IP Gotchas: Patenting Neural Networks

October 12, 2020

The demand for improvements in autonomous technology is accelerating. Memory and processing power has continued to grow exponentially cheaper, but the volume of data to process has exploded making it nearly impossible for traditional data analysis techniques to provide timely and cost effective guidance. Whether the data arrives in real time, or is acquired and stored for later analysis, the need for systems that can generate valuable insights from mountains of raw information will only increase. Consequently, neural network technology is increasingly valuable to organizations large and small making it a prime target for intellectual property protection. However, a number of misconceptions have arisen regarding the patentability of neural network technology.

Software isn’t patentable, so neural nets aren’t either: Software inventions are patentable, but the Patent Office and the Courts have narrowed the scope of what is patentable by requiring that the patent claims must be directed to something more than a well-known or abstract concept implemented on a computer. This is especially interesting where neural networks are involved because in some cases, the network itself is not new. The network topology (i.e. number of nodes, number of layers, the connections between them, etc.) may not be new, and perhaps the activation functions or backpropagation techniques used by the network are also not new. The training data sets or data preprocessing techniques may not be new either, but the outcome of using such a neural network may be truly revolutionary. Thus it is important to plan ahead in the drafting process to include aspects like tangible data sources and physical sensor input/output, control of physical objects or machines, and information about what technical problems are being overcome and how.

I didn’t invent a neural network so this is probably not patentable: Keeping the right focus on the invention is a fundamental issue that sometimes hampers patentability for software inventions, and it can be particularly problematic where neural networks are involved. Is the “magic” in the neural network or is the magic in how that network is used? For example, is the concept a new topology for a neural net that is more efficient, yields better results, or solves a particular problem? Is it a new activation function, or a new type of backpropagation scheme? Is the invention a new gradient descent algorithm that is optimized for a particular problem space? If any of these are the case, then the claims and disclosure should focus more on the network itself and how it is configured. On the other hand, is the invention a system that works better because it uses a neural network? If the invention is an improvement on neural networks, then more details about the network itself will be needed to show the technical problem and solution. If the invention is an improvement in some other field of endeavor that happens to involve a neural network, then more information about the inputs, outputs, and operation of the device will be needed, and perhaps less information about the neural network itself. Determining what the invention is will drive what kind of disclosure is needed in order to obtain a patent.

I’ll file the application but keep the real invention secret: The patent system grants the right to stop others from making, using, or selling patented inventions. In return, inventors are required to teach the world how to make the invention. In the case of neural networks, this can be tricky because many of the details about how a neural net reaches a given result are unknown until it is put to use, or they may be different from one execution to the next, or in some cases they are simply unknowable without extreme effort. In some cases, billions of permutations of inputs, outputs, and the corresponding weights for each node in the network could exist, but only after the network is put to use. That said, trying to patent a concept while keeping it secret is not permitted. The invention must be disclosed in such a way that a person of ordinary skill in the field could make and use the invention. The claims may be allowed to describe inputs and outputs at a high level, but at least some explanation is required as to how the system uses them and how they interact with other components of the system. Usually, more disclosure is better than less because failure to adequately explain the workings of a neural network, or the use of it, may cost both the opportunity to patent the concept, and the opportunity to protect the invention some other way, such as by trying to keep it a trade secret. With a little careful planning, both of these negative outcomes can usually be avoided.

Debunking the “Poor Man’s Patent Myth”

March 19, 2020

Myth: If I write down my invention and mail it to myself through the USPS, I will receive some level of patent protection on my invention. (In some versions of the myth, I will receive protection only if I do not open the envelope).

This is the Poor Man’s Patent Myth and is truly a myth. US patent protection is granted only to those who receive an issued US patent from the US Patent and Trademark Office (USPTO). Moreover, the US switched to a “first-to-file” system in 2013 under which-in the event two applicants file for patent protection on the same invention-the applicant who files first is given priority. This is why some refer to obtaining patent protection as being, in part, a “race to the patent office”. There are, of course, some narrow exceptions to the “first-to-file” rule (e.g., the earlier-filer derived the invention from the later-filer); however, in any event an application must be filed with the USPTO as a prerequisite to receive patent protection.

Prior to the switch to “first-to-file”, the US was under a “first-to-invent” system. Under the “first-to-invent” system, a later-filing inventor could obtain priority over an earlier-filing inventor who filed for protection on the same invention if the later-filing inventor could show he/she was the first to conceive the invention and meet certain requirements (e.g., diligence to reduction to practice). Today, however, most inventors are working under the “first-to-file” system (i.e., the law under the Leahy-Smith America Invents Act). Therefore, best practice is likely filing a patent application early.

If you have an invention you wish to protect, contact one of the attorneys at Woodard Emhardt. Writing down your invention is a good exercise, but don’t believe the myth that mailing it to yourself gives you patent protection.

This is not legal advice, nor should it be construed as forming an attorney-client relationship. If you wish to have either with our firm, please contact one of our attorneys to begin that process.

USPTO Increasing Patent Fees for 2018

November 16, 2017

The USPTO has issued a final rule, “Setting and Adjusting Patent Fees during Fiscal Year 2017” to set or adjust certain patent fees. Notably, the rule increases a large number of fees including filing fees, search fees, examination fees, and issue fees as well as fees for RCEs, PGRs, IPRs, Appeals, and Petitions. These fee increases will take effect on January 16, 2018.
For the full list of the patent fees that are changing and more information on fee setting and adjusting at the USPTO, please visit

The Story of the Coca-Cola Bottle

November 16, 2015

Today marks the 100th Anniversary of the design patent covering the classic Coca-Cola bottle. Developed in Terre Haute, Indiana, by the Root Glass Company, the classic Coca-Cola bottle was developed to provide a distinctive package for the popular drink. Versions of the original bottle are still available to this day and continue to provide a way to identify the popular Coca-Cola brand. For more information on this piece of intellectual property history, visit: The Coca-Cola Company.

Coca Cola Bottle

July 2015 Patent Case Update

July 31, 2015

The Supreme Court addressed a number of patent law topics this year, including the standard of review for patent claim construction (Teva Pharmaceuticals v. Sandoz), royalties on expired patents (Kimble v. Marvel Enterprises), and defenses to claims of induced infringement (Commil v. Cisco). The initial wave of appeals in post-issuance proceedings provided by the America Invents Act are now being decided by the Federal Circuit. These include cases involving whether the decision to institute is reviewable (In re Cuozzo Speed Technologies) and discussion of the claim construction standards that apply (Microsoft v. Proxyconn). The relationship between administrative challenges and related litigation is also an evolving area (ePlus v. Lawson Software). To view the presentation, visit: Recent Patent Cases from Woodard, Enhardt, Henry, Reeves & Wagner, LLP 

Circuit Split re: Patent Term Adjustment

November 15, 2013

35 U.S.C. § 154(b) allows for extension of patent term for design and utility patents filed after May 29, 2000. The statute provides for patent term adjustment for delays caused by the USPTO. “A Term” adjustments result from untimely responses from the USPTO, e.g. 14 month deadline for a first Office Action, and 4 month deadline for subsequent actions. “B Term” adjustments result from applications pending for more than 3 years. “C Term” adjustments result from delays due to interferences, secrecy orders, and appeals. The statute provides for one day of patent term adjustment for each day of delay caused by the USPTO, any delay caused by the applicant or overlap in terms is subtracted from the final term adjustment.

Recently, a split has arisen within the District Court for the Eastern District of Virginia regarding the calculation of B Term adjustment when a request for continued examination (RCE) has been filed. 35 U.S.C. § 154(b)(1)(B) guarantees no more than 3-year application pendency, providing for patent term adjustment in cases where prosecution extends beyond 3-years. However, the statute includes some limitations, including “any time consumed by continued examination of the application requested by the applicant under section 132(b).” The USPTO has interpreted the statute to require that patent term should not be extended, under B-term, for additional time spent during prosecution after filing of an RCE.

In Excelixis I, Exelixis, Inc. v. Mr. David Kappos , (E.D. Va. November 1, 2012), Judge Ellis found that when an RCE is filed after the 3-year period, the RCE should not have any effect on calculating B-Term adjustment. Following Excelixis I an application will accrue B Term adjustment every day following the 3-year anniversary of filing.

In Excelixis II, Exelixis v. Kappos (E.D. Va. 2013) Judge Brinkema found that “any time consumed by continued examination of the application requested by the applicant under section 132(b) (via the filing of an RCE) does not count toward that three-year period.” Following Excelixis II, an application will accrue B Term only during the period between the 3-year anniversary of filing and the date on which an RCE is filed.

The USPTO has appealed Excelixis I, the case is pending before the Court of Appeals for the Federal Circuit.

Indiana Farmer Loses U.S. Supreme Court Seed Patent Case

May 14, 2013

In a recent case arising out of Indiana, the United States Supreme Court addressed the patent exhaustion doctrine in the context of seed patents. Monsanto sued Indiana farmer Vernon Bowman alleging infringement of its patents covering its Roundup Ready seed. Bowman purchased and planted the Roundup Ready seeds and then saved seeds he recovered from that crop to use in future plantings. As a defense, Bowman argued that his actions were protected under the doctrine of patent exhaustion.

The exhaustion doctrine limits a patentee’s right to control what others can do with an article embodying or containing an invention. Under the doctrine, the initial authorized sale of a patented item terminates all patent rights to that item. By exhausting the patentee’s monopoly in that item, the sale confers on the purchaser, or any subsequent owner, the right to use or sell the item as desired. As the purpose of the patent law is fulfilled when the patentee has received money from the sale, the patent law affords no further basis for restraining the use and enjoyment of the thing sold. However, the doctrine does not extend so far as to allow a purchaser to make copies of the thing sold against the wishes of the patentee.

Despite Bowman’s argument that he used the seeds in a way that is normal for farmers, the Court decided that harvesting seeds from a planted crop was akin to making a copy of the patented seeds. The Court unanimously held that a farmer who buys patented seeds may not reproduce them through planting and harvesting without the patent holder’s permission. As noted by the Court, the applicability of the decision is limited and does not involve every situation with a self-replicating thing.

For more information on potential impacts of this case or patent law in general, feel free to contact the attorneys at Woodard, Emhardt, Moriarty, McNett & Henry LLP.

Tips for Mitigating Claims of Inequitable Conduct

September 21, 2010

Recent decisions by the Federal Circuit have affirmed findings of inequitable conduct based on a failure of the applicant to cite material references that were in the applicant’s possession and not cumulative of other previously-cited references.  The result of the inequitable conduct findings were that the patents-in-suit were rendered unenforceable.  In some instances, the material reference was cited (or created) by the United States Patent and Trademark Office (USPTO), during prosecution of another applications also owned by the patentee and copending with the application that issued as the patent-in-suit.  Additionally, the “other” application was not always a family member (related through common priority) of the application that issued as the patent-in suit.  References created by the USPTO, such as Office Actions (which may include positions opposite to those the applicant is currently taking) and Notices of Allowance (which may pertain to double-patenting issues), have also been considered sufficiently material and noncumulative to support a finding of inequitable conduct.  See McKesson Information Solutions, Inc. v. Bridge Medical, Inc., 47 F3d 897, 82 USPQ2d 1865, Fed Cir. 2007

To mitigate the risk of one of its patents being rendered unenforceable for failing to cite a material reference, businesses (especially those with large patent portfolios) may want to consider establishing procedures to determine whether a reference cited during prosecution in one application should be cited in another application, i.e., whether the reference is material to the claim in another application and noncumulative of previously-cited references.  One manner of doing this is to actively track pending applications and group them based on claimed subject matter.  The grouping should not necessarily be restricted to family members  and should include both U.S. and foreign applications.  When a references (including Office Actions and Notices of Allowance) is cited in one case within a group, consideration can be given to whether that reference should be cited in other cases in that group.  The applicant should also consider how a material reference is being characterized, since taking inconsistent positions with respect to material references whether made during U.S. or foreign prosecution has been used to support a finding of inequitable conduct.  See Therasense, Inc. v. Becton, Dickinson & Co., 593 F3d 1289, 93 USPQ2d 1489, Fed. Cir. 2010