Patent Dataset Research Articles

Spatial Interactions and the Commercialisation of Academic Patents: The Chinese Experience

Studies on academic patent commercialisation are well documented. However, spatial effects are seldom considered, which could lead to potentially misleading analytical results. This study addresses this concern by applying the spatial analysis method to investigate how university-level factors and commercialising academic patents are related. Using a comprehensive dataset of university patents from 1815 Chinese universities in 2016, we find that public research funding of universities, industrial research funding, the number of scientific research personnel, the number of monographs published, and the number of awarded achievements are positively related to the commercialisation of academic patents. However, the number of teaching and research personnel and the number of academic papers published are negatively related. There are positive spatial spillover effects in commercialising academic patents among neighbouring universities, but there will be negative spatial spillover effects in funding competition.

Cellulose-Based Hydrogels: Patent Analysis

Background: Among biopolymers, cellulose and its derivatives are the most commonly used for hydrogel formulations. The innovation and improvement of cellulose-based hydrogels concerned the raw materials, synthesis and methods of preparation, formulations and fabrication processes, as well as applications.
 Objectives: This study, in the form of patent analysis, presents the state by introducing what has been innovated and patented concerning cellulose-based hydrogels.
 Methods: Three databases have been used in this study: the Patentscope, the Espacenet, and the Lens patent data set. A detailed analysis has been provided regarding publication dates, patent families, jurisdictions, inventors, applicants, owners, and patent classifications.
 Results: A total of 8053 patent documents related to cellulose-based hydrogels have been published between 1965 and 2021. The United States leads the patent race in this sector, and the Massachusetts Institute of Technology is one of the top academic applicants.
 Conclusion: Based on patent classifications, most patent documents are related to medicinal preparations characterized by special physical forms. More specifically, the classification concerns materials for prostheses or coating prostheses, including cellulose derivatives characterized by their function or physical properties, such as macromolecular gels, hydrogels, or hydrocolloids.

Artificial Intelligence Innovation and Privacy Law: The Moderating Role of Culture

With the advent of artificial intelligence (AI) centred around a logic of commercial surveillance and data capitalism, scholars are increasingly being called upon to consider the implications of privacy law. We respond by exploring the relationship between privacy law and AI innovation and how that relationship between privacy law and AI innovation varies across countries with different culture. Employing a unique dataset of AI patents, we perform a difference-in-differences analysis by exploiting the introduction of General Data Protection Regulation (GDPR). Our results show that compared with the GDPR-unaffected countries, the GDPR-affected countries exhibited a lower AI innovation rate after the GDPR introduction. Further, we find this negative effect is weaker in GDPR-affected countries that have higher levels of individualism, masculinity, and power distance, but stronger with higher levels of uncertainty avoidance. The main contribution of our study is introducing the role of privacy law and its contingencies vis à vis national culture that constitute drivers of AI innovation.

How public investment fuels innovation: Clues from government-subsidized USPTO patents

• This study is among the first attempts to depict a full picture of the public-subsidized innovation scheme. • This study proposes a multi-dimension framework for learning public-subsidized innovation schemes—the innovation promotor, innovator, and innovation area. • Government agencies are more and more involved in collaborative relationships. • Non-government innovators have shown more powerful vitality than government innovators. • Federal agencies vary in preferences to select subsidized recipients and technological fields. Considerably little effort has been devoted to systematically exploring the landscape of government-funded innovation schemes, leaving policymakers and academics little empirical evidence with which to grasp the overall situation of national R&D funding strategies. In the United States (US), patents with a government right or financial interest are flagged, offering a window through which we can examine the landscape of today's patterns of government support for innovation. We conduct the examination from the perspectives of the leading government agencies, the promotor-recipient relationship, and the common concerns of different participants. The results show that the R&D project funding by federal agencies has, among other factors, contributed to the growth in the number of patent records. The Departments of Defense (DoD), Energy (DoE), and Health and Human Services (HHS), and the National Science Foundation (NSF) have issued the most funding and hold rights to more patent records than any other agencies. Moreover, reforms to the US patent system made 40 years ago to afford more protection to non-government assignees appear to have significantly impacted patenting activity. Universities, companies, and individuals have become far more active in fulfilling government-funded projects, with remarkable innovations as a result. In recent years, funding priorities have shifted toward research in biomedicine, fields related to human life, and computer science. However, the four dominant agencies responsible for most government funding differ in their investment priorities and frequently adjust which areas of innovation they support. This study is among the first attempts to support the landscape of the public-invested innovation using the full government-subsidized patent dataset. The analyses in this paper are instructive in understanding the dynamics of US government funding for innovation and informing effective policymaking around innovation.

A semantic main path analysis method to identify multiple developmental trajectories

• A semantic main path analysis approach is put forward to identify simultaneously multiple developmental trajectories in a target field. • To improve topical coherence of documents along a same trajectory, the conventional link weights are armed with the semantic information based ones. • After all paths are enumerated effectively with a dynamic programming based search algorithm, a density-based clustering method is used to divide them into several groups. • The source codes in Python language can be freely accessed at the GitHub and PyPi with detailed API documentation, thus to promote the related studies. Main Path Analysis (MPA) is widely used to trace the developmental trajectory of a technological field through a citation network. The citation-based traversal weight is usually utilized to cherry-pick the most significant path. However, the theme of documents along a main path may not be so coherent, and it is very possible to miss the main paths of significant sub-fields overall in a domain. Furthermore, the global path search algorithm in conventional MPA also suffers from high space complexity due to the exhaustive strategy. To address these limitations, a new method, named as semantic MPA (sMPA), is proposed by leveraging semantic information in two steps of candidate path generation and main path selection. In the meanwhile, the resulting source code can be freely accessed. To demonstrate the advantages of our method, extensive experiments are conducted on a patent dataset pertaining to lithium-ion battery in electric vehicle. Experimental results show that our sMPA is capable of discovering more knowledge flows from important sub-fields, and improving the topical coherence of candidate paths as well.

Talents and Cultures: Immigrant Inventors and Ethnic Diversity in the Age of Mass Migration

Abstract We investigate the importance of co-ethnic networks and diversity in determining immigrant inventors’ settlements in the United States by following the location choices of thousands of them across counties during the Age of Mass Migration. To do so, we combine a unique United States Patent and Trademark Office historical patent dataset on immigrants who arrived as adults with Census data and exploit exogenous variation in both immigration flows and diversity induced by former settlements, WWI, and the 1920s Immigration Acts. We find that co-ethnic networks play an important role in attracting immigrant inventors. Yet, we also find that immigrant diversity acts as an additional significant pull factor. This is mainly due to externalities that foster immigrant inventors’ productivity.

Where do capabilities reside? Analysis of related technological diversification in multi-locational firms

ABSTRACT Where do businesses source the capabilities that influence their diversification? From the resource-based perspective, capabilities are located within the firm. For many economic geographers, capabilities are place-based and flow between firms in local areas. Others claim the capabilities that count emerge from non-local collaboration. The value added of this paper is a firm–establishment–patent dataset that helps identify the sources of capabilities regulating technological diversification in the establishments of multi-locational firms. Results show that capabilities located within establishments themselves are most important to the process of diversification, followed by firm capabilities and then place-based capabilities.

Spatial Pattern and Evolution of Global Innovation Network from 2000 to 2019: Global Patent Dataset Perspective

In the era of the knowledge economy, the improvement of national innovation systems is playing a significant role in the global entrepreneurship ecosystem. Entrepreneurs are accelerating international intellectual property applications to be competitive. What remains to be explored is the evolution of international intellectual property network in the globe. With the application of social network analysis and intellectual property application database, the global innovation network structure from 2000 to 2019 is explored. Results showed that (1) in the period 2000–2019, the global innovation network has been expanding rapidly from a sparse network to a dense and complex one. (2) Patent application is unevenly distributed in the globe. Countries such as the US, China, and Canada have been the top countries flowing in, while Japan, Korea, EU, and Switzerland have been the main countries flowing out. (3) Global innovation network shows an obvious “core‐periphery” pattern. The distribution pattern presents a quadrilateral structure with the four core regions of “US, Japan, EU, and China” as the apex. This analysis contributes to the visualization of the global layout of intellectual property and the evolution trend by analyzing intellectual property application networks. This can provide important experience reference for enterprises to study the global entrepreneurship ecosystem.

Generating tacit knowledge across borders: international collaboration and design innovation in Norway

ABSTRACT We analyse the international dispersion of inventor networks engaged in design innovation, by studying a dataset of design patents linked to Norway. Specifically, we examine the extent to which the international dispersion of design innovation is sustained by firms´ organisation-based pipelines, innovation capabilities and star designers’ personal linkages. We find that, in traditional design sectors, greater orchestration and innovative capabilities help firms source knowledge globally, while star designers tend to work independently or collaborate locally. Design innovation activities in high-tech sectors show higher international connectedness, largely driven by star designers’ personal connections. Surprisingly, firms with lower innovative capability tend to be connected to inventor networks that are more internationally dispersed. While increasing attention is being given to the value added by design, little is known about international design innovation networks. Our paper opens potential avenues for research on the distribution and orchestration of design innovation across borders.

Identifying artificial intelligence (AI) invention: a novel AI patent dataset

Identifying artificial intelligence (AI) invention: a novel AI patent dataset

Patent Value Analysis Using Deep Learning Models—The Case of IoT Technology Mining for the Manufacturing Industry

The R&D output and global commercialization of intellectual properties (IPs), especially patents filed in many countries, have increased dramatically over the past decade. The overwhelming growth in research and IP activities has led to a major challenge to understand and forecast technology development insights and trends. Evidence-based data analytics is essential for technology mining. The assessment of patent values is a critical aspect of technology mining, which remains a highly subjective task performed by domain experts. As businesses become globalized, subjectivity in underlying assessments of large volumes of patent documents leads to overpriced or undervalued IP sales or licensing that exposes stakeholders to legal and financial risks. Thus, the development of intelligent methods for patent valuation requires new research emphasis. This article applies a deep learning analytical method for automatic and intelligent patent value estimation. Principal component analysis (PCA) is used to identify significant patent value indicators from the given patent dataset. Then, deep neural networks (DNN) for value prediction are modeled and trained using the training set. A detailed case study of 6466 manufacturing Internet of Things (IoT) patents is analyzed to demonstrate the improved results of building PCA-preprocessed DNN models to perform patent valuations. Finally, selected higher value IoT patents owned by leading Taiwan assignees are identified and analyzed to verify the technological competitive intelligence.

When Cyclodextrins Met Data Science: Unveiling Their Pharmaceutical Applications through Network Science and Text-Mining.

We present a data-driven approach to unveil the pharmaceutical technologies of cyclodextrins (CDs) by analyzing a dataset of CD pharmaceutical patents. First, we implemented network science techniques to represent CD patents as a single structure and provide a framework for unsupervised detection of keywords in the patent dataset. Guided by those keywords, we further mined the dataset to examine the patenting trends according to CD-based dosage forms. CD patents formed complex networks, evidencing the supremacy of CDs for solubility enhancement and how this has triggered cutting-edge applications based on or beyond the solubility improvement. The networks exposed the significance of CDs to formulate aqueous solutions, tablets, and powders. Additionally, they highlighted the role of CDs in formulations of anti-inflammatory drugs, cancer therapies, and antiviral strategies. Text-mining showed that the trends in CDs for aqueous solutions, tablets, and powders are going upward. Gels seem to be promising, while patches and fibers are emerging. Cyclodextrins’ potential in suspensions and emulsions is yet to be recognized and can become an opportunity area. This is the first unsupervised/supervised data-mining approach aimed at depicting a landscape of CDs to identify trending and emerging technologies and uncover opportunity areas in CD pharmaceutical research.

Deep learning based pipeline with multichannel inputs for patent classification

This work introduces a deep learning pipeline for automatic patent classification with multichannel inputs based on LSTM and word vector embeddings. Sophisticated text mining methods are used to extract the most important segments from patent texts, and a domain-specific pre-trained word embeddings model for the patent domain is developed; it was trained on a very large dataset of more than five million patents. The deep learning pipeline is using multiple parallel LSTM networks that read the source patent document using different input dimensions namely embeddings of different segments of patent texts, and sparse linear input of different metadata. Classifying patents into corresponding technical fields is selected as a use case. In this use case, a series of patent classification experiments are conducted on different patent datasets, and the experimental results indicate that using the segments of patent texts as well as the metadata as multichannel inputs for a deep neural network model, achieves better performance than one input channel.

Local Technology Adoption and Innovation: The Establishment of U.S. Airmail and the Organization of Aviation Innovation

This paper explores how technology adoption can shape local innovative activity in the context of the early 20th century rollout of airmail in the United States. Our theoretical framework highlights the mechanism of technological feedback and market incentives through which technology adoption induces local innovation, as well as the role of regional domain-specific capabilities in this process. Using our novel dataset of early 20th-century patents, we quantitatively examine these issues by measuring the effect of local airmail upon aviation innovation. Our results suggest that the establishment of an airmail route in a given county led to an increase in local aviation patenting by approximately 32%. Both individual and corporate inventors in treated counties had higher rates of aviation patenting in broad and diverse areas of aviation technology. We also find that the effect of airmail adoption was stronger in the treated counties with pre-existing local capabilities that specifically catered to, or could be readily applied to, the focal technological domain. This work contributes to the academic literature on antecedents and consequences of technology adoption and innovation. Our findings are also relevant for present-day managers and policymakers who are navigating the potential costs and benefits of new technology adoption.

Sequential patent trading recommendation using knowledge-aware attentional bidirectional long short-term memory network (KBiLSTM)

With the rapid development of the patent marketplace, patent trading recommendation is required to mitigate the technology searching cost of patent buyers. Current research focuses on the recommendation based on existing patents of a company; a few studies take into account the sequential pattern of patent acquisition activities and the possible diversity of a company’s business interests. Moreover, the profiling of patents based on solely patent documents fails to capture the high-order information of patents. To bridge the gap, we propose a knowledge-aware attentional bidirectional long short-term memory network (KBiLSTM) method for patent trading recommendation. KBiLSTM uses knowledge graph embeddings to profile patents with rich patent information. It introduces bidirectional long short-term memory network (BiLSTM) to capture the sequential pattern in a company’s historical records. In addition, to address a company’s diverse technology interests, we design an attention mechanism to aggregate the company’s historical patents given a candidate patent. Experimental results on the United States Patent and Trademark Office (USPTO) data set show that KBiLSTM outperforms state-of-the-art baselines for patent trading recommendation in terms of F1 and normalised discounted cumulative gain (nDCG). The attention visualisation of randomly selected company intuitively demonstrates the recommendation effectiveness.

A Detailed Forecast of the Technologies Based on Lifecycle Analysis of GMAW and CMT Welding Processes

In this study, GMAW and CMT welding technologies were evaluated in terms of their technological lifecycles based on their patent datasets together with the S-curve concept, and the joints were evaluated in terms of their welding characteristics. To predict the future trends for both technologies, different models based on the time-series and growth-curve methods were tested. From a process point of view, the results showed better performance and stability for the CMT process based on the heat input to the base material and the frequency of the short circuits. The temperature distribution in the sample revealed that the GMAW process delivers higher values and, consequently, greater heat transfer. Regarding the technological lifecycle, the analyses revealed that the CMT welding process, despite being recent, is already in its mature phase. Moreover, the GMAW welding process is positioned in the growth phase on the S-curve, indicating a possibility of advancement. The main findings indicated that through mathematical modelling, it is possible to predict, in a precise way, the inflection points and the maturity phases of each technology and chart their trends with expert opinions. The new perspectives for analysing maturity levels and welding characteristics presented herein will be essential for a broaden decision-making market process.

Similarity-based approach for inventive design solutions assistance

With the increasing demand for inventive products, finding out inventive design solutions hidden in different industrial engineering domains has always been a challenge for engineers. In addition, patent documents are full of the latest inventive knowledge inside. In this paper, we rely on the assumption that an engineering problem may have an inventive practical solution in another scientific domain as long as they are similarly described. Therefore, we focus on applying machine learning techniques, more particularly neural networks to determine the similarity between patent problems. Technically, a trained bidirectional LSTM neural network, called Manhattan LSTM is integrated into our approach named SAM-IDM to predict the similarity between sentences. We experimentally show that Manhattan LSTM outperforms other baseline approaches in a labelled sample dataset of SNLI corpus. We then experiment our approach on a real-world U.S. patent dataset and we demonstrate that it presents promising results in terms of sentence similarity matching and inventiveness. An inventive design case is detailed to illustrate its performance and practicality.

Patent landscape of not-in-kind active cooling technologies between 1998 and 2017

The rise in global cooling demand will have significant impacts on our efforts to limit climate change and foster sustainable development. Minimising these impacts requires further development of active cooling technologies, starting with an examination of existing technologies and their characteristics. This paper provides an overview of active cooling technologies using a patent landscape based on patent application data over the past 20 years. In developing the patent landscape, we searched the Derwent Innovation patent database, which contains more than 90 million patent documents, using technology-specific keywords and patent classification codes. A careful assessment of the resulting patent dataset shows a range of important trends about the evolution of cooling in the last two decades. Specifically, we discuss annual cooling technology patent application trends, countries of application, and patent assignees. The results show that patenting activity in the field of active cooling has increased significantly since 2010, mainly driven by the rapid patenting trend in China. Patents of the incumbent technology have been dominated by established manufacturers from Japan and the United States. Not-in-kind cooling technologies with a high number of patent applications include thermoelectric, magnetocaloric, evaporative, pulse-tube, absorption, and adsorption cooling. The results from examining this global patent help in understanding and shaping the global development of active cooling technology by providing new information on the major supplier countries, the key manufacturers, and maturity levels of alternative cooling technologies.

Historical patent data: A practitioner's guide

AbstractI provide a primer on six recent large‐scale historical patent data sets for use in innovation research. I discuss how each data set is constructed, the types of patent information included in each, and the quality and completeness of each. Throughout, I emphasize when our knowledge of the history of invention is dependent on the data source used and provide recommendations about which data set is most likely to be best for different contexts. Overall, these data sets paint a remarkably consistent picture of the history of U.S. invention. When the data sets do disagree, these differences tend to be minor, although I highlight some important exceptions. I further describe several “niche” historical patent data sets that allow researchers to study institutional contexts that cannot be studied using modern data. Finally, I discuss features of patent data that are not available for the historical patents but are available for modern patents.

Local Multipliers in New Firm Creation: Inter-Sectoral Spillovers of Entrepreneurship and Startup Agglomeration

ABSTRACT This paper investigates the local multiplier effects in new firm creation by analysing a unique panel dataset of new firms and patents at the regional level in Australia. It finds that new firms in the tradable sector such as low-technology manufacturing and knowledge-intensive services industries have local multiplier effects on new firm creation in the non-tradable sector such as other services industries. It also shows that the local multiplier effect in new firm creation exists between manufacturing and knowledge-intensive services industries in the tradable sector. The local multipliers are greater in regions with more patents than in other regions.