Similarity functions are a fundamental component of many learning algorithms. When dealing with string or tree-structured data, measures based on the edit distance are widely used, and there exist a few methods for learning them from data. In this context, we recently proposed GESL (Bellet et al., 2012 [3]), an approach to string edit similarity learning based on loss minimization which offers theoretical guarantees as to the generalization ability and discriminative power of the learned similarities. In this paper, we argue that GESL, which has been originally dedicated to deal with strings, can be extended to trees and lead to powerful and competitive similarities. We illustrate this claim on a music recognition task, namely melody classification, where each piece is represented as a tree modeling its structure as well as rhythm and pitch information. The results show that GESL outperforms standard as well as probabilistically-learned edit distances and that it is able to describe consistently the underlying melodic similarity model.

Research networks provide a framework for review, synthesis and systematic testing of theories by multiple scientists across international borders critical for addressing global-scale issues. In 2012, a GHG research network referred to as MAGGnet (Managing Agricultural Greenhouse Gases Network) was established within the Croplands Research Group of the Global Research Alliance on Agricultural Greenhouse Gases (GRA). With involvement from 46 alliance member countries, MAGGnet seeks to provide a platform for the inventory and analysis of agricultural GHG mitigation research throughout the world. To date, metadata from 315 experimental studies in 20 countries have been compiled using a standardized spreadsheet. Most studies were completed (74%) and conducted within a 1–3-year duration (68%). Soil carbon and nitrous oxide emissions were measured in over 80% of the studies. Among plant variables, grain yield was assessed across studies most frequently (56%), followed by stover (35%) and root (9%) biomass. MAGGnet has contributed to modeling efforts and has spurred other research groups in the GRA to collect experimental site metadata using an adapted spreadsheet. With continued growth and investment, MAGGnet will leverage limited-resource investments by any one country to produce an inclusive, globally shared meta-database focused on the science of GHG mitigation.