We will review some of the major results in random graphs and some of the more challenging open problems. We will cover algorithmic and structural questions. We will touch on newer models, including those related to the WWW.

Random graphs

Every year, the Conference on Computational Natural Language Learning (CoNLL) features a shared task, in which participants train and test their learning systems on the same data sets. In 2018, one of two tasks was devoted to learning dependency parsers for a large number of languages, in a real-world setting without any gold-standard annotation on test input. All test sets followed a unified annotation scheme, namely that of Universal Dependencies. This shared task constitutes a 2nd edition—the first one took place in 2017 (Zeman et al., 2017); the main metric from 2017 has been kept, allowing for easy comparison, also in 2018, and two new main metrics have been used. New datasets added to the Universal Dependencies collection between mid-2017 and the spring of 2018 have contributed to increased difficulty of the task this year. In this overview paper, we define the task and the updated evaluation methodology, describe data preparation, report and analyze the main results, and provide a brief categorization of the different approaches of the participating systems.

CoNLL 2018 Shared Task : Multilingual Parsing from Raw Text to Universal Dependencies

The Conference on Computational Natural Language Learning (CoNLL) features a shared task, in which participants train and test their learning systems on the same data sets. In 2017, the task was devoted to learning dependency parsers for a large number of languages, in a real-world setting without any gold-standard annotation on input. All test sets followed a unified annotation scheme, namely that of Universal Dependencies. In this paper, we define the task and evaluation methodology, describe how the data sets were prepared, report and analyze the main results, and provide a brief categorization of the different approaches of the participating systems.

/pdf/conll-2017-shared-task-multilingual-parsing-from-raw-text-to-45vgerk8my.pdf

CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies

This paper reviews research into social tagging and folksonomy (as reflected in about 180 sources published through December 2007). Methods of researching the contribution of social tagging and folksonomy are described, and outstanding research questions are presented. This is a new area of research, where theoretical perspectives and relevant research methods are only now being defined. This paper provides a framework for the study of folksonomy, tagging and social tagging systems. Three broad approaches are identified, focusing first, on the folksonomy itself (and the role of tags in indexing and retrieval); secondly, on tagging (and the behaviour of users); and thirdly, on the nature of social tagging systems (as socio-technical frameworks).

/pdf/studying-social-tagging-and-folksonomy-a-review-and-9l6ryly1t1.pdf

Studying Social Tagging and Folksonomy: A Review and Framework

Collaborative tagging systems allow users to assign keywords - so called “tags” - to resources. Tags are used for navigation, finding resources and serendipitous browsing and thus provide an immediate benefit for users. These systems usually include tag recommendation mechanisms easing the process of finding good tags for a resource, but also consolidating the tag vocabulary across users. In practice, however, only very basic recommendation strategies are applied.

In this paper we evaluate and compare several recommendation algorithms on large-scale real life datasets: an adaptation of user-based collaborative filtering, a graph-based recommender built on top of the FolkRank algorithm, and simple methods based on counting tag occurrences. We show that both FolkRank and collaborative filtering provide better results than non-personalized baseline methods. Moreover, since methods based on counting tag occurrences are computationally cheap, and thus usually preferable for real time scenarios, we discuss simple approaches for improving the performance of such methods. We show, how a simple recommender based on counting tags from users and resources can perform almost as good as the best recommender.

/pdf/tag-recommendations-in-social-bookmarking-systems-119gl0ynzi.pdf

Tag recommendations in social bookmarking systems

The Semantic Web initiative has led to an upsurge of the interest in rules as a general and powerful way of processing, combining, and analyzing semantic information. Since several of the technologies underlying rule-based systems are already quite mature, it is important to understand how such systems might perform on the Web scale. OpenRuleBench is a suite of benchmarks for analyzing the performance and scalability of different rule engines. Currently the study spans five different technologies and eleven systems, but OpenRuleBench is an open community resource, and contributions from the community are welcome. In this paper, we describe the tested systems and technologies, the methodology used in testing, and analyze the results.

/pdf/openrulebench-an-analysis-of-the-performance-of-rule-engines-1ncatnz7l6.pdf

OpenRuleBench: an analysis of the performance of rule engines

This short paper describes a novel technique for generating personalized tag recommendations for users of social book- marking sites such as del.icio.us. Existing techniques recommend tags on the basis of their popularity among the group of all users; on the basis of recent use; or on the basis of simple heuristics to extract keywords from the url being tagged. Our method is designed to complement these approaches, and is based on recommending tags from urls that are similar to the one in question, according to two distinct similarity metrics, whose principal utility covers complementary cases.

/pdf/personalized-tag-recommendations-via-tagging-and-content-1te6ljbdi3.pdf

Personalized Tag Recommendations via Tagging and Content-based Similarity Metrics.

We define logic programs with defaults and argumentation theories, a new framework that unifies most of the earlier proposals for defeasible reasoning in logic programming. We present a model-theoretic semantics and study its reducibility and well-behavior properties. We use the framework as an elegant and flexible foundation to extend and improve upon Generalized Courteous Logic Programs (GCLP) [19]--one of the popular forms of defeasible reasoning. The extensions include higher-order and object-oriented features of Hilog and F-Logic [7,21]. The improvements include much simpler, incremental reasoning algorithms and more intuitive behavior. The framework and its Courteous family instantiation were implemented as an extension to the FLORA-2 system.

/pdf/logic-programming-with-defaults-and-argumentation-theories-1lpmktui3f.pdf

Logic Programming with Defaults and Argumentation Theories

We introduce doc2dial, a new dataset of goal-oriented dialogues that are grounded in the associated documents. Inspired by how the authors compose documents for guiding end users, we first construct dialogue flows based on the content elements that corresponds to higher-level relations across text sections as well as lower-level relations between discourse units within a section. Then we present these dialogue flows to crowd contributors to create conversational utterances. The dataset includes over 4500 annotated conversations with an average of 14 turns that are grounded in over 450 documents from four domains. Compared to the prior document-grounded dialogue datasets, this dataset covers a variety of dialogue scenes in information-seeking conversations. For evaluating the versatility of the dataset, we introduce multiple dialogue modeling tasks and present baseline approaches.

doc2dial: A Goal-Oriented Document-Grounded Dialogue Dataset

Our work involves enriching the Stack-LSTM transition-based AMR parser (Ballesteros and Al-Onaizan, 2017) by augmenting training with Policy Learning and rewarding the Smatch score of sampled graphs. In addition, we also combined several AMR-to-text alignments with an attention mechanism and we supplemented the parser with pre-processed concept identification, named entities and contextualized embeddings. We achieve a highly competitive performance that is comparable to the best published results. We show an in-depth study ablating each of the new components of the parser.

/pdf/rewarding-smatch-transition-based-amr-parsing-with-bpq2nny9vv.pdf

Hui Wan

Papers

OpenRuleBench: an analysis of the performance of rule engines

Personalized Tag Recommendations via Tagging and Content-based Similarity Metrics.

Logic Programming with Defaults and Argumentation Theories

doc2dial: A Goal-Oriented Document-Grounded Dialogue Dataset

Rewarding Smatch: Transition-Based AMR Parsing with Reinforcement Learning