Welcome!

Machine Learning Authors: Pat Romanski, Yeshim Deniz, Liz McMillan, Elizabeth White, Zakia Bouachraoui

Related Topics: @DXWorldExpo, Cognitive Computing , Machine Learning

@DXWorldExpo: Article

Patent Data Quality | @CloudExpo #BigData #Analytics #AI #MachineLearning

Is clean data a pipe dream?

The United States Patent and Trademark Office (USPTO) recently announced an expansion of PatentsView, its visualization tool for US patents. First launched a few years ago, the intent behind the tool was to make 40 years of patent filing data available for free to those interested in examining "the dynamics of inventor patenting activity over time." In spite of being limited to patents (not applications) and with a focus only on the US, it offers some interesting visualizations around locations and citations.

In a blog post last month, USPTO director Michelle Lee said the PatentView tool is based on "the highest-quality patent data available," connecting 40 years' worth of information about inventors, their organizations, and their locations in unprecedented ways. The newly revamped interface presents three user-friendly starting points - relationship, locations, and comparison visualizations - which allow for deeper exploration and detailed views. However, through no fault of their own, the USPTO dataset is rife with spelling errors, doesn't reflect patent reassignments, and doesn't resolve company subsidiaries or acquisitions.

This issue is not unique to the USPTO. Other PTO offices around the world face similar barriers to presenting "clean" data. The first issue, spelling errors, merely reflects the fact that assignee information (among other fields like inventor names) is manually entered and hence prone to error and inconsistency. For example, "International Business Machines" has been spelled 1,200 different ways as a patent assignee over the last two decades in the USPTO data set.

In addition, PTO data doesn't get corrected or updated based on later corrections or patent reassignments. For example, patent US8176440 was originally - and incorrectly - assigned to Silicon Labs. My company, Innography, filed a certificate of correction to update the assignment, yet the USPTO data and PatentsView still don't reflect this. In fact, Innography research shows that nearly 20 percent of US patents are reassigned in their lifetimes, translating into a significant number of company portfolio errors based on this factor alone.

Finally, PTO data also doesn't reflect when companies purchase each other, when there's a spinoff, or when a subsidiary files patents. Microsoft, for example, now owns all LinkedIn's patents, even if the reassignments haven't been processed.

As a result, PTO data falls far short of reflecting reality, where patents and companies are bought and sold every day, and where data-entry errors exist and are corrected. The accuracy of the data is very low when it comes to representing company patent portfolios in the real world.

The Cost of Free Data
The USPTO aims to increase the transparency of patenting and invention processes. But if the quality of data and search results is questionable, what good is it to IP practitioners?

There is rich information available through the patenting process, including economic research, prior-art searching, and discovery of broader trends around filing patterns. However, it was never intended to be used as-is to inform strategic business decisions such as in and out licensing, merger and acquisition activities, or portfolio pruning and maintenance decisions.

It makes sense for PTOs to offer their data for free as a way to engage the community's interest in patenting processes. However, too many lightweight patent analytics tools use this flawed data verbatim to tout their "data quality" to IP professionals.

Many patent analyses start with a company's patent portfolio, such as competitive benchmarking, acquisition analysis, and negotiation preparation. In addition, just about every board-level question about patents requires accurate patent ownership information: "Are we ahead of or behind this competitor?" "What companies should we be worried about in this technology area?"

Poor data quality makes it difficult, if not impossible, to answer those questions accurately. To create the most accurate data set possible, companies must use other sources of information to crosscheck and improve patent data accuracy.

Innography data scientists process more than 2,000 company acquisitions annually, and our user base suggests another 5,000 updates each year. As a result, Innography has created more than 10 million data-correction rules over the last decade, which are continuously updated via machine learning and crowdsourcing.

Company leaders must be able to use patent reports to assess market opportunities and make strategic business decisions. This requires an IP analytics solution that reflects real-world changes, and doesn't rely on poor data quality from outdated PTO assignee information.

More Stories By Tyron Stading

Tyron Stading is president and founder of Innography, and chief data officer for CPA Global. He has been named one of the “World’s Leading IP Strategists" by IAM, and one of National Law Journal's "50 Intellectual Property Trailblazers & Pioneers". Before Innography, Tyron was an IBM worldwide industry solutions manager in the telecommunications and utilities sector, and worked at several start-ups focused on mobile communications and networks security. He has published multiple research papers and filed more than three dozen patents. Tyron has a BS in Computer Science from Stanford University and an MS in Technology Commercialization from The University of Texas.

CloudEXPO Stories
With more than 30 Kubernetes solutions in the marketplace, it's tempting to think Kubernetes and the vendor ecosystem has solved the problem of operationalizing containers at scale or of automatically managing the elasticity of the underlying infrastructure that these solutions need to be truly scalable. Far from it. There are at least six major pain points that companies experience when they try to deploy and run Kubernetes in their complex environments. In this presentation, the speaker will detail these pain points and explain how cloud can address them.
The deluge of IoT sensor data collected from connected devices and the powerful AI required to make that data actionable are giving rise to a hybrid ecosystem in which cloud, on-prem and edge processes become interweaved. Attendees will learn how emerging composable infrastructure solutions deliver the adaptive architecture needed to manage this new data reality. Machine learning algorithms can better anticipate data storms and automate resources to support surges, including fully scalable GPU-centric compute for the most data-intensive applications. Hyperconverged systems already in place can be revitalized with vendor-agnostic, PCIe-deployed, disaggregated approach to composable, maximizing the value of previous investments.
When building large, cloud-based applications that operate at a high scale, it's important to maintain a high availability and resilience to failures. In order to do that, you must be tolerant of failures, even in light of failures in other areas of your application. "Fly two mistakes high" is an old adage in the radio control airplane hobby. It means, fly high enough so that if you make a mistake, you can continue flying with room to still make mistakes. In his session at 18th Cloud Expo, Lee Atchison, Principal Cloud Architect and Advocate at New Relic, discussed how this same philosophy can be applied to highly scaled applications, and can dramatically increase your resilience to failure.
Machine learning has taken residence at our cities' cores and now we can finally have "smart cities." Cities are a collection of buildings made to provide the structure and safety necessary for people to function, create and survive. Buildings are a pool of ever-changing performance data from large automated systems such as heating and cooling to the people that live and work within them. Through machine learning, buildings can optimize performance, reduce costs, and improve occupant comfort by sharing information within the building and with outside city infrastructure via real time shared cloud capabilities.
As Cybric's Chief Technology Officer, Mike D. Kail is responsible for the strategic vision and technical direction of the platform. Prior to founding Cybric, Mike was Yahoo's CIO and SVP of Infrastructure, where he led the IT and Data Center functions for the company. He has more than 24 years of IT Operations experience with a focus on highly-scalable architectures.