John Shepherd

The University of New South Wales, Computer Science, Faculty Member

Followers

Following

Co-authors

Public Views

I'm a Senior Lecturer and Deputy Head of School (Education) in the School of Computer Science and Engineering at UNSW, Sydney, Australia. When I get a chance to do research, in amongst all of my teaching and admin, I work on information extraction.
Address: Australia

less

Michael Sheng

The University of Adelaide

Iaeme

Interests

Uploads

Papers by John Shepherd

CMVF: a novel dimension reduction scheme for efficient indexing in a large image database

International Conference on Management of Data, 2003

INTRODUCTION In recent years, due to the increasing volumes of multimedia data in the World Wide ... more INTRODUCTION In recent years, due to the increasing volumes of multimedia data in the World Wide Web, digital library, biomedicine and other applications, efficient content based similarity search in large image databases is gaining considerable research attentions. As a result, various indexing methods known as Spatial Access Methods (SAMs) and metric trees have been proposed to support this kind of

Download

The Carrick vision and computing education: four case studies in multi-institutional collaboration

Proceedings of the …, Jan 30, 2007

The Carrick Institute is an initiative of the Australia federal government. It is aimed at genera... more The Carrick Institute is an initiative of the Australia federal government. It is aimed at generating strategic change in Australian University education, via grants and other awards to approximately $20 million annually. By previous Australian standards, the potential funding for projects is large. However, the Carrick Institute has a well focused vision, and grant applications need to be aligned with that vision. This paper first describes some key aspects of the Carrick vision, before describing four multi-institutional computing ...

Download

Extracting Information from Semistructured Data

Lecture Notes in Computer Science, 2002

ABSTRACT

Information Extraction via Automatic Pattern Discovery in Identified Region

Lecture Notes in Computer Science, 2004

Information extraction using two-phase pattern discovery

Proceedings of the 27th annual international conference on Research and development in information retrieval - SIGIR '04, 2004

This paper presents a new two-phase pattern (2PP) discovery technique for information extraction.... more This paper presents a new two-phase pattern (2PP) discovery technique for information extraction. 2PP consists of orthographic pattern discovery (OPD) and semantic pattern discovery (SPD) where the OPD determines the structural features from an identified region of a document and the SPD discovers a dominant semantic pattern for the region via inference, apposition and analogy. Then the discovered pattern is applied back into the region to extract required data items through pattern matching. We evaluated 2PP using 6500 data items and obtained effective result.

Web object collection: here or there?

Although the advantages to performing the collection of web objects at the site where the objects... more Although the advantages to performing the collection of web objects at the site where the objects reside seem obvious, most of the popular search engines use their own centralized or coarsely distributed collectors (robots). Sending each object to the collector is almost always the worst option with respect to resource usage. An alternative is to distribute the collection process by

Efficient Nearest-Neighbour Searches Using Weighted Euclidean Metric

British National Conference on Databases, 1998

Building an index tree is a common approach to speed upthe k nearest neighbour search in large da... more Building an index tree is a common approach to speed upthe k nearest neighbour search in large databases of many-dimensionalrecords. Many applications require varying distance metrics by putting aweight on different dimensions. The main problem with k nearest neighboursearches using weighted euclidean metrics in a high dimensionalspace is whether the searches can be done efficiently We present a solutionto this

Download

Processbook: Towards Social Network-Based Personal Process Management

Lecture Notes in Business Information Processing, 2013

Robust User Community-Aware Landmark Photo Retrieval

by John Shepherd and Jolin Wu

Lecture Notes in Computer Science, 2015

Techniques for Supporting E cient Content-based Retrieval in Multimedia Databases

Teaching Novices to Debug

Knowledge Processing and Information Retrieval-Information Extraction via Automatic Pattern Discovery in Identified Region

Improved named entity recognition: patterns in columns model (PCM)

An Optimization Method for Proportionally Diversifying Search Results

by John Shepherd and Jolin Wu

The problem of diversifying search results has attracted much attention, since diverse results ca... more The problem of diversifying search results has attracted much attention, since diverse results can provide non-redundant information and cover multiple query-related topics. However, existing approaches typically assign equal importance to each topic. In this paper, we propose a novel method for diversification: proportionally diversifying search results. Specifically, we study the problem of returning a top-k ranked list where the number of candidates in each topic is proportional to the popularity degree of that topic with respect to the query. We obtain such a top-k proportionally diverse list by maximizing our proposed objective function and we prove that this is an NP-hard problem. We further propose a greedy heuristic to efficiently obtain a good approximate solution.

Download

Indexing schemes and file organisations for the NU-Prolog deductive database system

Co-ranking Images and Tags via Random Walks on a Heterogeneous Graph

by John Shepherd and Jolin Wu

Ranking on image search results has attracted considerable attentions. Despite many graph-based r... more Ranking on image search results has attracted considerable attentions. Despite many graph-based ranking algorithms have demonstrated remarkable success, most of their applications are limited to single image-networks such as the network of tags associated with images. In this paper, we investigate the problem of co-ranking images and tags attached in a heterogeneous network, which consists of three graphs: the image graph connecting images, the tag graph connecting tags attached to the images, as well as the image-tag graph connecting the above two graphs together. Observing that existing ranking approaches do not consider images and tags simultaneously, a novel co-ranking method via random walks on all three graphs is proposed to significantly improve the ranking effectiveness on both images and tags. Experimental results conducted on three benchmark data sets show that our approach outperforms the state-of-the-art local ranking approaches for image ranking and tag ranking and scales well on large scale data sets.

Download

Efficient image and tag co-ranking: a bregman divergence optimization method

by John Shepherd and Jolin Wu

Ranking on image search has attracted considerable attentions. Many graph-based algorithms have b... more Ranking on image search has attracted considerable attentions. Many graph-based algorithms have been proposed to solve this problem. Despite their remarkable success, these approaches are restricted to their separated image networks. To improve the ranking performance, one effective strategy is to work beyond the separated image graph by leveraging fruitful information from manual semantic labeling (i.e., tags) associated with images, which leads to the technique of co-ranking images and tags, a representative method that aims to explore the reinforcing relationship between image and tag graphs. The idea of co-ranking is implemented by adopting the paradigm of random walks. However, there are two problems hidden in co-ranking remained to be open: the high computational complexity and the problem of out-ofsample. To address the challenges above, in this paper, we cast the co-ranking process into a Bregman divergence optimization framework under which we transform the original random walk into an equivalent optimal kernel matrix learning problem. Enhanced by this new formulation, we derive a novel extension to achieve a better performance for both insample and out-of-sample cases. Extensive experiments are conducted to demonstrate the effectiveness and efficiency of our approach.

Download

Techniques for supporting efficient content-based retrieval in multimedia databases

This paper is a brief survey of the current state of the technology used in supporting efficientc... more This paper is a brief survey of the current state of the technology used in supporting efficientcontent-based retrieval in index-based multimedia databases that use feature vectors to representthe multimedia data. One of the main obstacles in performing efficient content-based multimediaretrieval is the high-dimensionality of the feature vectors used to characterise the data. Previoussurveys on multidimensional access structures tried to cope

Enhancing text classification using synopses extraction

This paper describes a novel approach to document classification that uses decision-tree machine ... more This paper describes a novel approach to document classification that uses decision-tree machine learning based on a succinct vector of important terms in each document. The succinct vector itself is generated by a machine-learning approach which builds parsers that can identify significant features in a document by partitioning it into regions based on low-level document characteristics. The fact that the

Automatic affiliation extraction from calls-for-papers

by John Shepherd and Rozita Rastan

CMVF: a novel dimension reduction scheme for efficient indexing in a large image database

International Conference on Management of Data, 2003

Download

The Carrick vision and computing education: four case studies in multi-institutional collaboration

Proceedings of the …, Jan 30, 2007

Download

Extracting Information from Semistructured Data

Lecture Notes in Computer Science, 2002

ABSTRACT

Information Extraction via Automatic Pattern Discovery in Identified Region