Filters








24 Hits in 9.0 sec

Author Profiling: Gender Prediction from Tweets and Images: Notebook for PAN at CLEF 2018

Yaakov HaCohen-Kerner, Yair Yigal, Elyashiv Shayovitz, Daniel Miller, Toby P. Breckon
2018 Conference and Labs of the Evaluation Forum  
' gender where for each author, 100 tweets and 10 images are provided.  ...  In this paper, we describe the participation of our teams (yigal18 and miller18, both teams contain the same people, but in another order) in the PAN 2018 shared task on author profiling, identifying authors  ...  This work was partially funded by the Jerusalem College of Technology (Lev Academic Center) and we gratefully acknowledge its support.  ... 
dblp:conf/clef/HaCohen-KernerY18 fatcat:y5k6qwemurh6tkvrjzq5wt4gc4

Stacked Gender Prediction from Tweet Texts and Images: Notebook for PAN at CLEF 2018

Giovanni Ciccone, Arthur Sultan, Léa Laporte, Elöd Egyed-Zsigmond, Alaa Alhamzeh, Michael Granitzer
2018 Conference and Labs of the Evaluation Forum  
This paper describes our participation at the PAN 2018 Author Profiling shared task. Given texts and images from some Twitter's authors, the goal is to estimate their genders.  ...  The first one, based on previous PAN Author Profiling editions, concerns gender prediction from texts.  ...  The objective of the 2018 Pan author profiling shared task and of our approach is thus to study if and how text and images taken from tweets can be used to predict the gender of the authors of those tweets  ... 
dblp:conf/clef/CicconeSLEAG18 fatcat:aczx2avambhpjny5s6cgbzkoey

Gender Prediction From Tweets With Convolutional Neural Networks: Notebook for PAN at CLEF 2018

Erhan Sezerer, Ozan Polatbilek, Özge Sevgili, Selma Tekir
2018 Conference and Labs of the Evaluation Forum  
This paper presents a system 1 developed for the author profiling task of PAN at CLEF 2018 .  ...  Our architecture was able to obtain competitive results on three languages provided by the PAN 2018 author profiling challenge with an average accuracy of 75.1% on local runs and 70.23% on the submission  ...  Conclusion We have described a system submitted to the author profiling task of PAN at CLEF 2018. A CNN architecture is proposed which takes the characters of each tweet's text as an input.  ... 
dblp:conf/clef/SezererPST18 fatcat:y76xpcszkzcbzhp5ne3wjd47fq

Word Unigram Weighing for Author Profiling at PAN 2018: Notebook for PAN at CLEF 2018

Pius von Däniken, Ralf Grubenmann, Mark Cieliebak
2018 Conference and Labs of the Evaluation Forum  
We present our system for the author profiling task at PAN 2018 on gender identification on Twitter.  ...  Our submission achieved accuracies of 77.42% for English, 74.64% for Spanish, and 73.20% for Arabic tweets. It ranked 15th out of 23 competitors.  ...  In this work, we describe our submission to the author profiling task at PAN 2018 [8, 9] on gender identification based on text and images posted by users of social media.  ... 
dblp:conf/clef/DanikenGC18 fatcat:o3qziu57jzdsxbf7txdhxkmsey

Multimodal Author Profiling for Twitter: Notebook for PAN at CLEF 2018

Braja Gopal Patra, Kumar Gourav Das, Dipankar Das
2018 Conference and Labs of the Evaluation Forum  
This paper describes the systems submitted to author profiling task at PAN-2018 using multimodal (textual and image) Twitter datasets provided by the organizers and the aim is to identify the author's  ...  An image captioning system was used to extract captions from images. Mainly latent semantic analysis, word embeddings, and stylistic features were extracted from tweets as well as captions.  ...  This time, the AP task at PAN-2018 is performed on three different languages (English, Spanish, Arabic), and the datasets contain tweets and images from Twitter.  ... 
dblp:conf/clef/PatraD018 fatcat:j2y24dgmabh3hnelfgrnfhznti

Author Profiling based on Text and Images: Notebook for PAN at CLEF 2018

Luka Stout, Robert Musters, Chris Pool
2018 Conference and Labs of the Evaluation Forum  
In this paper we describe our participation in the PAN 2018 shared task of Author Profiling. In this task we identify the gender of authors based on written text and shared images.  ...  The image classification is done by finding selfies and predicting the gender of the person on those images using CNNs.  ...  Dataset Description and Preprocessing The PAN 2018 Author Profiling [6] training set consists of text in three different languages and images grouped by authors, who are labeled by gender and language  ... 
dblp:conf/clef/StoutMP18 fatcat:6d25kwc2cbf4jhq43kzbnd5ewa

Multilingual Author Profiling using LSTMs: Notebook for PAN at CLEF 2018

Roy Khristopher Bayot, Teresa Gonçalves
2018 Conference and Labs of the Evaluation Forum  
This paper shows one approach of the Universidade de Évora for author profiling for PAN 2018. The approach mainly consists of using word vectors and LSTMs for gender classification.  ...  Using the PAN 2018 dataset, we achieved an accuracy of 67.60% for Arabic, 77.16% for English, and 68.73% for Spanish gender classification.  ...  Dataset In the current edition of PAN 2018 [34] for author profiling [25] , the task is to predict gender based on text, images, or both. The current dataset has 1500 Fig. 1 .  ... 
dblp:conf/clef/BayotG18 fatcat:7gjohpuljvbbjnfc75vuj7f5u4

A Straightforward Multimodal Approach for Author Profiling: Notebook for PAN at CLEF 2018

Mario Ezra Aragón, Adrián Pastor López-Monroy
2018 Conference and Labs of the Evaluation Forum  
In this paper we evaluate different strategies from the literature for text and image classification at PAN 2018.  ...  The main objective of this shared task is the identification of the gender of different users by using tweets and images posted.  ...  Conclusions In this notebook we presented an approach in order to determine the gender of a user using the tweets and images they post.  ... 
dblp:conf/clef/AragonL18 fatcat:xc6upgmhdbgxffaymd3m27hvwi

Combining Textual and Visual Representations for Multimodal Author Profiling: Notebook for PAN at CLEF 2018

Sebastián Sierra, Fabio A. González
2018 Conference and Labs of the Evaluation Forum  
Author Profiling studies the common use of language inside those demographic groups. This work describes our proposed method for the PAN 2018 Author Profiling shared task.  ...  This year's task consisted of evaluating gender using multimodal information (text and images) which was extracted from Twitter users.  ...  PAN Author Profiling 2018 Shared Task PAN-AP 2018 shared task consisted of classifying correctly the gender of an user of Twitter [19] .  ... 
dblp:conf/clef/SierraG18 fatcat:hmduiukzhzclppl526z4ob2q5q

Character-based Convolutional Neural Network and ResNet18 for Twitter Author Profiling: Notebook for PAN at CLEF 2018

Nils Schaetti
2018 Conference and Labs of the Evaluation Forum  
The evaluations are based on three collections of tweets and images (PAN AUTHOR PROFILING task at CLEF 2018).  ...  This paper describes and evaluates a mixing model for multimodal author profiling using character-based Convolutional Neural Networks (CNN) for tweet classification and ResNet18 for images.  ...  For the PAN CLEF 2018 evaluation campaign, three test collections of tweets and images were created, one for each of the following languages : English, Spanish and Arabic.  ... 
dblp:conf/clef/Schaetti18 fatcat:tlkr3hnnjve2pgkyxzrqxv7tza

Author Profiling using Word Embeddings with Subword Information: Notebook for PAN at CLEF 2018

Rafael Felipe Sandroni Dias, Ivandré Paraboni
2018 Conference and Labs of the Evaluation Forum  
We present a simple experiment on multilingual author profiling as proposed by the PAN-CLEF 2018 shared task, focusing on the issue of gender identification from Twitter text in English, Spanish and Arabic  ...  Our proposal makes use of word embeddings enriched with char n-gram information, and outperforms a majority class baseline.  ...  The second author received financial support from FAPESP grant nro. 2016/14223-0.  ... 
dblp:conf/clef/DiasP18 fatcat:qydhp4kztncyvb4omxkpa3rdta

Using Translated Data to Improve Deep Learning Author Profiling Models: Notebook for PAN at CLEF 2018

Robert Veenhoven, Stan Snijders, Daniël van der Hall, Rik van Noord
2018 Conference and Labs of the Evaluation Forum  
In this report on our participation in the PAN shared task on author profiling, we describe our attempt to identify the gender of authors using their posted tweets and images.  ...  The data of interest are tweets in the English, Spanish and Arabic languages as well as images.  ...  While new to the PAN Author profiling shared task, images have been used for author gender identification before, for example, [37] used a combination of image type and image content to classify author  ... 
dblp:conf/clef/VeenhovenSHN18 fatcat:n635j4g7fzhepo5qkxnroler64

Multilingual Gender Classification with Multi-view Deep Learning: Notebook for PAN at CLEF 2018

Matej Martinc, Blaz Skrlj, Senja Pollak
2018 Conference and Labs of the Evaluation Forum  
We present the results of a gender identification performed on the data set of tweets and images prepared for the PAN 2018 Author profiling shared task.  ...  The proposed approach was 8 th in the global ranking of PAN 2018 Author profiling shared task.  ...  The best gender profiling approaches within the last year's PAN shared task on tweets [17] achieved the accuracy of 0.8233 for English and 0.8321 for Spanish.  ... 
dblp:conf/clef/MartincSP18 fatcat:luxhvxxnfjh5vali26rxubqiou

Text and Image Synergy with Feature Cross Technique for Gender Identification: Notebook for PAN at CLEF 2018

Takumi Takahashi, Takuji Tahara, Koki Nagatani, Yasuhide Miura, Tomoki Taniguchi, Tomoko Ohkuma
2018 Conference and Labs of the Evaluation Forum  
This paper describes a neural network model for the author profiling task of PAN@CLEF 2018.  ...  We tackle the author profiling task using neural networks for texts and images.  ...  PAN 2018: Author Profiling Task [13] is identifying the user's gender from tweets that are contained texts and images in three languages (English, Spanish, and Arabic).  ... 
dblp:conf/clef/TakahashiTNMTO18 fatcat:yvrbku6fpffolltlnsxekb6yca

Twitter Text and Image Gender Classification with a Logistic Regression N-Gram Model: Notebook for PAN at CLEF 2018

Moniek Nieuwenhuis, Jeroen Wilkens
2018 Conference and Labs of the Evaluation Forum  
We present our participation in the PAN 2018 Author Profiling shared task, classifying authors on gender for English, Arabic and Spanish.  ...  Our highest scores on the PAN 2018 test dataset are accuracies of 81.2% for English using only text-based features, 78.7% for Arabic using both text-and image-based features and 80.3% for Spanish using  ...  In this paper, we describe our approach for the Author Profiling shared task at PAN 2018 [17] .  ... 
dblp:conf/clef/NieuwenhuisW18 fatcat:mtrxwihhkbdnnjsoijesccssju
« Previous Showing results 1 — 15 out of 24 results