Kaiqi Zhao

Who, Where, When, and What: A Nonparametric Bayesian Approach to Context-aware Recommendation and Search for Twitter Users

Micro-blogging services and location-based social networks, such as Twitter, Weibo, and Foursquare, enable users to post short messages with timestamps and geographical annotations. The rich spatial-temporal-semantic information of individuals embedded in these geo-annotated short messages provides exciting opportunity to develop many context-aware applications in ubiquitous computing environments. Example applications include contextual recommendation and contextual search. To obtain accurate recommendations and most relevant search results, it is important to capture users’ contextual information (e.g., time and location) and to understand users’ topical interests and intentions. While time and location can be readily captured by smartphones, understanding user’s interests and intentions calls for effective methods in modeling user mobility behavior. Here, user mobility refers to who visits which place at what time for what activity. That is, user mobility behavior modeling must consider user (Who), spatial (Where), temporal (When), and activity (What) aspects. Unfortunately, no previous studies on user mobility behavior modeling have considered all of the four aspects jointly, which have complex interdependencies. In our preliminary study, we propose the first solution named W4 (short for Who, Where, When, and What) to discover user mobility behavior from the four aspects. In this article, we further enhance W4 and propose a nonparametric Bayesian model named EW4 (short for Enhanced W4). EW4 requires no parameter tuning and achieves better results over W4 in our experiments. Given some of the four aspects of a user (e.g., time), our model is able to infer information of the other aspects (e.g., location and topical words). Thus, our model has a variety of context-aware applications, particularly in contextual search and recommendation. Experimental results on two real-world datasets show that the proposed model is effective in discovering users’ spatial-temporal topics. The model also significantly outperforms state-of-the-art baselines for various tasks including location prediction for tweets and requirement-aware location recommendation.

SAR: A sentiment-aspect-region model for user preference analysis in geo-tagged reviews

Many location based services, such as FourSquare, Yelp, TripAdvisor, Google Places, etc., allow users to compose reviews or tips on points of interest (POIs), each having a geographical coordinates. These services have accumulated a large amount of such geo-tagged review data, which allows deep analysis of user preferences in POIs. This paper studies two types of user preferences to POIs: topical-region preference and category aware topical-aspect preference. We propose a unified probabilistic model to capture these two preferences simultaneously. In addition, our model is capable of capturing the interaction of different factors, including topical aspect, sentiment, and spatial information. The model can be used in a number of applications, such as POI recommendation and user recommendation, among others. In addition, the model enables us to investigate whether people like an aspect of a POI or whether people like a topical aspect of some type of POIs (e.g., bars) in a region, which offer explanation for recommendations. Experiments on real world datasets show that the model achieves significant improvement in POI recommendation and user recommendation in comparison to the state-of-the-art methods. We also propose an efficient online recommendation algorithm based on our model, which saves up to 90% computation time.

Topic Exploration in Spatio-Temporal Document Collections

Huge amounts of data with both spatial and temporal information (e.g., geo-tagged tweets) are being generated, and are often used to share and spread personal updates, spontaneous ideas, and breaking news. We refer to such data as spatio-temporal documents. It is of great interest to explore topics in a collection of spatio-temporal documents. In this paper, we study the problem of efficiently mining topics from spatio-temporal documents within a user specified bounded region and timespan, to provide users with insights about events, trends, and public concerns within the specified region and time period. We propose a novel algorithm that is able to efficiently combine two pre-trained topic models learnt from two document sets with a bounded error, based on which we develop an efficient approach to mining topics from a large number of spatio-temporal documents within a region and a timespan. Our experimental results show that our approach is able to improve the runtime by at least an order of magnitude compared with the baselines. Meanwhile, the effectiveness of our proposed method is close to the baselines.

Annotating Points of Interest with Geo-tagged Tweets

Microblogging services like Twitter contain abundant of user generated content covering a wide range of topics. Many of the tweets can be associated to real-world entities for providing additional information for the latter. In this paper, we aim to associate tweets that are semantically related to real-world locations or Points of Interest (POIs). Tweets contain dynamic and real-time information while POIs contain relatively static information. The tweets associated with POIs provide complementary information for many applications like opinion mining and POI recommendation; the associated POIs can also be used as POI tags in Twitter. We define the research problem of annotating POIs with tweets and propose a novel supervised Bayesian Model (sBM). The model takes into account the textual, spatial features and user behaviors together with the supervised information of whether a tweet is POI-related. It is able to capture user interests in latent regions for the prediction of whether a tweet is POI-related and the association between the tweet and its most semantically related POI. On tweets and POIs collected for two cities (New York City and Singapore), we demonstrate the effectiveness of our models against baseline methods.

Towards personalized maps: mining user preferences from geo-textual data

Rich geo-textual data is available online and the data keeps increasing at a high speed. We propose two user behavior models to learn several types of user preferences from geo-textual data, and a prototype system on top of the user preference models for mining and search geo-textual data (called PreMiner) to support personalized maps. Different from existing recommender systems and data analysis systems, PreMiner highly personalizes user experience on maps and supports several applications, including user mobility & interests mining, opinion mining in regions, user recommendation, point-of-interest recommendation, and querying and subscribing on geo-textual data.

Clustering image search results by entity disambiguation

Existing key-word based image search engines return images whose title or immediate surrounding text contains the search term as a keyword. When the search term is ambiguous and means different things, the results often come in a mixed bag of different entities. This paper proposes a novel framework that understands the context and thus infers the most likely entity in the given image by disambiguating the terms in the context into the corresponding concepts from external knowledge in a process called conceptualization. The images can subsequently be clustered by the most likely associated entities. This approach outperforms the best competing image clustering techniques by 29.2% in NMI score. In addition, the framework automatically annotates each cluster of images by its key entities which allows users to quickly identify the images they want.

Querying and mining geo-textual data for exploration: Challenges and opportunities

Geo-textual data (e.g., geo-tagged tweets) is becoming increasingly available on the Web. This paper reviews recent studies on searching and mining geo-textual data for exploration, and discusses future directions along with open problems.

Efficient Similar Region Search with Deep Metric Learning

With the proliferation of mobile devices and location-based services, rich geo-tagged data is becoming prevalent and this offer great opportunities to understand different geographical regions (e.g., shopping areas). However, the huge number of regions with complicated spatial information are expensive for people to explore and understand. To solve this issue, we study the problem of searching similar regions given a user specified query region. The problem is challenging in both similarity definition and search efficiency. To tackle the two challenges, we propose a novel solution equipped by (1) a deep learning approach to learning the similarity that considers both object attributes and the relative locations between objects; and (2) an efficient branch and bound search algorithm for finding top-N similar regions. Moreover, we propose an approximation method to further improve the efficiency by slightly sacrificing the accuracy. Our experiments on three real world datasets demonstrate that our solution improves both the accuracy and search efficiency by a significant margin compared with the state-of-the-art methods.

ANR: Aspect-based Neural Recommender

Textual reviews, which are readily available on many e-commerce and review websites such as Amazon and Yelp, serve as an invaluable source of information for recommender systems. However, not all parts of the reviews are equally important, and the same choice of words may reflect a different meaning based on its context. In this paper, we propose a novel end-to-end Aspect-based Neural Recommender (ANR) to perform aspect-based representation learning for both users and items via an attention-based component. Furthermore, we model the multi-faceted process behind how users rate items by estimating the aspect-level user and item importance by adapting the neural co-attention mechanism. Our proposed model concurrently address several shortcomings of existing recommender systems, and a thorough experimental study on 25 benchmark datasets from Amazon and Yelp shows that ANR significantly outperforms recently proposed state-of-the-art baselines such as DeepCoNN, D-Attn and ALFM.

A system for region search and exploration

With the increasing popularity of mobile devices and location based services, massive amount of geo-textual data (e.g., geo-tagged tweets) is being generated everyday. Compared with traditional spatial data, the textual dimension of geo-textual data greatly enriches the data. Meanwhile, the spatial dimension of geo-textual data also adds a semantically rich new aspect to textual data. The large volume, together with its rich semantics, calls for the need for data exploration. First, it has many applications to retrieve a region for exploration that satisfies user-specified conditions (e.g., the size and shape of the region) while maximizing some other conditions (e.g., the relevance to the query keywords of the objects in the region). Second, it is useful to mine and explore the topics of the geo-textual data within a (specified or retrieved) region and perhaps a timespan. This demonstration proposal presents the main ideas of our system, the R eg I on S earch and E xploration System (RISE), for efficiently supporting region search and exploration, and our demonstration plan.