Hongzhao Huang

Collective Tweet Wikification based on Semi-supervised Graph Regularization

Wikification for tweets aims to automatically identify each concept mention in a tweet and link it to a concept referent in a knowledge base (e.g., Wikipedia). Due to the shortness of a tweet, a collective inference model incorporating global evidence from multiple mentions and concepts is more appropriate than a noncollecitve approach which links each mention at a time. In addition, it is challenging to generate sufficient high quality labeled data for supervised models with low cost. To tackle these challenges, we propose a novel semi-supervised graph regularization model to incorporate both local and global evidence from multiple tweets through three fine-grained relations. In order to identify semanticallyrelated mentions for collective inference, we detect meta path-based semantic relations through social networks. Compared to the state-of-the-art supervised model trained from 100% labeled data, our proposed approach achieves comparable performance with 31% labeled data and obtains 5% absolute F1 gain with 50% labeled data.

Be Appropriate and Funny: Automatic Entity Morph Encoding

Internet users are keen on creating different kinds of morphs to avoid censorship, express strong sentiment or humor. For example, in Chinese social media, users often use the entity morph “¹ ? b (Instant Noodles)” to refer to “h 8 · (Zhou Yongkang)” because it shares one character “· (Kang)” with the well-known brand of instant noodles “·� (Master Kang)”. We developed a wide variety of novel approaches to automatically encode proper and interesting morphs, which can effectively pass decoding tests 1 .

Context-aware Entity Morph Decoding

People create morphs, a special type of fake alternative names, to achieve certain communication goals such as expressing strong sentiment or evading censors. For example, “Black Mamba”, the name for a highly venomous snake, is a morph that Kobe Bryant created for himself due to his agility and aggressiveness in playing basketball games. This paper presents the first end-to-end context-aware entity morph decoding system that can automatically identify, disambiguate, verify morph mentions based on specific contexts, and resolve them to target entities. Our approach is based on an absolute “cold-start” it does not require any candidate morph or target entity lists as input, nor any manually constructed morph-target pairs for training. We design a semi-supervised collective inference framework for morph mention extraction, and compare various deep learning based approaches for morph resolution. Our approach achieved significant improvement over the state-of-the-art method (Huang et al., 2013), which used a large amount of training data. 1

Cross-media Cross-genre Information Ranking based on Multi-media Information Networks

Current web technology has brought us a scenario that information about a certain topic is widely dispersed in data from different domains and data modalities, such as texts and images from news and social media. Automatic extraction of the most informative and important multimedia summary (e.g. a ranked list of inter-connected texts and images) from massive amounts of cross-media and cross-genre data can significantly save users’ time and effort that is consumed in browsing. In this paper, we propose a novel method to address this new task based on automatically constructed Multi-media Information Networks (MiNets) by incorporating cross-genre knowledge and inferring implicit similarity across texts and images. The facts from MiNets are exploited in a novel random walk-based algorithm to iteratively propagate ranking scores across multiple data modalities. Experimental results demonstrated the effectiveness of our MiNets-based approach and the power of cross-media cross-genre inference.

Free-form text summarization in social sensing

This demonstration illustrates an information aggregation and summarization service for social sensing applications. Social sensing, using mobile phones and other networked devices in the possession of individuals, has gained significant popularity in recent years. In some cases, the information collected is structured, such as numeric data from temperature sensors, accelerometers, or GPS devices. Aggregate statistical properties, such as expected values, standard deviations, and outliers, can be easily computed, and can be used to summarize the data set. In other cases, however, the collection includes unstructured data types such as text or images with textual annotations. The concepts of expected values and outliers are harder to define, yet it is still important to be able to aggregate and summarize the data. We demonstrate a system which can automatically summarize real-time textual data common to social sensing applications. Specifically, we focus on text messages that describe events in the environment. The output of our service provides a reliable summary of observations that can be used in many contexts from military intelligence to participatory sensing campaigns.

Demo abstract: Free-form text summarization in social sensing

This demonstration illustrates an information aggregation and summarization service for social sensing applications. Social sensing, using mobile phones and other networked devices in the possession of individuals, has gained significant popularity in recent years. In some cases, the information collected is structured, such as numeric data from temperature sensors, accelerometers, or GPS devices. Aggregate statistical properties, such as expected values, standard de-viations, and outliers, can be easily computed, and can be used to summarize the data set. In other cases, however, the collection includes unstructured data types such as text or images with textual annotations. The concepts of expected values and outliers are harder to define, yet it is still important to be able to aggregate and summarize the data. We demonstrate a system which can automatically summarize real-time textual data common to social sensing applications. Specifically, we focus on text messages that describe events in the environment. The output of our service provides a reliable summary of observations that can be used in many contexts from military intelligence to participatory sensing campaigns.