GenWiki: A Dataset of 1.3 Million Content-Sharing Text and Graphs for Unsupervised Graph-to-Text Generation

Jin, Zhijing; Guo, Qinglan; Qiu, Xipeng; Zhang, Zheng

doi:10.18653/v1/2020.coling-main.217

Cited by 20 publications

(33 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It also used in-domain unlabeled documents during training, which we do not use. Jin et al (2020) demonstrated that the choice of seed keywords has a significant impact on the model's accuracy. STM,S label is the result of STM using only unigrams in the category name as seed keywords.…”

Section: Results Of Coarse-grained Contextual Classificationmentioning

confidence: 99%

“…To investigate the contribution of the in-domain unlabeled document to STM's superior performance, we trained an STM model with the manually-curated keywords in Jin et al (2020) and the Wikipedia dataset we used to train wiki2cat (denoted as STM,D wiki ). There is a noticeable decrease in performance in STM,D wiki without indomain unlabeled documents.…”

Section: Results Of Coarse-grained Contextual Classificationmentioning

confidence: 99%

“…We evaluated our method using three contextual classification datasets. The first two are coarsegrained evaluation datasets published by Jin et al (2020) covering all IAB tier-1 categories except for "News" (totaling 22 categories). The datasets are collected using different methods (news-crawl-v2 dataset (nc-v2) by mapping from news categories; browsing dataset by manual labelling) and contain 2,127 and 1,501 documents separately 7 .…”

Section: Evaluation Datasetsmentioning

confidence: 99%

“…For coarse-grained classification, STM used hand-curated seed words while STM,S label used category names as seed words. Both were trained by Jin et al (2020) on a private in-domain dataset. We also trained STM using our Wikipedia dataset, referred to as STM,D wiki .…”

Section: Contextual-eval-datasetmentioning

confidence: 99%

“…Both Doc2vec and STM require unlabeled training corpus. We copy the coarsegrained classification result for Doc2vec, STM, and STM,S label from Jin et al (2020). For fine-grained classification, we train Doc2vec and STM,S label using the same set of Wikipedia documents as in wiki2cat.…”

Section: Contextual-eval-datasetmentioning

confidence: 99%

See 4 more Smart Citations

Proceedings of the Fifteenth Workshop on Graph-Based Methods for Natural Language Processing (TextGraphs-15)

2021

View full text Add to dashboard Cite

show abstract

Section: Results Of Coarse-grained Contextual Classificationmentioning

confidence: 99%

Section: Results Of Coarse-grained Contextual Classificationmentioning

confidence: 99%

Section: Evaluation Datasetsmentioning

confidence: 99%

Section: Contextual-eval-datasetmentioning

confidence: 99%

Section: Contextual-eval-datasetmentioning

confidence: 99%

See 3 more Smart Citations

Proceedings of the Fifteenth Workshop on Graph-Based Methods for Natural Language Processing (TextGraphs-15)

2021

View full text Add to dashboard Cite

show abstract

Adversarial Machine Learning in Text Processing: A Literature Survey

et al. 2022

View full text Add to dashboard Cite

Machine learning algorithms represent the intelligence that controls many information systems and applications around us. As such, they are targeted by attackers to impact their decisions. Text created by machine learning algorithms has many types of applications, some of which can be considered malicious especially if there is an intention to present machine-generated text as human-generated. In this paper, we surveyed major subjects in adversarial machine learning for text processing applications. Unlike adversarial machine learning in images, text problems and applications are heterogeneous. Thus, each problem can have its own challenges. We focused on some of the evolving research areas such as: malicious versus genuine text generation metrics, defense against adversarial attacks, and text generation models and algorithms. Our study showed that as applications of text generation will continue to grow in the near future, the type and nature of attacks on those applications and their machine learning algorithms will continue to grow as well. Literature survey indicated an increasing trend in using pre-trained models in machine learning. Word/sentence embedding models and transformers are examples of those pre-trained models. Adversarial models may utilize same or similar pre-trained models as well. In another trend related to text generation models, literature showed effort to develop universal text perturbations to be used in both black-and whitebox attack settings. Literature showed also using conditional GANs to create latent representation for writing types. This usage will allow for a seamless lexical and grammatical transition between various writing styles. In text generation metrics, research trends showed developing successful automated or semi-automated assessment metrics that may include human judgement. Literature showed also research trends of designing and developing new memory models that increase performance and memory utilization efficiency without validating real-time constraints. Many research efforts evaluate different defense model approaches and algorithms. Researchers evaluated different types of targeted attacks, and methods to distinguish human versus machine generated text.

show abstract