Lupu M., Mayer K., Tait J., Trippe A.J. (eds.) Current Challenges in Patent Information Retrieval

Springer, 2011. — 431 p.Patent Information Retrieval is an economically important activity. Today’s economy is becoming increasingly knowledge-based and intellectual property in the form of patents plays a vital role in this growth. Between 1998 and 2008, the number of patent applications filed worldwide grew by more than 50 percent. The number of granted patents worldwide continues to increase, albeit at a slower rate than at its peak in 2006 (18%), when some 727,000 patents were granted. The substantial increase in patents granted is due, in part, to efforts by patent offices to reduce backlogs as well as the significant growth in the number of patents granted by China and, to a lesser extent in the more recent years, by the Republic of Korea. According to these statistics, the total number of patents in force worldwide at the end of 2008 was approximately 6.7 million (WIPO report 2010). A prior art search might have to cover as many as 70 million patents. By combining data from Ocean Tomo’s Intangible Asset Market Value Survey, and Standard and Poor’s 1200 Index we can estimate that the global value of patents exceeds US$10 trillion in 2009.
A patent is a bargain between the inventor and the state. The inventor must teach the community how to make the product, and use the techniques he/she has invented in return for a limited monopoly which gives him a set time to exploit his invention and realise its value. Patents are used for many reasons, e.g. to protect inventions, to create value and to monitor competitive activities in a field. Much knowledge is distilled through patents, which is never published elsewhere. Thus patents form an important knowledge resource—e.g. much technical information represented in patents is not represented in scientific literature—and are at the same time important legal documents.
Despite the overall increase in patent applications and grants, a situation of economic downturn, such as the one the world has experienced in 2008, leads to a reduction in patent applications and grants (as indicated by preliminary figures published by WIPO for 2009). This is, to some extent, explained by the high costs involved in applying for a patent, particularly for small enterprises. The costs of the pre-application process, the long duration of the application process and the corresponding uncertainty in the long-term economy in such periods of economic downturn need to be addressed by changing the way we search the patent and non-patent literature. Both the Intellectual Property (IP) professionals and the Information Retrieval (IR) scientists can see this book as a challenge: for the former, in terms of adapting to new tools; for the latter, in terms of creating better tools for an obviously difficult task; for both, in terms of engaging in exchange and cooperation.
In the past 10 or 15 years, general information retrieval and Web search engines have made tremendous advances. And still, we see a huge gap between the technologies which, on the one hand, were emerging from research labs and in use by major internet search engines, in e-commerce, and in enterprise search systems, and, on the other, the systems in day-to-day use by the patent search communities.
It has been estimated that since 1991, when the US Federal National Institute of Standards and Technology (NIST) began its Text Retrieval Conference (TREC) evaluation campaign, the available information retrieval and search systems have improved 40% or more in their ability to find relevant documents. And yet the technologies underlying the patent search system were largely unaffected by these changes. Patent searchers generally use the same technology as in the 1980s. Boolean specification of searches and set-based retrieval are the norm rather than the ranked retrieval systems used by Google and the like. Tools in some areas have moved on significantly: some providers have semantic analysis tools, others effective visualisation mechanisms for patent documents. And yet there has not been the kind of revolution in patent search which Google had represented for Web search.
In the past few years, the Information Retrieval Facility (a not-for-profit research institution based in Vienna, Austria) has organised a series of events to bring together leading researchers in IR with those who practice and use patent search, to establish the interdisciplinary dialogue between the IR and the IP communities and to create a discursive as well as empirical space for sustainable discussion and innovation.
In the first Information Retrieval Facility Symposium in Vienna in 2007 (www.irfs.at), a distinguished audience of information retrieval scientists and patent search specialists started to explore the reasons for the knowledge gap. It turned out that academic researchers were often unaware of the specialised needs of the patent searchers: for example, they needed a degree of transparency quite unlike the casual Web searchers, upon which the academics mainly focussed. The patent searchers were often unaware of the advances made in other areas, and how they had been achieved. There were difficulties in finding (and using) a common, comprehensible vocabulary. In the course of that first Symposium, and through subsequent IRF symposia and other joint activities, such as the CLEF-IP and TREC-CHEM tracks, the PaIR and Aspire workshops, major progress has been made in developing a common understanding, and even an agenda between search researchers and technologists and the patent search community.
This book is part of the development of that joint understanding. Its origins lie in the idea of producing post-proceedings for the first IRF Symposium. That idea was not fully followed up, in part because of pressure to produce more practical, actionoriented work, and in part because many of the participants felt their approaches were at too early a stage for formal publication. In the course of the following years it became apparent there really was a demand to produce a volume which was accessible to both the patent search community and to the information retrieval research community; to provide a collected and organized introduction to the work and views of the two sides of the emerging patent search research and innovation community; and to provide a coherent and organised view of what has been achieved and, perhaps even more significantly, of what remains to be achieved.
We have already noted the need for transparency (or at least defensibility) of search processes from the patent search community. We hope this book will allow the IR researchers to better understand why such transparency is needed, and what it means in practise. Furthermore, it is our hope that this book will also be a valuable resource for IP professionals in learning about current approaches of IR in the patent domain. It has often been difficult to reconcile the focus on useful technological innovation from the IP community, with the demands for scientific rigour and to proceed on the basis of sound empirical evidence, which is such an important feature of IR (in contrast to some other areas of computer science).
Moreover, patent search is an inherently multilingual and multinational topic: the novelty of a patent may be dismissed by finding a document describing the same idea in any language anywhere in the world. Patents are complex legal documents, even less accessible than the scientific literature. These are just some of the characteristics of the patent system, which make it an important challenge for the search, information retrieval and information access communities.
The book has had a lengthy and difficult gestation: the list of authors has been revised many times as a result of changes in institutional, occupational and private circumstances. Although we, the editors, do feel we have succeeded in producing a volume which will provide important perspectives of the issues affecting patent search research and innovation at the time of writing, as well as a useful, brief introduction to the outlook and literature of the community accessible to its members, regardless of their background, we would have liked to cover several topics not represented here.
In particular it was disappointing we could not include a chapter on NTCIR, the first of the evaluation campaigns to focus seriously on patents. Also, a chapter on the use of Latent Semantic Indexing for the patent domain had been planned, which ultimately could not appear in this book.Introduction to Patent Searching
Introduction to Patent Searching
An Introduction to Contemporary Search Technology
Evaluating Patent Retrieval
Overview of Information Retrieval Evaluation
Evaluating Information Retrieval in the Intellectual Property Domain: The CLEF–IP Campaign
Evaluation of Chemical Information Retrieval Tools
Evaluating Real Patent Retrieval Effectiveness
High Recall Search
Measuring and Improving Access to the Corpus
Measuring Effectiveness in the TREC Legal Track
Large-Scale Logical Retrieval: Technology for Semantic Modelling of Patent Search
Patent Claim Decomposition for Improved Information Extraction
From Static Textual Display of Patents to Graphical Interactions
Classification
Automated Patent Classification
Phrase-based Document Categorization
Using Classification Code Hierarchies for Patent Prior Art
Semantic Search
Information Extraction and Semantic Annotation for Multi-Paradigm Information Management
Intelligent Information Access from Scientific Papers
Representation and Searching of Chemical-Structure Information in Patents
Offering New Insights by Harmonizing Patents, Taxonomies and Linked Data
Automatic Translation of Scholarly Terms into Patent Terms
Future Patent Search