18 February 2020
Wolf, Greenfield & Sacks, P.C.
To print this article, all you need is to be registered or login on Mondaq.com.
Today, companies are developing artificial intelligence (AI)systems to meaningfully analyze the deluge of biomedical data. Asubstantial investment in building and deploying machine learning(ML) technologythe most active area of AI technology beingdeveloped todaywarrants carefully considering how to protectthe resulting intellectual property (IP), but there are challengesto doing so. In this article, we explore strategies of protectingIP for ML technology, including what aspects to consider patentinggiven current and ongoing changes to U.S. patent law, and when toconsider trade secret protection.
Generally, developing an ML system involves creating anddeploying a computer program having a model whose performance onsome task improves as additional data is used to train the model.In the life sciences, such data can include medical images, genomicdata, and electronic health records.
For example, an ML model may be trained on magnetic resonance(MR) images to recognize whether a previously unseen MR image of apatient's brain shows a hemorrhage. As another example, an MLmodel may be trained on genomic data for individuals with aparticular cancer to predict whether a patient's genome hasfeatures indicative of the cancer.
Today, neural networks are a popular class of ML models widelyused, and are often referred to as "deep learning" in anod to their multi-layer (deep) structure. Other ML models includeBayesian models, decision trees, random forests, and graphicalmodels. Indeed, rapid development of various ML tools has led to anexplosion of activity in applying them to new problems acrossdiverse fields.
Deploying an ML system typically involves: (1) selecting/designing an ML model, (2) training the ML model using data, and(3) deploying and using the trained ML model in an application.Valuable IP may be generated at each of these stages, and it'sworth considering protecting it through patents. There, however,are a number of challenges in patenting ML systems.
An invention must be new and non-obvious to be patented. Thismakes it difficult to patent the use of off-the-shelf ML technologyeven if in the context of a new application. Simply downloadingfreely available ML software, providing it with data, anddisplaying the results (e.g., to a doctor or researcher) may beviewed by the U.S. Patent and Trademark Office (USPTO) as failingto clear the non-obviousness hurdle. After all, the freelyavailable ML software is distributed precisely so that people canperform this exact processwhy, then, would it not be obviousto do so?
But in reality, building and deploying ML systems requires morework beyond simply downloading and running software. Focusingpatent claims on the results of such efforts will lead to greatersuccess. Here are three examples of potentially patentable aspectsof an ML system:
To see the full article click here
The content of this article is intended to provide a generalguide to the subject matter. Specialist advice should be soughtabout your specific circumstances.
POPULAR ARTICLES ON: Intellectual Property from United States
Global Advertising Lawyers Alliance (GALA)
While my seats afforded me only a so-so sight-line to the stage, I had no trouble seeing the ocean of cell phones, in the hands of adoring fans, simultaneously recording (without authorization)...
Weintraub Tobin Chediak Coleman Grodin Law Corporation
Generally, the title to a single motion picture is not entitled to trademark protection. This is the same for the title to single books, songs and other singular creative works
Cowan Liebowitz & Latman PC
What can you do to protect your goodwill if you unknowingly select an unfortunate brand name, or through no fault of your own...
See more here:
Strategies For Patenting Artificial Intelligence Innovations In The Life Sciences - Mondaq News Alerts