Manifold-ranking Process Research Articles

A Novel Biased Diversity Ranking Model for Query-Oriented Multi-Document Summarization

Query-oriented multi-document summarization (QMDS) attempts to generate a concise piece of text byextracting sentences from a target document collection, with the aim of not only conveying the key content of that corpus, also, satisfying the information needs expressed by that query. Due to its great applicable value, QMDS has been intensively studied in recent decades. Three properties are supposed crucial for a good summary, i.e., relevance, prestige and low redundancy (orso-called diversity). Unfortunately, most existing work either disregarded the concern of diversity, or handled it with non-optimized heuristics, usually based on greedy sentences election. Inspired by the manifold-ranking process, which deals with query-biased prestige, and DivRank algorithm which captures query-independent diversity ranking, in this paper, we propose a novel biased diversity ranking model, named ManifoldDivRank, for query-sensitive summarization tasks. The top-ranked sentences discovered by our algorithm not only enjoy query-oriented high prestige, more importantly, they are dissimilar with each other. Experimental results on DUC2005and DUC2006 benchmark data sets demonstrate the effectiveness of our proposal.

Ranking on Data Manifold with Sink Points

Ranking is an important problem in various applications, such as Information Retrieval (IR), natural language processing, computational biology, and social sciences. Many ranking approaches have been proposed to rank objects according to their degrees of relevance or importance. Beyond these two goals, diversity has also been recognized as a crucial criterion in ranking. Top ranked results are expected to convey as little redundant information as possible, and cover as many aspects as possible. However, existing ranking approaches either take no account of diversity, or handle it separately with some heuristics. In this paper, we introduce a novel approach, Manifold Ranking with Sink Points (MRSPs), to address diversity as well as relevance and importance in ranking. Specifically, our approach uses a manifold ranking process over the data manifold, which can naturally find the most relevant and important data objects. Meanwhile, by turning ranked objects into sink points on data manifold, we can effectively prevent redundant objects from receiving a high rank. MRSP not only shows a nice convergence property, but also has an interesting and satisfying optimization explanation. We applied MRSP on two application tasks, update summarization and query recommendation, where diversity is of great concern in ranking. Experimental results on both tasks present a strong empirical performance of MRSP as compared to existing ranking approaches.

Iterative Feedback Based Manifold-Ranking for Update Summary

The update summary as defined for the DUC2007 new task aims to capture evolving information of a single topic over time. It delivers focused information to a user who has already read a set of older documents covering the same topic. This paper presents a novel manifold-ranking frame based on iterative feedback mechanism to this summary task. The topic set is extended by using the summarization of previous timeslices and the first sentences of documents in current timeslice. Iterative feedback mechanism is applied to model the dynamically evolving characteristic and represent the relay propagation of information in temporally evolving data. Modified manifold-ranking process also can naturally make use of both the relationships among all the sentences in the documents and relationships between the topic and the sentences. The ranking score for each sentence obtained in the manifold-ranking process denotes the importance of sentence biased towards topic, and then the greedy algorithm is employed to rerank the sentences for removing the redundant information. The summary is produced by choosing the sentences with high ranking score. Experiments on dataset of DUC2007 update task demonstrate the encouraging performance of the proposed approach.

Towards a unified approach to document similarity search using manifold-ranking of blocks

Document similarity search (i.e. query by example) aims to retrieve a ranked list of documents similar to a query document in a text corpus or on the Web. Most existing approaches to similarity search first compute the pairwise similarity score between each document and the query using a retrieval function or similarity measure (e.g. Cosine), and then rank the documents by the similarity scores. In this paper, we propose a novel retrieval approach based on manifold-ranking of document blocks (i.e. a block of coherent text about a subtopic) to re-rank a small set of documents initially retrieved by some existing retrieval function. The proposed approach can make full use of the intrinsic global manifold structure of the document blocks by propagating the ranking scores between the blocks on a weighted graph. First, the TextTiling algorithm and the VIPS algorithm are respectively employed to segment text documents and web pages into blocks. Then, each block is assigned with a ranking score by the manifold-ranking algorithm. Lastly, a document gets its final ranking score by fusing the scores of its blocks. Experimental results on the TDT data and the ODP data demonstrate that the proposed approach can significantly improve the retrieval performances over baseline approaches. Document block is validated to be a better unit than the whole document in the manifold-ranking process.