Luis Gravano

Snowball : extracting relations from large plain-text collections

Text documents often contain valuable structured data that is hidden Yin regular English sentences. This data is best exploited infavailable as arelational table that we could use for answering precise queries or running data mining tasks.We explore a technique for extracting such tables from document collections that requires only a handful of training examples from users. These examples are used to generate extraction patterns, that in turn result in new tuples being extracted from the document collection.We build on this idea and present our Snowball system. Snowball introduces novel strategies for generating patterns and extracting tuples from plain-text documents.At each iteration of the extraction process, Snowball evaluates the quality of these patterns and tuples without human intervention,and keeps only the most reliable ones for the next iteration. In this paper we also develop a scalable evaluation methodology and metrics for our task, and present a thorough experimental evaluation of Snowball and comparable techniques over a collection of more than 300,000 newspaper documents.

Beyond Trending Topics: Real-World Event Identification on Twitter

User-contributed messages on social media sites such as Twitter have emerged aspowerful, real-time means of information sharing on the Web. These short messages tend to reflect a variety of events in real time, making Twitter particularly well suited as a source of real-time event content. In this paper, we explore approaches for analyzing the stream of Twitter messages to distinguish between messages about real-world events andnon-event messages. Our approach relies on a rich family of aggregatestatistics of topically similar message clusters. Large-scale experiments over millions of Twitter messages show the effectiveness of our approach for surfacing real-world event content on Twitter.

GlOSS : text-source discovery over the Internet

The dramatic growth of the Internet has created a new problem for users: location of the relevant sources of documents. This article presents a framework for (and experimentally analyzes a solution to) this problem, which we call the text-source discovery problem. Our approach consists of two phases. First, each text source exports its contents to a centralized service. Second, users present queries to the service, which returns an ordered list of promising text sources. This article describes GlOSS, Glossary of Servers Server, with two versions: bGlOSS, which provides a Boolean query retrieval model, and vGlOSS, which provides a vector-space retrieval model. We also present hGlOSS, which provides a decentralized version of the system. We extensively describe the methodology for measuring the retrieval effectiveness of these systems and provide experimental evidence, based on actual data, that all three systems are highly effective in determining promising text sources for a given query.

STHoles: a multidimensional workload-aware histogram

Attributes of a relation are not typically independent. Multidimensional histograms can be an effective tool for accurate multiattribute query selectivity estimation. In this paper, we introduce STHoles, a “workload-aware” histogram that allows bucket nesting to capture data regions with reasonably uniform tuple density. STHoles histograms are built without examining the data sets, but rather by just analyzing query results. Buckets are allocated where needed the most as indicated by the workload, which leads to accurate query selectivity estimations. Our extensive experiments demonstrate that STHoles histograms consistently produce good selectivity estimates across synthetic and real-world data sets and across query workloads, and, in many cases, outperform the best multidimensional histogram techniques that require access to and processing of the full data sets during histogram construction.

Computing Geographical Scopes of Web Resources

Many information resources on the web are relevant primarily to limited geographical communities. For instance, web sites containing information on restaurants, theaters, and apartment rentals are relevant primarily to web users in geographical proximity to these locations. In contrast, other information resources are relevant to a broader geographical community. For instance, an on-line newspaper may be relevant to users across the United States. Unfortunately, current web search engines largely ignore the geographical scope of web resources. In this paper, we introduce techniques for automatically computing the geographical scope of web resources, based on the textual content of the resources, as well as on the geographical distribution of hyperlinks to them. We report an extensive experimental evaluation of our strategies using real web data. Finally, we describe a geographicallyaware search engine that we have built to showcase our techniques.

The Stanford Digital Library Metadata Architecture

Abstract. The overall goal of the Stanford Digital Library project is to provide an infrastructure that affords interoperability among heterogeneous, autonomous digital library services. These services include both search services and remotely usable information processing facilities. In this paper, we survey and categorize the metadata required for a diverse set of Stanford Digital Library services that we have built. We then propose an extensible metadata architecture that meets these requirements. Our metadata architecture fits into our established infrastructure and promotes interoperability among existing and de-facto metadata standards. Several pieces of this architecture are implemented; others are under construction. The architecture includes attribute model proxies, attribute model translation services, metadata information facilities for search services, and local metadata repositories. In presenting and discussing the pieces of the architecture, we show how they address our motivating requirements. Together, these components provide, exchange, and describe metadata for information objects and metadata for information services. We also consider how our architecture relates to prior, relevant work on these two types of metadata.

STARTS: Stanford proposal for Internet meta-searching

Document sources are available everywhere, both within the internal networks of organizations and on the Internet. Even individual organizations use search engines from different vendors to index their internal document collections. These search engines are typically incompatible in that they support different query models and interfaces, they do not return enough information with the query results for adequate merging of the results, and finally, in that they do not export metadata about the collections that they index (e.g., to assist in resource discovery). This paper describes STARTS, an emerging protocol for Internet retrieval and search that facilitates the task of querying multiple document sources. STARTS has been developed in a unique way. It is not a standard, but a group effort coordinated by Stanfords Digital Library project, and involving over 11 companies and organizations. The objective of this paper is not only to give an overview of the STARTS protocol proposal, but also to discuss the process that led to its definition.

The effectiveness of GIOSS for the text database discovery problem

The popularity of on-line document databases has led to a new problem: finding which text databases (out of many candidate choices) are the most relevant to a user. Identifying the relevant databases for a given query is the text database discovery problem. The first part of this paper presents a practical solution based on estimating the result size of a query and a database. The method is termed GlOSS—Glossary of Servers Server. The second part of this paper evaluates the effectiveness of GlOSS based on a trace of real user queries. In addition, we analyze the storage cost of our approach.

Text joins in an RDBMS for web data integration

The integration of data produced and collected across autonomous, heterogeneous web services is an increasingly important and challenging problem. Due to the lack of global identifiers, the same entity (e.g., a product) might have different textual representations across databases. Textual data is also often noisy because of transcription errors, incomplete information, and lack of standard formats. A fundamental task during data integration is matching of strings that refer to the same entity. In this paper, we adopt the widely used and established cosine similarity metric from the information retrieval field in order to identify potential string matches across web sources. We then use this similarity metric to characterize this key aspect of data integration as a join between relations on textual attributes, where the similarity of matches exceeds a specified threshold. Computing an exact answer to the text join can be expensive. For query processing efficiency, we propose a sampling-based join approximation strategy for execution in a standard, unmodified relational database management system (RDBMS), since more and more web sites are powered by RDBMSs with a web-based front end. We implement the join inside an RDBMS, using SQL queries, for scalability and robustness reasons. Finally, we present a detailed performance evaluation of an implementation of our algorithm within a commercial RDBMS, using real-life data sets. Our experimental results demonstrate the efficiency and accuracy of our techniques.

Probe, count, and classify: categorizing hidden web databases

The contents of many valuable web-accessible databases are only accessible through search interfaces and are hence invisible to traditional web “crawlers.” Recent studies have estimated the size of this “hidden web” to be 500 billion pages, while the size of the “crawlable” web is only an estimated two billion pages. Recently, commercial web sites have started to manually organize web-accessible databases into Yahoo!-like hierarchical classification schemes. In this paper, we introduce a method for automating this classification process by using a small number of query probes. To classify a database, our algorithm does not retrieve or inspect any documents or pages from the database, but rather just exploits the number of matches that each query probe generates at the database in question. We have conducted an extensive experimental evaluation of our technique over collections of real documents, including over one hundred web-accessible databases. Our experiments show that our system has low overhead and achieves high classification accuracy across a variety of databases.