Conference
Graph-based Methods for Natural Language Processing
About: Graph-based Methods for Natural Language Processing is an academic conference. The conference publishes majorly in the area(s): Graph (abstract data type) & Computer science. Over the lifetime, 105 publications have been published by the conference receiving 1531 citations.
Topics: Graph (abstract data type), Computer science, Task (project management), Semantic similarity, Inference
Papers
1 Jun 2018
TL;DR: This article proposed to learn word and sentence embeddings in hyperbolic space in an unsupervised manner from text corpora, which can encode certain intuitive notions of hierarchy, such as word-context frequency and phrase constituency.
Abstract: Natural language text exhibits hierarchical structure in a variety of respects. Ideally, we could incorporate our prior knowledge of this hierarchical structure into unsupervised learning algorithms that work on text data. Recent work by Nickel and Kiela (2017) proposed using hyperbolic instead of Euclidean embedding spaces to represent hierarchical data and demonstrated encouraging results when embedding graphs. In this work, we extend their method with a re-parameterization technique that allows us to learn hyperbolic embeddings of arbitrarily parameterized objects. We apply this framework to learn word and sentence embeddings in hyperbolic space in an unsupervised manner from text corpora. The resulting embeddings seem to encode certain intuitive notions of hierarchy, such as word-context frequency and phrase constituency. However, the implicit continuous hierarchy in the learned hyperbolic space makes interrogating the model’s learned hierarchies more difficult than for models that learn explicit edges between items. The learned hyperbolic embeddings show improvements over Euclidean embeddings in some – but not all – downstream tasks, suggesting that hierarchical organization is more useful for some tasks than others.
173 citations
7 Aug 2009
TL;DR: The algorithm aggregates local relatedness information via a random walk over a graph constructed from an underlying lexical resource that forms a "semantic signature" that can be compared to another such distribution to get a relat-edness score for texts.
Abstract: Many tasks in NLP stand to benefit from robust measures of semantic similarity for units above the level of individual words. Rich semantic resources such as WordNet provide local semantic information at the lexical level. However, effectively combining this information to compute scores for phrases or sentences is an open problem. Our algorithm aggregates local relatedness information via a random walk over a graph constructed from an underlying lexical resource. The stationary distribution of the graph walk forms a "semantic signature" that can be compared to another such distribution to get a relat-edness score for texts. On a paraphrase recognition task, the algorithm achieves an 18.5% relative reduction in error rate over a vector-space baseline. We also show that the graph walk similarity between texts has complementary value as a feature for recognizing textual entailment, improving on a competitive baseline system.
124 citations
24 Sep 2019
TL;DR: This work presents a novel method that traces and visualizes features that contribute to a classification decision in the visible and hidden layers of a GCN, and exposes hidden cross-layer dynamics in the input graph structure.
Abstract: Representations in the hidden layers of Deep Neural Networks (DNN) are often hard to interpret since it is difficult to project them into an interpretable domain. Graph Convolutional Networks (GCN) allow this projection, but existing explainability methods do not exploit this fact, i.e. do not focus their explanations on intermediate states. In this work, we present a novel method that traces and visualizes features that contribute to a classification decision in the visible and hidden layers of a GCN. Our method exposes hidden cross-layer dynamics in the input graph structure. We experimentally demonstrate that it yields meaningful layerwise explanations for a GCN sentence classifier.
89 citations
1 Nov 2019
TL;DR: In this paper, a graph neural network-based model was proposed to detect fake news and misleading information through online media outlets, which does away with the need of feature engineering for fine grained fake news classification.
Abstract: The rising growth of fake news and misleading information through online media outlets demands an automatic method for detecting such news articles. Of the few limited works which differentiate between trusted vs other types of news article (satire, propaganda, hoax), none of them model sentence interactions within a document. We observe an interesting pattern in the way sentences interact with each other across different kind of news articles. To capture this kind of information for long news articles, we propose a graph neural network-based model which does away with the need of feature engineering for fine grained fake news classification. Through experiments, we show that our proposed method beats strong neural baselines and achieves state-of-the-art accuracy on existing datasets. Moreover, we establish the generalizability of our model by evaluating its performance in out-of-domain scenarios. Code is available at https://github.com/MysteryVaibhav/fake_news_semantics.
84 citations
7 Aug 2009
TL;DR: This paper approaches Label Propagation as solution to a system of linear equations which can be implemented as a scalable parallel algorithm using the map-reduce framework and provides empirical evidence to that effect using two natural language tasks -- lexical relat-edness and polarity induction.
Abstract: Label Propagation, a standard algorithm for semi-supervised classification, suffers from scalability issues involving memory and computation when used with large-scale graphs from real-world datasets. In this paper we approach Label Propagation as solution to a system of linear equations which can be implemented as a scalable parallel algorithm using the map-reduce framework. In addition to semi-supervised classification, this approach to Label Propagation allows us to adapt the algorithm to make it usable for ranking on graphs and derive the theoretical connection between Label Propagation and PageRank. We provide empirical evidence to that effect using two natural language tasks -- lexical relat-edness and polarity induction. The version of the Label Propagation algorithm presented here scales linearly in the size of the data with a constant main memory requirement, in contrast to the quadratic cost of both in traditional approaches.
56 citations
Performance Metrics
| Year | Papers |
|---|---|
| 2021 | 20 |
| 2020 | 14 |
| 2019 | 23 |
| 2018 | 6 |
| 2013 | 12 |
| 2012 | 9 |