A text-embedding-based approach to measuring patent-to-patent technological similarity

Hain, Daniel S.; Jurowetzki, Roman; Buchmann, Tobias; Wolf, Patrick

doi:10.1016/j.techfore.2022.121559

Cited by 46 publications

(23 citation statements)

References 90 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The use of embeddings allows the representations of text while preserving its original meaning 18 . The algorithm is comparable to those in Whalen et al (2020) and Hain et al (2022), but we additionally exploit document‐level interdependencies with Scientific Paper Embeddings using Citation‐informed TransformERs (SPECTER) model weights. Although embeddings have proven to be very powerful in measuring semantic similarity, they may fail to account for certain peculiarities (in patents and standard documents) that do not exist in the training text corpus (i.e., scientific publications).…”

Section: Methodsmentioning

confidence: 99%

Truly standard‐essential patents? A semantics‐based analysis

Brachtendorf

Gaessler

Harhoff

2022

Economics Manag Strategy

View full text Add to dashboard Cite

Standard-essential patents (SEPs) have become a key element of technical coordination in standardsetting organizations. Yet, in many cases, it remains unclear whether a declared SEP is truly standardessential. To date, there is no automated procedure that allows for a scalable and objective assessment of SEP status. This paper introduces a semantics-based method for approximating the standard essentiality of patents. We provide details on the procedure that generates the measure of standard essentiality and present the results of several validation exercises. In a first empirical application we illustrate the measure's usefulness in estimating the share of true SEPs in firm patent portfolios for several mobile telecommunication standards. We find firm-level differences that are statistically significant and economically substantial. Furthermore, we observe a general decline in the average share of presumably true SEPs between successive standard generations.

show abstract

Section: Methodsmentioning

confidence: 99%

Truly standard‐essential patents? A semantics‐based analysis

Brachtendorf

Gaessler

Harhoff

2022

Economics Manag Strategy

View full text Add to dashboard Cite

show abstract

“…For instance, when the research object is shifted to the textual information of the patent, for example, patent abstract, the model can also be extended to measure patent similarity. Although previous research has widely applied text‐matching techniques in calculating patent‐to‐patent similarities (Arts et al, 2017; Hain et al, 2022), the semantic‐enhanced approaches have been neglected. In the future, we will conduct further experiments to validate whether the semantic‐enhanced model proposed in this paper can achieve better performance than the traditional text representation approaches.…”

Section: Methodsmentioning

confidence: 99%

A lightweight semantic‐enhanced interactive network for efficient short‐text matching

Xue

Luo

et al. 2022

Asso for Info Science & Tech

View full text Add to dashboard Cite

Knowledge‐enhanced short‐text matching has been a significant task attracting much attention in recent years. However, the existing approaches cannot effectively balance effect and efficiency. Effective models usually consist of complex network structures leading to slow inference speed and the difficulties of applications in actual practice. In addition, most knowledge‐enhanced models try to link the mentions in the text to the entities of the knowledge graphs—the difficulties of entity linking decrease the generalizability among different datasets. To address these problems, we propose a lightweight Semantic‐Enhanced Interactive Network (SEIN) model for efficient short‐text matching. Unlike most current research, SEIN employs an unsupervised method to select WordNet's most appropriate paraphrase description as the external semantic knowledge. It focuses on integrating semantic information and interactive information of text while simplifying the structure of other modules. We conduct intensive experiments on four real‐world datasets, that is, Quora, Twitter‐URL, SciTail, and SICK‐E. Compared with state‐of‐the‐art methods, SEIN achieves the best performance on most datasets. The experimental results proved that introducing external knowledge could effectively improve the performance of the short‐text matching models. The research sheds light on the role of lightweight models in leveraging external knowledge to improve the effect of short‐text matching.

show abstract

“…Meanwhile, the network homogeneity of LDGIN is higher than that of RDGIN. The homogeneity of a DGI network can promote the absorption of digital green knowledge and the improvement of the DGI performance of manufacturing enterprises [46].…”

Section: Ldgin and Dgi Of Manufacturing Enterprisesmentioning

confidence: 99%

Insight into the Balancing Effect of a Digital Green Innovation (DGI) Network to Improve the Performance of DGI for Industry 5.0: Roles of Digital Empowerment and Green Organization Flexibility

Yang

Yin

2022

Systems

View full text Add to dashboard Cite

Under the background of double carbon target and digital intelligence era, the innovation resources and innovation environment that manufacturing enterprises rely on have shown exponential growth. Digital green innovation (DGI) has gradually become the mainstream paradigm of innovation. How to achieve a balance between a local DGI network (LDGIN) and a remote DGI network (RDGIN) and how to use the role of digital empowerment and green organization flexibility to improve the performance of DGI are very important issues facing manufacturing enterprises at present. However, this problem has not been fully addressed in the existing research. In this study, the influence mechanism of LDGIN and RDGIN on the DGI performance of manufacturing enterprises was revealed, considering the moderating role of digital empowerment and green organization flexibility. The linear regression method was used to analyze the 562 valid data obtained by questionnaire survey. The results of this study are as follows. The effect of the DGI network on manufacturing enterprises’ DGI performance is heterogeneous because of LDGIN and RDGIN. The establishment of embedded links in a DGI network inevitably requires manufacturing enterprises to pay the corresponding costs. The over-embedding of manufacturing enterprises into RDGIN will have a negative impact on DGI performance. The balance between LDGIN and RDGIN has an important impact on manufacturing enterprises’ operation. The comprehensive balance and relative balance indexes constructed in this paper show that an appropriate balance can promote the improvement of the DGI performance of manufacturing enterprises. Digital transformation and organizational structure innovation are changing the business model of manufacturing enterprises and can regulate the relationship between the LDGIN and RDGIN and the DGI performance of manufacturing enterprises. The balance of DGI network embedding in practice shows the important role and enlightening significance of local and remote search in developing countries.

show abstract

A text-embedding-based approach to measuring patent-to-patent technological similarity

Cited by 46 publications

References 90 publications

Truly standard‐essential patents? A semantics‐based analysis

Truly standard‐essential patents? A semantics‐based analysis

A lightweight semantic‐enhanced interactive network for efficient short‐text matching

Insight into the Balancing Effect of a Digital Green Innovation (DGI) Network to Improve the Performance of DGI for Industry 5.0: Roles of Digital Empowerment and Green Organization Flexibility

Contact Info

Product

Resources

About