Adversarial Learning for Weakly-Supervised Social Network Alignment

Li, Chaozhuo; Wang, Senzhang; Wang, Yukun; Yu, Philip S.; Liang, Yanbo; Liu, Yun; Li, Zhoujun

doi:10.1609/aaai.v33i01.3301996

Cited by 77 publications

(38 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the inception and extensive applications of knowledge representation of natural or complex data such as languages and multi-media, the GAN framework is also widely applied to classification missions for data in representation spaces (e.g., vector spaces of matrices) rather than just for the original data in feature spaces. Some researchers also extended applications of GAN to more challenging problems such as recommendation systems [11] or social network alignment problems [12].…”

Section: Adversarial Learning Of Neural Networkmentioning

confidence: 99%

Sentiment-Guided Adversarial Learning for Stock Price Prediction

Zhang

Wang

et al. 2021

Front. Appl. Math. Stat.

View full text Add to dashboard Cite

Prediction of stock prices or trends have attracted financial researchers’ attention for many years. Recently, machine learning models such as neural networks have significantly contributed to this research problem. These methods often enable researchers to take stock-related factors such as sentiment information into consideration, improving prediction accuracies. At present, Long Short-Term Memory (LSTM) networks is one of the best techniques known to learn knowledge from time-series data and to predict future tendencies. The inception of generative adversarial networks (GANs) also provides researchers with diversified and powerful methods to explore the stock prediction problem. A GAN network consists of two sub-networks known as generator and discriminator, which work together to minimize maximum loss on both actual and simulated data. In this paper, we developed a sentiment-guided adversarial learning and predictive models of stock prices, adopting a popular variation of GAN called conditional GAN (CGAN). We adopted an LSTM network in the generator and a multilayer perceptron (MLP) network in the discriminator. After extensively pre-processing historical stock price datasets, we analyzed the sentiment information from daily tweets and computed sentiment scores as an additional model feature. Our experiments demonstrated that the average forecast accuracies of the CGAN models were improved with sentiment data. Moreover, our GAN and CGAN models outperformed LSTM and other traditional methods on 11 out of 36 processed stock price datasets, potentially playing a part in ensemble methods.

show abstract

Section: Adversarial Learning Of Neural Networkmentioning

confidence: 99%

Sentiment-Guided Adversarial Learning for Stock Price Prediction

Zhang

Wang

et al. 2021

Front. Appl. Math. Stat.

View full text Add to dashboard Cite

show abstract

“…[28][29][30]. Also, there have been various network embedding methods [31][32][33][34] that exhibited the state-of-the-art performance. However, our story embedding studies [12,13] keep applying the Word2Vec-based methods [26,27].…”

Section: The Character Network Is a Representation Of Relationships Bmentioning

confidence: 99%

“…There are two main reasons. First, the existing methods mostly aim at social networks [31,34] or knowledge graphs [35,36]. They mainly attempted to embed nodes or edges for the link prediction, community detection, and so forth.…”

Section: The Character Network Is a Representation Of Relationships Bmentioning

confidence: 99%

Learning Hierarchical Representations of Stories by Using Multi-Layered Structures in Narrative Multimedia

Lee

Jung

Kim

2020

Sensors

View full text Add to dashboard Cite

Narrative works (e.g., novels and movies) consist of various utterances (e.g., scenes and episodes) with multi-layered structures. However, the existing studies aimed to embed only stories in a narrative work. By covering other granularity levels, we can easily compare narrative utterances that are coarser (e.g., movie series) or finer (e.g., scenes) than a narrative work. We apply the multi-layered structures on learning hierarchical representations of the narrative utterances. To represent coarser utterances, we consider adjacency and appearance of finer utterances in the coarser ones. For the movies, we suppose a four-layered structure (character roles ∈ characters ∈ scenes ∈ movies) and propose three learning methods bridging the layers: Char2Vec, Scene2Vec, and Hierarchical Story2Vec. Char2Vec represents a character by using dynamic changes in the character’s roles. To find the character roles, we use substructures of character networks (i.e., dynamic social networks of characters). A scene describes an event. Interactions between characters in the scene are designed to describe the event. Scene2Vec learns representations of a scene from interactions between characters in the scene. A story is a series of events. Meanings of the story are affected by order of the events as well as their content. Hierarchical Story2Vec uses sequential order of scenes to represent stories. The proposed model has been evaluated by estimating the similarity between narrative utterances in real movies.

show abstract

“…DeepLink [Zhou et al, 2018] A deep model for alignment. SNNA [Li et al, 2019a] A WGAN-based alignment method. moana [Zhang et al, 2019a] An algorithm for alignment at multiple levels.…”

Section: The Behavior Analysis Objectivementioning

confidence: 99%

BANANA: when Behavior ANAlysis meets social Network Alignment

Ren

Zhang

et al. 2020

Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence

View full text Add to dashboard Cite

Recently, aligning users among different social networks has received significant attention. However, most of the existing studies do not consider users’ behavior information during the aligning procedure and thus still suffer from the poor learning performance. In fact, we observe that social network alignment and behavior analysis can benefit from each other. Motivated by such an observation, we propose to jointly study the social network alignment problem and user behavior analysis problem. We design a novel end-to-end framework named BANANA. In this framework, to leverage behavior analysis for social network alignment at the distribution level, we design an earth mover’s distance based alignment model to fuse users’ behavior information for more comprehensive user representations. To further leverage social network alignment for behavior analysis, in turn, we design a temporal graph neural network model to fuse behavior information in different social networks based on the alignment result. Two models above can work together in an end-to-end manner. Through extensive experiments on real-world datasets, we demonstrate that our proposed approach outperforms the state-of-the-art methods in the social network alignment task and the user behavior analysis task, respectively.

show abstract

Adversarial Learning for Weakly-Supervised Social Network Alignment

Cited by 77 publications

References 18 publications

Sentiment-Guided Adversarial Learning for Stock Price Prediction

Sentiment-Guided Adversarial Learning for Stock Price Prediction

Learning Hierarchical Representations of Stories by Using Multi-Layered Structures in Narrative Multimedia

BANANA: when Behavior ANAlysis meets social Network Alignment

Contact Info

Product

Resources

About