How Reliable is AI in Patent Analytics?

A Rise in Patent Applications

A recently conducted study found that the average number of patents filed in a year has increased by 1800% since 2012. The speed of innovation has increased by multiple folds, so the researchers and strategists need faster hands or technologies to provide quick yet reliable insights. Millions of patents are getting added every year, and with increasing data, converting them into insights is very difficult and time-consuming. Patent text isn’t standard, and different formats require different approaches. Patents use jargon and abbreviations that common language doesn’t use. Also in patent analytics, for an expert to analyze patents and categorize them into multi-level taxonomies or categories and perform landscape studies for different objectives becomes really difficult as the count of patents becomes more than 2000. It becomes an impossible task in case of broad domain studies when you have more than 5000 patents.

Emerging Problems

Companies have also started to test AI and NLP technologies to cluster patents for quick insights. However, the black-box approach has raised some issues:

1. Experts don’t have any control of the clusters created by the machine. Sometimes the concepts are too random.
2. Reliability and comprehensiveness of concepts.
3. There is no way to see the trends through our own technology lens

Focusing on these challenges, XLSCOUT, an AI-based patent analytics platform has developed explainable AI technologies to extract hidden insights from competitors and technology patents and presents an interactive dashboard by connecting standard and proprietary parameters.

XLSCOUT developed a taxonomy builder that can classify the patents into expert-defined taxonomies with a verified accuracy of 95%. Our patent classification techniques are tested and verified. We have compared our results with ones from manual expert classification.

The deep analytics that XLSCOUT provides are powerful indicators of emerging technologies, problems that companies are trying to solve, competitive trends, and major trends in R&D. It combines multiple preprocessed parameters of the patent with expert-curated taxonomies to create a storyboard and extract hidden insights.

Using XLSCOUT technology, you can find new R&D trends in competitors or technology newcomers faster than current patent analysis techniques. You can also detect emerging market trends by analyzing patents of startups on emerging tech.

Our algorithm has already analyzed over 136 million patents to develop the world’s largest problems dataset and predict specific technologies on which competitors/startups are placing their bets.

About XLSCOUT: XLSCOUT engineered data with best-in-class Artificial Intelligence (AI) technologies to develop the most accurate, comprehensive, and intelligent patent database. Our AI technologies are based on a glass-box approach, thus making it explainable to the users.

We believe that this transparency helps in the adoption of this new age technology by giving more confidence to our users.