Screening patents of ICT in construction using deep learning and NLP techniques

PurposeThis study proposes an approach to solve the fundamental problem in using query-based methods (i.e. searching engines and patent retrieval tools) to screen patents of information and communication technology in construction (ICTC). The fundamental problem is that ICTC incorporates various techniques and thus cannot be simply represented by man-made queries. To investigate this concern, this study develops a binary classifier by utilizing deep learning and NLP techniques to automatically identify whether a patent is relevant to ICTC, thus accurately screening a corpus of ICTC patents.Design/methodology/approachThis study employs NLP techniques to convert the textual data of patents into numerical vectors. Then, a supervised deep learning model is developed to learn the relations between the input vectors and outputs.FindingsThe validation results indicate that (1) the proposed approach has a better performance in screening ICTC patents than traditional machine learning methods; (2) besides the United States Patent and Trademark Office (USPTO) that provides structured and well-written patents, the approach could also accurately screen patents form Derwent Innovations Index (DIX), in which patents are written in different genres.Practical implicationsThis study contributes a specific collection for ICTC patents, which is not provided by the patent offices.Social implicationsThe proposed approach contributes an alternative manner in gathering a corpus of patents for domains like ICTC that neither exists as a searchable classification in patent offices, nor is accurately represented by man-made queries.Originality/valueA deep learning model with two layers of neurons is developed to learn the non-linear relations between the input features and outputs providing better performance than traditional machine learning models. This study uses advanced NLP techniques lemmatization and part-of-speech POS to process textual data of ICTC patents. This study contributes specific collection for ICTC patents which is not provided by the patent offices.

Patent retrieval architecture based on document retrieval. Sketching out the Spanish patent landscape

A patent is a property granted to any new shape, configuration or arrangement of elements, of any device, tool, instrument, mechanism or other object or part thereof, that allows for a better or different operation, use or manufacture of the object that incorporates it or that provides it with some utility, advantage or technical effect that it did not have before. As a document, a patent really is a title that recognizes the right to exploit the patented invention exclusively, preventing others from making, selling or using it without the consent of the owner. The fact of making a patent is motivated by the fact of promoting creativity, hindering competition in the market as only one person holds the patent, thus protecting the initial investment and fighting against plagiarism. Patents are available to the public for dissemination and general knowledge. It is generally recognized in the specialized literature that patents can be used as an indicator to calculate the results generated by research and development activities, being a very useful indicator to measure various social, economic or technological aspects. For this reason, it is of relevant interest to have tools or systems that allow us to obtain the patents developed in a specific period of time and to carry out analyses of various economic and social factors. These analyses can serve to obtain a social perspective of society’s progress in the technological field, and this is why an analysis of patents is of our interest. This paper proposes a platform specifically designed to obtain knowledge about patents as an indicator of Spanish social, economic or technological aspects. For this purpose, the platform retrieves, analyses and visualizes functionalities that represent data on the landscape of patents obtained from the Spanish Patent and Trademark Office (OEPM) as a particular case of study.

Effect of Technical Domains and Patent Structure on Patent Information Retrieval

Patents are critical intellectual assets for any competitive business. With ever increasing patent filings, effective patent prior art search has become an inevitably important task in patent retrieval which is a subfield of information retrieval (IR). The goal of the prior art search is to find and rank documents related to a query patent. Query formulation is a key step in prior art search in which patent structure is exploited to generate queries using various fields available in patent text. As patent encodes multiple technical domains, this work argues that technical domains and patent structure have their combined effect on the effectiveness of patent retrieval. The study uses international patent classification codes (IPC) to categorize query patents in eight technical domains and also explores eighteen different combination of patent fields to generate search queries. A total of 144 extensive retrieval experiments have been carried out using BM25 ranking algorithm. Retrieval performance is evaluated in terms of recall score of top 1000 records. Empirical results support our assumption. A two-way analysis of variance is also conducted to validate the hypotheses. The findings of this work may be helpful for patent information retrieval professionals to develop domain specific patent retrieval systems exploiting the patent structure.

Engineering Knowledge Graph for Keyword Discovery in Patent Search

Patent retrieval and analytics have become common tasks in engineering design and innovation. Keyword-based search is the most common method and the core of integrative methods for patent retrieval. Often searchers intuitively choose keywords according to their knowledge on the search interest which may limit the coverage of the retrieval. Although one can identify additional keywords via reading patent texts from prior searches to refine the query terms heuristically, the process is tedious, time-consuming, and prone to human errors. In this paper, we propose a method to automate and augment the heuristic and iterative keyword discovery process. Specifically, we train a semantic engineering knowledge graph on the full patent database using natural language processing and semantic analysis, and use it as the basis to retrieve and rank the keywords contained in the retrieved patents. On this basis, searchers do not need to read patent texts but just select among the recommended keywords to expand their queries. The proposed method improves the completeness of the search keyword set and reduces the human effort for the same task.