-
From Data Fusion to Knowledge Fusion
Authors:
Xin Luna Dong,
Evgeniy Gabrilovich,
Geremy Heitz,
Wilko Horn,
Kevin Murphy,
Shaohua Sun,
Wei Zhang
Abstract:
The task of {\em data fusion} is to identify the true values of data items (eg, the true date of birth for {\em Tom Cruise}) among multiple observed values drawn from different sources (eg, Web sites) of varying (and unknown) reliability. A recent survey\cite{LDL+12} has provided a detailed comparison of various fusion methods on Deep Web data. In this paper, we study the applicability and limitat…
▽ More
The task of {\em data fusion} is to identify the true values of data items (eg, the true date of birth for {\em Tom Cruise}) among multiple observed values drawn from different sources (eg, Web sites) of varying (and unknown) reliability. A recent survey\cite{LDL+12} has provided a detailed comparison of various fusion methods on Deep Web data. In this paper, we study the applicability and limitations of different fusion techniques on a more challenging problem: {\em knowledge fusion}. Knowledge fusion identifies true subject-predicate-object triples extracted by multiple information extractors from multiple information sources. These extractors perform the tasks of entity linkage and schema alignment, thus introducing an additional source of noise that is quite different from that traditionally considered in the data fusion literature, which only focuses on factual errors in the original sources. We adapt state-of-the-art data fusion techniques and apply them to a knowledge base with 1.6B unique knowledge triples extracted by 12 extractors from over 1B Web pages, which is three orders of magnitude larger than the data sets used in previous data fusion papers. We show great promise of the data fusion approaches in solving the knowledge fusion problem, and suggest interesting research directions through a detailed error analysis of the methods.
△ Less
Submitted 1 March, 2015;
originally announced March 2015.
-
Dynamic Routing for Flying Ad Hoc Networks
Authors:
S. Rosati,
K. Kruzelecki,
G. Heitz,
D. Floreano,
B. Rimoldi
Abstract:
This paper reports experimental results on self-organizing wireless networks carried by small flying robots. Flying ad hoc networks (FANETs) composed of small unmanned aerial vehicles (UAVs) are flexible, inexpensive and fast to deploy. This makes them a very attractive technology for many civilian and military applications. Due to the high mobility of the nodes, maintaining a communication link b…
▽ More
This paper reports experimental results on self-organizing wireless networks carried by small flying robots. Flying ad hoc networks (FANETs) composed of small unmanned aerial vehicles (UAVs) are flexible, inexpensive and fast to deploy. This makes them a very attractive technology for many civilian and military applications. Due to the high mobility of the nodes, maintaining a communication link between the UAVs is a challenging task. The topology of these networks is more dynamic than that of typical mobile ad hoc networks (MANETs) and of typical vehicle ad hoc networks (VANETs). As a consequence, the existing routing protocols designed for MANETs partly fail in tracking network topology changes. In this work, we compare two different routing algorithms for ad hoc networks: optimized link-state routing (OLSR), and predictive-OLSR (P-OLSR). The latter is an OLSR extension that we designed for FANETs; it takes advantage of the GPS information available on board. To the best of our knowledge, P-OLSR is currently the only FANET-specific routing technique that has an available Linux implementation. We present results obtained by both Media Access Control (MAC) layer emulations and real-world experiments. In the experiments, we used a testbed composed of two autonomous fixed-wing UAVs and a node on the ground. Our experiments evaluate the link performance and the communication range, as well as the routing performance. Our emulation and experimental results show that P-OLSR significantly outperforms OLSR in routing in the presence of frequent network topology changes.
△ Less
Submitted 18 March, 2015; v1 submitted 17 June, 2014;
originally announced June 2014.
-
Convex Point Estimation using Undirected Bayesian Transfer Hierarchies
Authors:
Gal Elidan,
Ben Packer,
Geremy Heitz,
Daphne Koller
Abstract:
When related learning tasks are naturally arranged in a hierarchy, an appealing approach for coping with scarcity of instances is that of transfer learning using a hierarchical Bayes framework. As fully Bayesian computations can be difficult and computationally demanding, it is often desirable to use posterior point estimates that facilitate (relatively) efficient prediction. However, the hierarch…
▽ More
When related learning tasks are naturally arranged in a hierarchy, an appealing approach for coping with scarcity of instances is that of transfer learning using a hierarchical Bayes framework. As fully Bayesian computations can be difficult and computationally demanding, it is often desirable to use posterior point estimates that facilitate (relatively) efficient prediction. However, the hierarchical Bayes framework does not always lend itself naturally to this maximum aposteriori goal. In this work we propose an undirected reformulation of hierarchical Bayes that relies on priors in the form of similarity measures. We introduce the notion of "degree of transfer" weights on components of these similarity measures, and show how they can be automatically learned within a joint probabilistic framework. Importantly, our reformulation results in a convex objective for many learning problems, thus facilitating optimal posterior point estimation using standard optimization techniques. In addition, we no longer require proper priors, allowing for flexible and straightforward specification of joint distributions over transfer hierarchies. We show that our framework is effective for learning models that are part of transfer hierarchies for two real-life tasks: object shape modeling using Gaussian density estimation and document classification.
△ Less
Submitted 13 June, 2012;
originally announced June 2012.
