Filters








278 Hits in 5.4 sec

Learning to Describe Video with Weak Supervision by Exploiting Negative Sentential Information

Haonan Yu, Jeffrey Siskind
2015 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this paper, we learn to describe video by discriminatively training positive sentential labels against negative ones in a weakly supervised fashion: the meaning representations (i.e., HMMs) of individual  ...  It is more appealing from a linguistic point of view, for word models for all parts of speech to be learned simultaneously from whole sentences, a hypothesis suggested by some linguists for child language  ...  Government is authorized to reproduce and distribute reprints for Government purposes, notwithstanding any copyright notation herein.  ... 
doi:10.1609/aaai.v29i1.9790 fatcat:jxie2ez4mvgzjpdjtcz2dkn5au

Sentence Directed Video Object Codiscovery

Haonan Yu, Jeffrey Mark Siskind
2017 International Journal of Computer Vision  
Video object codiscovery can leverage the weak semantic constraint implied by sentences that describe the video content.  ...  Although the semantic information employed is usually simple and weak, it can greatly boost performance by constraining the hypothesized object locations.  ...  , and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  ... 
doi:10.1007/s11263-017-1018-6 fatcat:7jfvkgvvureavdjjqcszuyrcie

A Compositional Framework for Grounding Language Inference, Generation, and Acquisition in Video

Haonan Yu, N. Siddharth, Andrei Barbu, Jeffrey Mark Siskind
2015 The Journal of Artificial Intelligence Research  
In the third, a corpus of video clips is paired with sentences which describe some of the events in those clips and the meanings of the words in those sentences are learned.  ...  In the first, a video clip along with a sentence are taken as input and the participants in the event described by the sentence are highlighted, even when the clip depicts multiple similar simultaneous  ...  Government is authorized to reproduce and distribute reprints for Government purposes, notwithstanding any copyright notation herein.  ... 
doi:10.1613/jair.4556 fatcat:ruegimpgvva37ddmsaxuru4fty

A review of affective computing: From unimodal analysis to multimodal fusion

Soujanya Poria, Erik Cambria, Rajiv Bajpai, Amir Hussain
2017 Information Fusion  
With the proliferation of videos posted online (e.g., on YouTube, Facebook, Twitter) for product reviews, movie reviews, political views, and more, affective computing research has increasingly evolved  ...  As part of this review, we carry out an extensive study of different categories of state-of-the-art fusion techniques, followed by a critical analysis of potential performance improvements with multimodal  ...  learning to develop robust features, in both supervised and unsupervised settings.  ... 
doi:10.1016/j.inffus.2017.02.003 fatcat:ytebhjxlz5bvxcdghg4wxbvr6a

An overview of event extraction and its applications [article]

Jiangwei Liu, Liangyu Min, Xiaohong Huang
2021 arXiv   pre-print
With the rapid development of information technology, online platforms have produced enormous text resources.  ...  As a particular form of Information Extraction (IE), Event Extraction (EE) has gained increasing popularity due to its ability to automatically extract events from human language.  ...  Dor et al. [102] use rules to automatically extract weak labels for event mentions describing economic events.  ... 
arXiv:2111.03212v1 fatcat:o3oagnjrybh3vapvvp7twgjtuu

Extracting Lexically Divergent Paraphrases from Twitter

Wei Xu, Alan Ritter, Chris Callison-Burch, William B. Dolan, Yangfeng Ji
2014 Transactions of the Association for Computational Linguistics  
Using this principled latent variable model alone, we achieve the performance competitive with a state-of-the-art method which combines a latent space model with a feature-based supervised classifier.  ...  We present MULTIP (Multi-instance Learning Paraphrase Model), a new model suited to identify paraphrases within the short messages on Twitter.  ...  Yangfeng Ji is supported by a Google Faculty Research Award awarded to Jacob Eisenstein.  ... 
doi:10.1162/tacl_a_00194 fatcat:54bwppjf55asjmpi7cicq46vt4

A Survey of Opinion Mining and Sentiment Analysis [chapter]

Bing Liu, Lei Zhang
2012 Mining Text Data  
Moreover, it is also known that human analysis of text information is subject to considerable biases, e.g., people often pay greater attention to opinions that are consistent with their own preferences  ...  The average human reader will have difficulty identifying relevant sites and accurately summarizing the information and opinions contained in them.  ...  Classification based on Supervised Learning Sentiment classification obviously can be formulated as a supervised learning problem with three classes, positive, negative and neutral.  ... 
doi:10.1007/978-1-4614-3223-4_13 fatcat:vpgpw7fvpzglfmla5w2x4wapwu

A Neural Multi-sequence Alignment TeCHnique (NeuMATCH)

Pelin Dogan, Boyang Li, Leonid Sigal, Markus Gross
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
The alignment of heterogeneous sequential data (video to text) is an important and challenging problem.  ...  This flexible architecture supports a large variety of alignment tasks, including one-to-one, one-to-many, skipping unmatched elements, and (with extensions) non-monotonic alignment.  ...  Weak supervision can be introduced as optimization constraints.  ... 
doi:10.1109/cvpr.2018.00912 dblp:conf/cvpr/DoganLSG18 fatcat:ou3dvllpinelvjeeov3zahiz24

The Customer is Always Right: Sentiment Analysis for Bank Service Quality
Клиент всегда прав: анализ тональности текста в отзывах о качестве банковского обслуживания

Elena G. Brunova, Yulia V. Bidulya
2017 Tyumen State University Herald Humanities Research Humanitates  
However, with the Web, especially with the explosive growth of the usergenerated content on the Web in the past few years, the world has been transformed.  ...  In this chapter, we only focus on opinion expressions that convey people's positive or negative sentiments.  ...  I thank my former and current students for working with me on this fascinating topic: Xiaowen Ding, Murthy Ganapathibhotla, Minqing Hu, Nitin Jindal, Guang Qiu (visiting student from Zhejiang University  ... 
doi:10.21684/2411-197x-2017-3-1-72-89 fatcat:f6kqhljzybbd5ks3dp5bvdzcsu

A Neural Multi-sequence Alignment TeCHnique (NeuMATCH) [article]

Pelin Dogan, Boyang Li, Leonid Sigal, Markus Gross
2018 arXiv   pre-print
The alignment of heterogeneous sequential data (video to text) is an important and challenging problem.  ...  This flexible architecture supports a large variety of alignment tasks, including one-to-one, one-to-many, skipping unmatched elements, and (with extensions) non-monotonic alignment.  ...  Weak supervision can be introduced as optimization constraints.  ... 
arXiv:1803.00057v2 fatcat:tyuaiprkuzdedlp2jzv43fhrza

Sentiment Analysis and Opinion Mining [chapter]

Lei Zhang, Bing Liu
2017 Encyclopedia of Machine Learning and Data Mining  
Extraction by exploiting opinion and target relations 3. Extraction using supervised learning 4.  ...  The authors also experimented with a semi-supervised learning method exploiting the idea that a spammer tends to write many fake reviews.  ... 
doi:10.1007/978-1-4899-7687-1_907 fatcat:iy5ty44cyzbrtodxfo7osy3iu4

Contrastive Learning of Sociopragmatic Meaning in Social Media [article]

Chiyu Zhang, Muhammad Abdul-Mageed, Ganesh Jawahar
2022 arXiv   pre-print
To bridge this gap, we propose a novel framework for learning task-agnostic representations transferable to a wide range of sociopragmatic tasks (e.g., emotion, hate speech, humor, sarcasm).  ...  Our framework outperforms other contrastive learning frameworks for both in-domain and out-of-domain data, across both the general and few-shot settings.  ...  InfoDCL is a distantly supervised contrastive learning (DCL) framework that exploits surrogate labels as a proxy for gold labels and incorporates corpus-level information to capture inter-class relationships  ... 
arXiv:2203.07648v2 fatcat:6zmhiogvirdlznoaqonyuesc54

Sentiment/Subjectivity Analysis Survey for Languages other than English [article]

Mohammed Korayem, Khalifeh Aljadda, David Crandall
2016 arXiv   pre-print
The paper presents a separate section devoted to Arabic sentiment analysis.  ...  The second type of systems involves reusing or transferring sentiment resources from English to the target language. The third type of methods is based on using language independent methods.  ...  After creating lexicons of positive and negative words, they introduced Semi-supervised learning to build sentiment classifier.  ... 
arXiv:1601.00087v3 fatcat:zi4ikkge7ng4tfumhia4ytulwe

Human/Human Conversation Understanding [chapter]

Gokhan Tur, Dilek Hakkani-Tür
2011 Spoken Language Understanding  
Human/Human Conversation Understanding 3 which has access to extra information tries to help the speaker.  ...  technical seminars dominated by lecture-style presentations.  ...  They concluded that while active learning does not help significantly for this task, exploiting unlabeled data by using minimal supervision is effective when the dialog act tag sequence is also modeled  ... 
doi:10.1002/9781119992691.ch9 fatcat:2quzefll5re2lp2ytbasw6nthu

Sentic patterns: Dependency-based rules for concept-level sentiment analysis

Soujanya Poria, Erik Cambria, Grégoire Winterstein, Guang-Bin Huang
2014 Knowledge-Based Systems  
information associated with natural language.  ...  structure information that is key for properly detecting the polarity conveyed by natural language opinions.  ...  Acknowledgements This work was supported by a grant from Singapore Ministry of Defence (MINDEF) under Project MINDEF-NTU-JPP/12/02/05 (M406340000).  ... 
doi:10.1016/j.knosys.2014.05.005 fatcat:awgw4u7uobg6roujumzf6hqvzq
« Previous Showing results 1 — 15 out of 278 results