We consider here the problem of building a never-ending language learner; that is, an intelligent computer agent that runs forever and that each day must (1) extract, or read, information from the web to populate a growing structured knowledge base, and (2) learn to perform this task better than on the previous day. In particular, we propose an approach and a set of design principles for such an agent, describe a partial implementation of such a system that has already learned to extract a knowledge base containing over 242,000 beliefs with an estimated precision of 74% after running for 67 days, and discuss lessons learned from this preliminary attempt to build a never-ending learning agent.

Traditional information extraction systems adopt pipeline strategies, which are highly ineffective and suffer from several problems such as error propagation. Typically, pipeline models fail to produce highly-accurate final output. On the other hand, there has been growing interest in integrated or joint models which explore mutual benefits and perform multiple subtasks simultaneously to avoid problems caused by pipeline models. However, building such systems usually increases computational complexity and requires considerable engineering. This paper presents a general, strongly-coupled, and bidirectional architecture based on discriminatively trained factor graphs for information extraction. First we introduce joint factors connecting variables of relevant subtasks to capture dependencies and interactions between them. We then propose a strong bidirectional MCMC sampling inference algorithm which allows information to flow in both directions to find the approximate MAP solution for all subtasks. Extensive experiments on entity identification and relation extraction using real-world data illustrate the promise of our approach.

Research on information extraction (IE) seeks to distill relational tuples from natural language text, such as the contents of the WWW. Most IE work has focussed on identifying static facts, encoding them as binary relations. This is unfortunate, because the vast majority of facts are fluents, only holding true during an interval of time. It is less helpful to extract PresidentOf(Bill-Clinton, USA) without the temporal scope 1/20/93 — 1/20/01. This paper presents TIE, a novel, information-extraction system, which distills facts from text while inducing as much temporal information as possible. In addition to recognizing temporal relations between times and events, TIE performs global inference, enforcing transitivity to bound the start and ending times for each event. We introduce the notion of temporal entropy as a way to evaluate the performance of temporal IE systems and present experiments showing that TIE outperforms three alternative approaches.

It is often desirable to extract structured information from raw web pages for better information browsing, query answering, and pattern mining. many such Information Extraction (IE) technologies are costly and applying them at the web-scale is impractical. In this paper, we propose a novel prioritization approach where candidate pages from the corpus are ordered according to their expected contribution to the extraction results and those with higher estimated potential are extracted earlier. Systems employing this approach can stop the extraction process at any time when the resource gets scarce (i.e., not all pages in the corpus can be processed), without worrying about wasting extraction effort on unimportant pages. More specifically, we define a novel notion to measure the value of extraction results and design various mechanisms for estimating a candidate page’s contribution to this value. We further design and build the Extraction Prioritization (EP) system with efficient scoring and scheduling algorithms, and experimentally demonstrate that EP significantly outperforms the naive approach and is more flexible than the classifier approach.

With the development of Web 2.0, sentiment analysis has now become a popular research problem to tackle. Recently, topic models have been introduced for the simultaneous analysis for topics and the sentiment in a document. These studies, which jointly model topic and sentiment, take the advantage of the relationship between topics and sentiment, and are shown to be superior to traditional sentiment analysis tools. However, most of them make the assumption that, given the parameters, the sentiments of the words in the document are all independent. In our observation, in contrast, sentiments are expressed in a coherent way. The local conjunctive words, such as “and” or “but”, are often indicative of sentiment transitions. In this paper, we propose a major departure from the previous approaches by making two linked contributions. First, we assume that the sentiments are related to the topic in the document, and put forward a joint sentiment and topic model, i.e. Sentiment-LDA. Second, we observe that sentiments are dependent on local context. Thus, we further extend the Sentiment-LDA model to Dependency-Sentiment-LDA model by relaxing the sentiment independent assumption in Sentiment-LDA. The sentiments of words are viewed as a Markov chain in Dependency-Sentiment-LDA. Through experiments, we show that exploiting the sentiment dependency is clearly advantageous, and that the Dependency-Sentiment-LDA is an effective approach for sentiment analysis.

Vast amounts of text on the Web are unstructured and ungrammatical, such as classified ads, auction listings, forum postings, etc. We call such text posts. Despite their inconsistent structure and lack of grammar, posts are full of useful information. This paper presents work on semi-automatically building tables of relational information, called reference sets, by analyzing such posts directly. Reference sets can be applied to a number of tasks such as ontology maintenance and information extraction. Our reference-set construction method starts with just a small amount of background knowledge, and constructs tuples representing the entities in the posts to form a reference set. We also describe an extension to this approach for the special case where even this small amount of background knowledge is impossible to discover and use. To evaluate the utility of the machine-constructed reference sets, we compare them to manually constructed reference sets in the context of reference-set-based information extraction. Our results show the reference sets constructed by our method outperform manually constructed reference sets. We also compare the reference-set-based extraction approach using the machine-constructed reference set to supervised extraction approaches using generic features. These results demonstrate that using machine-constructed reference sets outperforms the supervised methods, even though the supervised methods require training data.

We use quantitative media (blogs, and news as a comparison) data generated by a large-scale natural language processing (NLP) text analysis system to perform a comprehensive and comparative study on how company related news variables anticipates or reflects the company's stock trading volumes and financial returns. Building on our findings, we give a sentiment-based market-neutral trading strategy which gives consistently favorable returns with low volatility over a long period. Our results are significant in confirming the performance of general blog and news sentiment analysis methods over broad domains and sources. Moreover, several remarkable differences between news and blogs are also identified.

We investigate the impact of a discussion snippet's overall sentiment on a user's willingness to read more of a discussion. Using sentiment analysis, we constructed positive, neutral, and negative discussion snippets using the discussion topic and a sample comment from discussions taking place around content on an enterprise social networking site. We computed personalized snippet recommendations for a subset of users and conducted a survey to test how these recommendations were perceived. Our experimental results show that snippets with high sentiments are better discussion "teasers."

This paper addresses the problem of automatically generating domain-specific sentiment clues. The main idea is to bootstrap from a small seed set and generate new clues by using dependencies and collocation information between sentiment clues and sentence-level topics that would be a primary subject of sentiment expression (e.g., event, company, and person). The experiments show that the aggregated clues are effective for sentiment classification.

We connect measures of public opinion measured from polls with sentiment measured from text. We analyze several surveys on consumer confidence and political opinion over the 2008 to 2009 period, and find they correlate to sentiment word frequencies in contempora- neous Twitter messages. While our results vary across datasets, in several cases the correlations are as high as 80%, and capture important large-scale trends. The re- sults highlight the potential of text streams as a substi- tute and supplement for traditional polling. consumer confidence and political opinion, and can also pre- dict future movements in the polls. We find that temporal smoothing is a critically important issue to support a suc- cessful model.