State-of-the-art patent risk assessment for innovators and patent lawyers.
RISKPAT is a state-of-the-art, user-friendly web application designed to assist innovators and patent lawyers in conducting freedom-to-operate and patent risk assessments in the United States.
Understanding patents at the core of potential litigation is crucial to avoiding infringement. However, this process is complex and challenging due to several factors:
Identify the origin and destination of patent transactions for each applicant, as well as the transacted patents themselves.
Explore →Identify similar patents: For each existing patent, users can identify the most similar patents based on their abstracts using advanced semantic analysis.
Explore →Ideal Abstract Check: Users can write their ideal abstract and directly check for the most similar patents.
Explore →RISKPAT significantly reduces the time required for patent clearance with its two unique features:
Approximately 80% of litigated patents are transacted at least once. With RISKPAT, users can identify the origin and destination of patent transactions for each applicant, as well as the transacted patents themselves. The application generates a comprehensive patent transaction network, detailing the relationships between sellers and buyers. Additionally, RISKPAT provides advanced filtering options based on various criteria, such as:
Understanding patent proximity — how closely patents resemble each other — is vital for detecting and anticipating patent infringement. It helps identify potential infringers early, informs strategic decisions on licensing and litigation, optimizes resource allocation through efficient analysis, and safeguards innovations by navigating patent landscapes effectively. Overall, patent proximity serves as a proactive tool in managing intellectual property risks and supporting competitive advantage in the market.
RISKPAT addresses these challenges with two key functionalities:
Our services leverage advanced AI methods. Specifically, for patent similarity, we have developed proprietary natural language processing (NLP) models. Utilizing Sentence Transformers, a BERT-based model, we employ the Siamese triplet architecture to enhance text similarity calculations. Additionally, we have created a text-based algorithm that combines NLP, network theory, and rule-based techniques. This algorithm harmonizes the name variants of companies registered as patent assignees and predicts patent litigation (infringement).