New Regularized Algorithms for Transductive Learning

Talukdar, Partha; Crammer, Koby

doi:10.1007/978-3-642-04174-7_29

Cited by 130 publications

(145 citation statements)

References 9 publications

Supporting

Mentioning

145

Contrasting

Order By: Relevance

“…Our contributions are as follows: (1) we propose the use of Modified Adsorption (Talukdar and Crammer, 2009) as a baseline network-based geolocation model, and show that it outperforms previous network-based approaches (Jurgens, 2013;Rahimi et al, 2015); (2) we demonstrate that removing "celebrity" nodes (nodes with high indegrees) from the network increases geolocation accuracy and dramatically decreases network edge size; and (3) we integrate text-based geolocation priors into Modified Adsorption, and show that our unified geolocation model outperforms both text-only and network-only approaches, and achieves state-of-the-art results over three standard datasets.…”

Section: Introductionmentioning

confidence: 99%

Twitter User Geolocation Using a Unified Text and Network Prediction Model

Rahimi¹,

Cohn²,

Baldwin³

2015

Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Confere

View full text Add to dashboard Cite

We propose a label propagation approach to geolocation prediction based on Modified Adsorption, with two enhancements: (1) the removal of "celebrity" nodes to increase location homophily and boost tractability; and (2) the incorporation of text-based geolocation priors for test users. Experiments over three Twitter benchmark datasets achieve state-of-the-art results, and demonstrate the effectiveness of the enhancements.

show abstract

Section: Introductionmentioning

confidence: 99%

Twitter User Geolocation Using a Unified Text and Network Prediction Model

Rahimi¹,

Cohn²,

Baldwin³

2015

Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Confere

View full text Add to dashboard Cite

show abstract

“…We will utilize their random walk approach and other graph-based techniques, such as modified adsorption (Talukdar and Crammer 2009) to generate seed distributions. We also plan an endto-end evaluation, for instance, by employing the extracted bilingual lexicon in an MT system.…”

Section: Resultsmentioning

confidence: 99%

Extracting Translation Pairs from Comparable Corpora through Graph-based Label Propagation

Tamura

Watanabe

Sumita

et al. 2013

Journal of Natural Language Processing

View full text Add to dashboard Cite

This paper proposes a novel method for bilingual lexicon extraction from comparable corpora using graph-based label propagation. A previous study found that performance drastically decreases when the coverage of a seed lexicon is small. We address this problem by using indirect relations with bilingual seeds together with direct relations, in which each word is represented by a distribution of lexical seeds. The seed distributions are propagated over a graph that represents relations among words. Translation pairs are extracted by identifying word pairs with high similarities in the seed distributions. We propose two types of graphs: (1) a co-occurrence graph, representing co-occurrence relations between words; and (2) a similarity graph, representing context similarities between words. Evaluations on comparable corpora of English and Japanese patent documents show that our proposed graph propagation method outperforms conventional methods. Further, the similarity graph improved performance by clustering synonyms into the same translation.

show abstract

“…We use Modified Adsorption (MAD) algorithm, a graph based semi-supervised approach for our task (Talukdar and Crammer, 2009). MAD fits to our requirements specifically on two aspects.…”

Section: Methodsmentioning

confidence: 99%

“…But, lack of labeled data and other resources led us to development of a semi supervised approach for identification and analysis of derived words in Sanskrit. We use the Modified Adsorption algorithm (Talukdar and Crammer, 2009), a variant of the label propagation algorithm for the task. In this task, we effectively combine the diverse features ranging from rules in As .…”

Section: Introductionmentioning

confidence: 99%

A Graph Based Semi-Supervised Approach for Analysis of Derivational Nouns in Sanskrit

Krishna

Satuluri²,

Ponnada³

et al. 2017

Proceedings of TextGraphs-11: The Workshop on Graph-Based Methods For Natural Language Processing

View full text Add to dashboard Cite

Derivational nouns are widely used in Sanskrit corpora and is a prevalent means of productivity in the language. Currently there exists no analyser that identifies the derivational nouns. We propose a semi supervised approach for identification of derivational nouns in Sanskrit. We not only identify the derivational words, but also link them to their corresponding source words. The novelty of our work is primarily in its design of the network structure for the task. The edge weights are featurised based on the phonetic, morphological, syntactic and the semantic similarity shared between the words to be identified. We find that our model is effective for the task, even when we employ a labelled dataset which is only 5 % to that of the entire dataset.

show abstract

New Regularized Algorithms for Transductive Learning

Cited by 130 publications

References 9 publications

Twitter User Geolocation Using a Unified Text and Network Prediction Model

Twitter User Geolocation Using a Unified Text and Network Prediction Model

Extracting Translation Pairs from Comparable Corpora through Graph-based Label Propagation

A Graph Based Semi-Supervised Approach for Analysis of Derivational Nouns in Sanskrit

Contact Info

Product

Resources

About