67,083 Hits in 2.8 sec

Discriminative Neural Topic Models [article]

Gaurav Pandey, Ambedkar Dukkipati
2017 arXiv   pre-print
We propose a neural network based approach for learning topics from text and image datasets.  ...  The model makes no assumptions about the conditional distribution of the observed features given the latent topics.  ...  Using the proposed model, we obtain a mean purity of 49.3% for the learned topics.  ... 
arXiv:1701.06796v2 fatcat:r7ly6dgyxffxpec7qe4oiirzxm

Cross Domain Regularization for Neural Ranking Models Using Adversarial Learning [article]

Daniel Cohen, Bhaskar Mitra, Katja Hofmann, W. Bruce Croft
2018 arXiv   pre-print
We use an adversarial discriminator and train our neural ranking model on a small set of domains.  ...  Unlike traditional learning to rank models that depend on hand-crafted features, neural representation learning models learn higher level features for the ranking task by training on large datasets.  ...  However, in the presence of the adversarial discriminator both the models show significant improvement in performance on all held out topics.  ... 
arXiv:1805.03403v1 fatcat:sl2tum3325etdjtroyigsqg6xy

Neural Topic Modeling with Deep Mutual Information Estimation [article]

Kang Xu and Xiaoqiu Lu and Yuan-fang Li and Tongtong Wu and Guilin Qi and Ning Ye and Dong Wang and Zheng Zhou
2022 arXiv   pre-print
The emerging neural topic models make topic modeling more easily adaptable and extendable in unsupervised text mining.  ...  In this paper, we propose a neural topic model which incorporates deep mutual information estimation, i.e., Neural Topic Modeling with Deep Mutual Information Estimation(NTM-DMIE).  ...  [3] is a Gaussian Softmax topic model parameterized with neural networks. • NTM [17] is a neural topic model which incorporates a topic coherence objective. • Scholar [16] is a supervised neural  ... 
arXiv:2203.06298v1 fatcat:vls2jisyzjgbhhy2ws77yfuti4

Modeling documents with Generative Adversarial Networks [article]

John Glover
2016 arXiv   pre-print
We propose a model that is based on the recently proposed Energy-Based GAN, but instead uses a Denoising Autoencoder as the discriminator network.  ...  Document representations are extracted from the hidden layer of the discriminator and evaluated both quantitatively and qualitatively.  ...  This was one of the factors that led to the development of an autoregressive neural topic model called DocNADE [11] , which is based on the NADE model [12] .  ... 
arXiv:1612.09122v1 fatcat:klpmak2gmncbpc25gbjf64ptja

Discriminative Topic Modeling with Logistic LDA [article]

Iryna Korshunova, Hanchen Xiong, Mateusz Fedoryszak, Lucas Theis
2020 arXiv   pre-print
In contrast to other recent topic models designed to handle arbitrary inputs, our model does not sacrifice the interpretability and principled motivation of LDA.  ...  In particular, our model can easily be applied to groups of images, arbitrary text embeddings, and integrates well with deep neural networks.  ...  Discussion and conclusion We presented logistic LDA, a neural topic model that preserves most of LDA's inductive biases while giving up its generative component in favour of a discriminative approach,  ... 
arXiv:1909.01436v2 fatcat:qfg3ihatv5ac7kikgk2ifyier4

Neural Topic Modeling with Bidirectional Adversarial Training [article]

Rui Wang, Xuemeng Hu, Deyu Zhou, Yulan He, Yuxuan Xiong, Chenchen Ye, Haiyang Xu
2020 arXiv   pre-print
for neural topic modeling.  ...  Recent years have witnessed a surge of interests of using neural topic models for automatic topic extraction from text, since they avoid the complicated mathematical derivations for model inference as  ...  One possible way in addressing this limitation is through neural topic models which employ blackbox inference mechanism with neural networks.  ... 
arXiv:2004.12331v1 fatcat:xsbtuzkfk5fvvmfm57ddt4izg4

Prelims [chapter]

2021 Developing an Effective Model for Detecting Trade-based Market Manipulation  
In Chapter 7, a hybrid model using advanced data mining techniques like Artificial Neural Network and Genetic Algorithm is developed.  ...  Further, the performance of this hybrid model is compared with a conventional standalone model based on Quadratic Discriminant Function (QDF).  ... 
doi:10.1108/978-1-80117-396-420211022 fatcat:r66t3gca75cqnmecpa7kvrphdu

Natural Language Processing. A Machine Learning Perspective

Julia Ive
2021 Computational Linguistics  
Following clear logic, generative models are introduced before discriminative ones (e.g., Chapter 7 from Part II introduces generative sequence labelling and Chapter 8 introduces discriminative sequence  ...  This textbook introduces NLP from the ML standpoint elaborating on fundamental approaches and algorithms used in the field: such as statistical and deep learning models, generative and discriminative,  ...  Then discriminative models and reranking are elaborated on.  ... 
doi:10.1162/coli_r_00423 fatcat:rbztldsm6bgjlemmajt7wvpb6y

ATM:Adversarial-neural Topic Model [article]

Rui Wang and Deyu Zhou and Yulan He
2019 arXiv   pre-print
To address these limitations, we propose a topic modeling approach based on Generative Adversarial Nets (GANs), called Adversarial-neural Topic Model (ATM).  ...  Topic models are widely used for thematic structure discovery in text. But traditional topic models often require dedicated inference procedures for specific tasks at hand.  ...  Adversarial-neural Topic Model We propose the Advesarial-neural Topic Model (ATM) as shown in Figure 1 .  ... 
arXiv:1811.00265v2 fatcat:ydx5bt3pxvcobmv5uhrkflrx2u

Efficient text generation of user-defined topic using generative adversarial networks [article]

Chenhan Yuan, Yi-chin Huang, Cheng-Hung Tsai
2020 arXiv   pre-print
Then, the second discriminator is re-trained with the generator if the topic or sentiment for text generation is modified.  ...  the topic.  ...  In order to make these models fit the distribution of real text data better, the number of parameters of text generation models based on neural network are increased, which means that training these neural  ... 
arXiv:2006.12005v1 fatcat:66rfgm5pcvghbi2mxbku5yuor4

Contrastive Learning for Neural Topic Model [article]

Thong Nguyen, Anh Tuan Luu
2021 arXiv   pre-print
words; (2) it restricts the ability to integrate external information, such as sentiments of the document, which has been shown to benefit the training of neural topic model.  ...  Experimental results show that our framework outperforms other state-of-the-art neural topic models in three common benchmark datasets that belong to various domains, vocabulary sizes, and document lengths  ...  by the neural topic model.  ... 
arXiv:2110.12764v1 fatcat:2oz4s2hfnvdjbpne6inqclpufa

A Unified Neural Coherence Model

Han Cheol Moon, Tasnim Mohiuddin, Shafiq Joty, Chi Xu
2019 Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)  
Recently, neural approaches to coherence modeling have achieved state-of-the-art results in several evaluation tasks.  ...  With extensive experiments on local and global discrimination tasks, we demonstrate that our proposed model outperforms existing models by a good margin, and establish a new state-of-the-art.  ...  In this paper, we propose a unified neural model that incorporates sentence grammar (intentional structure), discourse relations, attention and topic structures in a single framework.  ... 
doi:10.18653/v1/d19-1231 dblp:conf/emnlp/MoonMJC19 fatcat:py6riojf25huvfgmvsco3thgfq

Neural Net Models for Open-Domain Discourse Coherence [article]

Jiwei Li, Dan Jurafsky
2017 arXiv   pre-print
We study both discriminative models that learn to distinguish coherent from incoherent discourse, and generative models that produce coherent text, including a novel neural latent-variable Markovian generative  ...  In this paper, we describe domain-independent neural models of discourse coherence that are capable of measuring multiple aspects of coherence in existing sentences and can maintain coherence while generating  ...  The extended version of the discriminative model described in this work significantly outperforms the parse-tree based recursive models presented in Li and Hovy (2014) as well as all non-neural baselines  ... 
arXiv:1606.01545v3 fatcat:ylniiicirjb63abbtitntbdsqe

A Supervised Neural Autoregressive Topic Model for Simultaneous Image Classification and Annotation [article]

Yin Zheng, Yu-Jin Zhang, Hugo Larochelle
2013 arXiv   pre-print
Recently, a new type of topic model called the Document Neural Autoregressive Distribution Estimator (DocNADE) was proposed and demonstrated state-of-the-art performance for document modeling.  ...  Specifically, we propose SupDocNADE, a supervised extension of DocNADE, that increases the discriminative power of the hidden topic features by incorporating label information into the training objective  ...  What distinguishes DocNADE from other topic models is its reliance on a neural network architecture.  ... 
arXiv:1305.5306v1 fatcat:tymbgeiadzagro45xwq6ntmm7i

Learning Term Discrimination

Jibril Frej, Philippe Mulhem, Didier Schwab, Jean-Pierre Chevallet
2020 Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval  
Along with tf (that only reflects the importance of a term in a document), traditional IR models use term discrimination values (TDVs) such as inverse document frequency (idf) to favor discriminative terms  ...  In this work, we propose to learn TDVs for document indexing with shallow neural networks that approximate traditional IR ranking functions such as TF-IDF and BM25.  ...  • AP88-89 with topics 51-200 and 15 856 positive qrels • FT91-94 with topics 251-450 and 6 486 positive qrels • LA with topics 301-450 and 3 535 positive qrels We use title of topics as queries.  ... 
doi:10.1145/3397271.3401211 dblp:conf/sigir/FrejMSC20 fatcat:7dqdoez63rb3xcewhcpouz7lwq
« Previous Showing results 1 — 15 out of 67,083 results