated with each individual, and that these agents will work together efficiently to manage individual and joint activities, while maintaining autonomy and privacy to the extent possible.

Finally, explore possible methods for decision making in general, have received a lot uses of combinatorial aggregation in sequential voting, of attention in the AI community in recent years. The reasons and discuss theoretical generalisations to more complex logical for this focus are clear: SCT provides tools for the analysis of languages and practical applications. Particularly close to the interests of AI is the to study binary aggregation procedures, inspired by research problem of social choice in combinatorial domains (Chevaleyre in AI. As long as we do not know the intended application of et al., 2008), where the space of alternatives the individuals the model, there is no appropriate set of axioms to concentrate have to choose from has a combinatorial structure. Instead, we prove characterisation results concerning one Definition 1.

The paper describes a method of relation extraction, which is based on parsing the input text using a combination of a generic HPSG-based grammar and a highly focused domain- and relation-specific lexicon. We also show a method of unsupervised acquisition of such a lexicon from a large unlabeled corpus. Together, the methods introduce a novel approach to the “Open IE” task, which is superior in accuracy and in quality of relation identification to the existing approaches.

Another example is items to users based on their historical ratings. In that, although many people don't like animations, they may real-world scenarios, user interests may drift over still have interests in emerging 3-D animations because of the time since they are affected by moods, contexts, fantastic 3-D visual effects. These observations show that, and pop culture trends. This leads to the fact that although many aspects of user interests can be found based a user's historical ratings comprise many aspects of on users' historical ratings, at a certain time slice, one user's user interests spanning a long time period. However, interest may only focus on one or a couple of aspects. Thus, at a certain time slice, one user's interest may the static CF methods built on the entire historical ratings are only focus on one or a couple of aspects. Thus, inadequate to capture user-interest drift. In order to track user CF techniques based on the entire historical ratings interests and create comprehensive user profiles such that different may recommend inappropriate items. In this paper, recommendation strategies can be used for consistenttaste we consider modeling user-interest drift over time users and changing-taste users, a CF method that can based on the assumption that each user has multiple model user interests over time is required.

Recommender systems are becoming tools of choice to select the online information relevant to a given user. Collaborative filtering is the most popular approach to building recommender systems and has been successfully employed in many applications. With the advent of online social networks, the social network based approach to recommendation has emerged. This approach assumes a social network among users and makes recommendations for a user based on the ratings of the users who have direct or indirect social relations with the given user. As one of their major benefits, social network based approaches have been shown to reduce the problems with cold start users. In this paper, we explore a model-based approach for recommendation in social networks, employing matrix factorization techniques. Advancing previous work, we incorporate the mechanism of trust propagation into the model in a principled way. Trust propagation has been shown to be a crucial phenomenon in the social sciences, in social network analysis and in trust-based recommendation. We have conducted experiments on two real life data sets. Our experiments demonstrate that modeling trust propagation leads to a substantial increase in recommendation accuracy, in particular for cold start users.

We study the existence and computational complexity of coalitional stability concepts based on social networks. Our concepts represent a natural and rich combinatorial generalization of a recent notion termed partition equilibrium. We assume that players in a strategic game are embedded in a social (or, communication) network, and there are coordination constraints defining the set of coalitions that can jointly deviate in the game. A main feature of our approach is that players act in a "considerate" fashion to ignore potentially profitable (group) deviations if the change in their strategy may cause a decrease of utility to their neighbors in the network. We explore the properties of such considerate equilibria in application to the celebrated class of resource selection games (RSGs). Our main result proves existence of a super-strong considerate equilibrium in all symmetric RSGs with strictly increasing delays, for any social network among the players and feasible coalitions represented by the set of cliques. The existence proof is constructive and yields an efficient algorithm. In fact, the computed considerate equilibrium is a Nash equilibrium for a standard RSG, thus showing that there exists a state that is stable against selfish and considerate behavior simultaneously. Furthermore, we provide results on convergence of considerate dynamics.

Extracting the relations that exist between two entities is an important step in numerousWeb-related tasks such as information extraction.A supervised relation extraction system that is trained to extract a particular relation type might not accurately extract a new type of a relation for which it has not been trained.However, it is costly to create training data manually for every new relation type that one might want to extract.We propose a method to adapt an existing relation extraction system to extractnew relation types with minimum supervision. Our proposed method comprises two stages: learning a lower-dimensional projection between different relations, and learning a relational classifier for the target relation type with instance sampling. We evaluate the proposed method using a dataset that contains 2000 instances for 20 different relation types. Our experimental results show that the proposed method achieves a statistically significant macro-average F-score of 62.77. Moreover, the proposed method outperforms numerous baselines and a previously proposed weakly-supervised relation extraction method.