Quantum Computing Patents: Mid-Year PTAB-Appeal Decisions on Enablement
We analyzed approximately 377 Patent Trial and Appeal Board (PTAB) appeal decisions from January 01 to June 25, 2024, specifically reviewing examiner’s rejection based on 35 U.S.C. § 112(a) enablement ground. Within this period, we identified only one decision pertaining to quantum computing technology and enablement. We provide our analysis and offer recommendations to mitigate the risk of enablement rejection.
Enablement Requirement under 35 U.S.C. § 112(a)
In U.S. patent law, enablement is a requirement outlined in the Manual of Patent Examining Procedure (MPEP), specifically in MPEP § 2164. It means that a patent application must describe the invention clearly and in enough detail that someone skilled in the relevant field can make and use the invention without needing to do extensive extra work or guesswork, i.e., without undue experimentation.
The Wands factors, derived from the case In re Wands, 858 F.2d 731 (Fed. Cir. 1988), are used to determine whether a patent application meets the enablement requirement. These factors include:
· The Quantity of Experimentation Needed: How much work is required to practice the invention?
· The Amount of Direction or Guidance Provided: Does the patent give clear instructions?
· The Presence or Absence of Working Examples: Are there practical examples demonstrating the invention?
· The Nature of the Invention: How complex is the invention?
· The State of the Prior Art: What is the existing knowledge in the field?
· The Relative Skill of Those in the Art: What is the skill level of someone in the relevant field?
· The Predictability or Unpredictability of Art: How predictable is the technology area?
· The Breadth of the Claims: How broad are the claims in the patent?
These factors are considered together to determine if the disclosure in the patent application is sufficient for a person skilled in the art to make and use the invention without undue experimentation.
Appeal Decision
Appeal No. 2023002850: Quantum Entanglement and external deep learning (January 23, 2024)
Background:
The case involves a patent application (No. 16/204,784) filed by IBM. The invention pertains to a method where a computing device selects layers from a plurality of external deep learning models, concatenates these layers to form a core deep learning model, trains this core model, and synchronizes its layers with the external models using quantum entanglement.
The application was rejected by the Examiner on several grounds under 35 U.S.C. § 112, relating to issues of indefiniteness, written description, and enablement. IBM appealed these rejections to the PTAB.
Representative Claim:
A method comprising:
· Selecting by a computing device layers from a plurality of external deep learning models;
· Concatenating by the computing device the selected layers from the plurality of external deep learning models to form a core deep learning model;
· Training by the computing device the core deep learning model; and
· Synchronizing by the computing device layers in the core deep learning model with the layers from the plurality of external deep learning models using quantum entanglement.
This claim outlines a method involving the selection, concatenation, training, and synchronization of deep learning model layers, with a step of using quantum entanglement for synchronization. This specific step of using quantum entanglement became a significant point of contention in the enablement rejection, as the Examiner found the specification lacked sufficient detail on how quantum entanglement was to be practically applied in this context.
Key Examiner's Arguments:
Regarding the enablement requirement under 35 U.S.C. § 112(a), the examiner argued that the specification lacked a detailed description of how quantum entanglement is used to synchronize the layers of deep learning models. While the claims included this step, the specification did not provide enough technical detail or guidance on how to implement this process.
According to the Examiner, quantum entanglement is a complex physical phenomenon, and applying it to synchronize deep learning models requires a thorough understanding of both quantum physics and deep learning technologies. The Examiner noted that the specification did not describe any specific starting materials, conditions, or methods for achieving quantum entanglement in the context of deep learning models.
The Examiner conducted an analysis based on the Wands factors which included:
· Quantity of Experimentation Necessary: The Examiner concluded that significant experimentation would be needed to implement the claimed invention.
· Amount of Direction or Guidance Presented: The specification provided minimal guidance on how to achieve synchronization using quantum entanglement.
· Presence or Absence of Working Examples: The specification lacked concrete examples or experimental results demonstrating the use of quantum entanglement in this context.
· Nature of the Invention: The invention's nature involves complex and cutting-edge technology requiring detailed and specific disclosure.
· State of the Prior Art: The Examiner cited references to show that the use of quantum entanglement in such applications is not straightforward or well-understood in the art. The references included:
o Wikipedia Articles: These provided general information on quantum entanglement and qubits, illustrating the complexity and challenges of the technology.
o Cheng et al. Model Compression and Acceleration for Deep Neural Networks (2018): Highlighted the intricacies of deep learning models.
o Phys.org Articles: Discussed advancements and challenges in quantum entanglement, indicating that the technology is still in the experimental stage and not easily applicable to deep learning.
· Relative Skill of Those in the Art: Given the high level of skill required in both quantum mechanics and deep learning, the specification did not provide sufficient details for a person skilled in these areas to implement the invention.
· Predictability or Unpredictability of the Art: Quantum entanglement is highly unpredictable, further complicating the implementation.
· Breadth of the Claims: The claims are broad and cover a wide scope, increasing the need for detailed guidance in the specification.
The Examiner concluded that the specification did not enable the full scope of the claimed invention as it did not provide sufficient details on how to practically apply quantum entanglement to synchronize layers in deep learning models.
Key Arguments Presented by the Appellant:
The appellant argued that the specification includes detailed technical disclosures explaining how the layers in the core deep learning model are synchronized with layers from external deep learning models using quantum entanglement. Specific paragraphs of the specification were cited to support this argument.
Additionally, the appellant referenced the article "Deep Learning and Quantum Entanglement: Fundamental Connections with Implications to Network Design" by Levine, arguing that it provides evidence that quantum entanglement measures are well understood by those skilled in the art.
The appellant asserted that quantum entanglement measurements are employed in the analysis of correlations modeled by deep convolutional networks. The specification’s use of quantum entanglement to synchronize layers in deep learning models aligns with well-established principles in the field.
In response to the Wands Factors Analysis, the appellant contended that the Examiner's application of the Wands factors was flawed. They argued that:
· The amount of experimentation required was not undue given the state of the art and the skill level of practitioners in the relevant fields.
· The specification offered sufficient guidance and direction.
· The absence of working examples did not necessarily imply non-enablement, as the detailed descriptions in the specification should suffice.
The appellant further provided examples and analogies to illustrate how quantum entanglement can be practically applied in synchronizing deep learning model layers. They argued that similar techniques are already being researched and developed in the field, supporting the feasibility of their approach.
PTAB Response
The Board concurred with the Examiner’s assessment that the specification did not enable the full scope of the claimed invention without undue experimentation. They found the Examiner’s analysis of the Wands factors thorough and convincing.
Insufficient Technical Details:
The Board emphasized that the specification lacked sufficient technical details on how to practically implement quantum entanglement for synchronizing deep learning model layers. The references to paragraphs in the specification made by the appellant were not found to provide the necessary guidance.
Quantum Entanglement Complexity:
Quantum entanglement is a complex and advanced concept, especially when applied to synchronizing deep learning model layers. The Board agreed with the Examiner that the specification did not describe any specific starting materials, conditions, or methods for achieving quantum entanglement in this context. The Board pointed out that achieving quantum entanglement typically involves sophisticated hardware and precise conditions, which were not addressed in the specification.
Reliance on External Knowledge:
The Board noted that while the appellant referenced external sources (such as the Levine article) to support their argument, these sources did not compensate for the lack of detailed implementation guidance within the patent application itself. The specification must independently enable the invention without requiring excessive reliance on external sources or additional experimentation.
The Board examined the appellant’s reliance on the Levine article, which discusses the theoretical connections between deep learning and quantum entanglement. The Board found that while the article provided some theoretical background, it did not offer practical implementation details relevant to the claimed invention. The Board emphasized that the specification must show possession of the invention by explaining how the claimed function (synchronizing using quantum entanglement) is achieved, which the Levine article did not do.
Conclusion:
In summary, the PTAB affirmed the Examiner's rejection of the claims under the enablement requirement. The Board found that the specification did not provide sufficient technical details or guidance to enable a person skilled in the art to make and use the claimed invention without undue experimentation. The appellant’s arguments and references to external sources were not adequate to overcome the deficiencies identified by the Examiner.
Strategies to Overcome Enablement Challenges
· Detailed Technical Descriptions: Ensure the specification includes comprehensive descriptions of the invention, specifying starting materials, methods, and conditions. Clear and detailed technical explanations help demonstrate practical implementation.
· Multiple Working Examples: Provide multiple examples illustrating the full scope of the claimed invention. These examples should cover diverse scenarios to demonstrate feasibility across different applications.
· Expert Testimony and Declarations: Include expert testimony, declarations, and post-filing evidence together to support the enablement of complex technologies requiring specialized knowledge.
· Communication with Inventors: Engage with inventors to understand the level of ordinary skill in the art and ensure the specification reflects this level, including necessary technical details and context.
By following these strategies, patent practitioners can better navigate the complexities of quantum computing patents and address enablement issues.
The opinions and recommendations expressed are those of the author and do not reflect the views of Leo Patent Law Office or its clients. This document is for general information purposes only and is not intended to be and should not be taken as legal advice.