Dickinson Wright: A 3-part series of Key Takeaways From Claim Examples In The 2024 AI Patent Eligibility Guideline
Part 1021001010On 17 July 2024, the U.S. Patent and Trademark Office released a new guidance on subject matters eligibility entitled “The 2024 Patent Subject Matter Eligibility Update Including Artificial Intelligence (2024 AI SME update). The 2024 AI SME Update provides patent prosecutors with a valuable tool for prosecuting AI patents. It includes three sets of claim illustrations that show the differences between subject matter eligible and ineligible claims for patent applications directed towards artificial intelligence (AI). This three-part blog series looks at the examples provided with the 2024 AI SME Update in-depth and provides key takeaways for each of the provided examples.
Background of the Office Action Rejection
The subject matter eligibility question arises when the Office issues a Section 101 rejection (35 U.S.C. 101) of the claims recited within a pending utility application. To satisfy a “prima facie” case, the rejection must assert the claims are defective within the bounds set forth in Section 101 rejection. This involves a two-step analysis. Step 1 of the Office’s subject-matter eligibility analysis asks whether the claimed invention fits into one of the four categories outlined in Section 101. Step 2 of ‘Office’s analysis of subject matter eligibility applies the Supreme ‘Court’s framework of two parts (Alice/Mayo), to identify claims directed at a judicial exemption (Step2A Prong One), as well as evaluate if other elements of the claim provide a inventive concept (Step2A Prong Two). In Step 2A, if the claim is determined to be directed towards a judicial exemption, then the analysis continues to Step 2B to evaluate whether or not the claimed additional elements amount to “significantly more” than judicial exception itself. In this case, the analysis proceeds to Step 2B in order to determine whether the claimed elements are “significantly more than” the recited exception itself.
In reality, the rejection is usually based on the two prongs under Step 2A. In Step 2A Prong One for example, a claim is said to “recite” a judicial exemption when it “sets forth” or “describes” the judicial difference in the claim. If the claim does recite a legal exception, then it is eligible and the eligibility analysis is over. MPEP SS 210604, subparagraph II.A.1. If the claim does not cite a judicial exemption, the analysis of eligibility continues to the second prong in Step 2A. This prong (Step 2, Prong Two) is used for determining whether the claim integrates a judicial exception recited into a “practical” application of the exception. The rejection will analyze the claim according to the judicial considerations listed in MPEP SS 210604(d), subsection 1, 2106.04 (d)(1), 2106.04 (d)(2), and 2106.05 (a)-(c), and 2106.05 (e)-h), for example whether the extra element(s), is(are), is(are), merely instructions to apply an exemption; or whether it reflects an improved functioning of a computing device or improvement to another technical field or technology. If the additional element(s) in the claim integrates the judicial exception into a practical application of the exception, the claim is not “directed to” the judicial exception, and the claim is eligible.
2024 AI SME Update: Examples 47-49
With the 2024 AI SME Update, the Office drafts hypothetical and illustrative examples showing the claim analysis performed under MPEP SS 2106. The 2024 AI SME Update also provides a helpful Issue Spotting Chart that is reproduced below for reference:
The Office notes that Examples 47-49 should be interpreted based on the fact patterns that are included with each example and that different fact patterns may have different eligibility outcomes. However, the Office asserts that it is “not necessary” for eligible claims to mirror a claim provided with the examples.
Example 47. Anomaly detection
Key takeaway for Claim 1
: Example47, Claim 1, is an effective example that Applicants can cite to argue patent eligibility for software-based AI inventions, where the claim describes specific hardware components. Because such components do not amount recitation of abstract ideas, the analysis ends at Step 2A Prong One.Claim 1 An application specific integrated circuit (ASIC) for an artificial neural network (ANN), the ASIC comprising:
a plurality of neurons organized in an array, wherein each neuron comprises a register, a microprocessor, and at least one input; and
a plurality of synaptic circuits, each synaptic circuit including a memory for storing a synaptic weight, wherein each neuron is connected to at least one other neuron via one of the plurality of synaptic circuits.
Background for Claim 1:
The application describes structural features of an artificial neural network (ANN). The structure of an ANN, for example, has a series layers, with each layer having neurons arranged into neuron arrays. In this example, the neurons consist of a register, microprocessor and at least one input. Each neuron produces a signal, or activation, using an activation function which uses the outputs from the previous layer as inputs, and a set weights. Each neuron of a neuron network may be connected via a synaptic loop to another neuron. A synaptic network may include a memory that stores a synaptic value. In some embodiments an ANN can be implemented by a application-specific integrated chip (ASIC). ASICs can be customized for a particular artificial intelligence application. They also provide superior computing capabilities and reduce electricity consumption when compared to traditional processors. The application further provides methods for training an ANN that lead to faster training times and a more accurate model for detecting anomalies.
SME Holding for Claim 1:
Claim 1 is eligible. The claim mentions an ASIC that is used for an ANN. While the background explains that “
n ANN can be realized through software, hardware, or a combination of software and hardware,” the broadest reasonable interpretation of the claimed ANN requires hardware because the claimed ASIC is a physical circuit.
Applicant’s Sample Response to a Section 101 Rejection of Claim 1:[a]Step 1: Under MPEP SS 2106.03, the analysis determines whether the claim falls within any statutory category, including processes, machines, manufactures, and compositions of matter. The claim here is directed at a physical device, which is a machine or manufacture and falls under one of the statutory categories for invention. (Step 1: Yes).
Step 2, Prong One: According to MPEP SS 2106.04, the analysis determines if the claim recites an exception. The claim “recites a judicial exemption” when the judicial “exception” is “set forth” in the claim. In this case, the claim does not cite any judicial exception. The claim mentions a plurality neurons, which are hardware elements comprising a microprocessor and a register, and a multitude of synaptic networks, which together make up an ANN. The claim does cite any abstract ideas as defined by MPEP SS 210604(a)(2). This includes a mathematical concept, a mental process, a method for organizing human activity such as managing interactions between people, or a fundamental economic idea. The claim does not mention any mathematical concepts, even though ANNs can be trained with mathematics. The claim is eligible, but the analysis does not proceed to Step 2A Prong Two or Step 2B. The claim is eligible, but the analysis does not proceed to Step 2A Prong Two or Step 2B.
Key Takeaway for Claim 2
:
The reliance on only software processing features and a general “computer,” without enumerating how the software performs or achieves its processing features, dooms the claim to being ineligible.Claim 2. A method of using an artificial neural network (ANN) comprising:(a) receiving, at a computer, continuous training data;
(b) discretizing, by the computer, the continuous training data to generate input data;
(c) training, by the computer, the ANN based on the input data and a selected training algorithm to generate a trained ANN, wherein the selected training algorithm includes a backpropagation algorithm and a gradient descent algorithm;
(d) detecting one or more anomalies in a data set using the trained ANN;
(e) analyzing one or more detected anomalies using the trained ANN to generate anomaly data; and
(f) outputting the anomaly data from the trained ANN.
Background for Claim 2:
The ANN introduced with Claim 1 is trained in Claim 2. The application describes the typical AI training operations and functionalities. The application, for example, describes how the training data is received at a computer as continuous data and then discretized by the computer. Machine learning models can benefit from being trained using discrete data with a limited set of values rather than continuous data. To convert continuous data into discrete data, any discretization method can be used, including binning and clustering. The ANN can be trained using any training technique, including a conventional gradient descent or backpropagation algorithm. The trained ANN then monitors incoming data sets in order to detect anomalies. The trained ANN can detect anomalies and analyze them to generate anomaly information. This data can be used to retrain the ANN or to output to a user. For example, the anomaly data may explain the type of anomaly or the cause of the anomaly.
Anomaly detection is an important task that impacts any industry that benefits from identifying abnormal data that deviates from expected data or from a general pattern. An intrusion detection system, for example, may use the disclosed method of anomaly detection to improve the detection malicious network packets. In order to detect anomalies, a system needs to accurately classify data and define the boundary between normal and anomalous data. It can be difficult to distinguish between normal and anomalous data when the cases are close to a boundary or based on a domain-specific application. Minor variations can trigger the identification of an anomaly when it comes to network security or medicine. However, more significant deviations could be considered normal for less sensitive applications. Malicious actors may also try to make anomalies look like normal activity. This application provides solutions to use a trained ANN for quickly and accurately identifying anomalies compared to anomaly identification performed using traditional methods. The ANN can determine if a detected anomaly in network traffic is linked to a malicious packet if it detects one or several anomalies. If the detected anomaly has been associated with a malignant packet, the ANN can cause a network device drop the malicious package and block future traffic coming from the sender. The present invention improves network security through automatic detection of network intrusions and other malicious attacks. This allows for proactive, automatic remediation of attacks. In some embodiments, a system can use different detection techniques to detect potentially malicious packets. It can also alert a network administrator about potential problems. The system can detect the source of potentially malicious network packets through a trace operation or software tool. The disclosed system detects intrusions into networks and takes remedial measures, such as automatically dropping suspicious packets or blocking traffic from suspicious source address without needing to alert a network admin. The disclosed system improves network security by eliminating the delay of waiting for a network administrator to react to a network intrusion. It does this by automatically dropping suspicious packets, and blocking traffic from suspicious sources addresses based on anomalies identified by the ANN in real time. The disclosed system realizes an improvement in network security by avoiding the delay involved in waiting on a network administrator to react to a network intrusion by automatically dropping suspicious packets and blocking traffic from suspicious source addresses based on anomalies identified by the ANN in real time.
SME Holding for Claim 2:
Claim 2 is ineligible.
Step 1: Under MPEP SS 2106.03, the analysis determines whether the claim falls within any statutory category, including processes, machines, manufactures, and compositions of matter. The claim cites at least one act or step, such as receiving continuous training data. The claim is directed at a process which is a statutory category of invention. Step 2A, Prong 1: Under MPEP 2106.04(II), a claim “recites a judicial exemption” when the judicial Exception is “set forth” (described) in the claim. Claim Elements b), d), and e) fall under the mental process groupings abstract ideas, because they cover concepts performed by the human mind. These include observation, evaluation, judgement, and opinion. See MPEP SS 210604(a)(2) (III). Claim elements (b) and(c) are directed at mathematical concepts. Step 2A Prong One YES.
Step 2, Prong Two : Under MPEP SS 210604(d), it is determined whether the entire claim integrates the judicial exemption into a practical application or if the claim “directs to” the exception. The claim cites the additional elements “(a), receiving, at a computing device, continuous training data”, “using the trained ANN”, in Claim Elements d) and e), and “f), outputting anomaly data from trained ANN”. Claim Elements b) and c) are performed using a general computer. Claim Elements a) and f) are merely data gathering and output, recited with a high degree of generality. They are therefore insignificant extra-solution activities. Claim Elements a, b, and c are recited at a high generality level as being performed by computers. Claim Elements d) and e) recite that “using the trained ANN”, which is nothing more than a simple instruction to implement an abstract concept on a generic PC. Claim Elements d) and e) merely indicate the field of use or the technological environment in which a judicial exception is performed. This merely confines use of the abstract concept to a specific technological environment (neural network) and fails to add inventive concepts to the claims. (Step 1A Prong One: NO).
Step 3: Under MPEP SS 210605, the analysis evaluates if the claim as a entire amounts to “significantly” more than the recited exemption, i.e. whether any additional element or combination of additional components adds an innovative concept to the claims. As a general practice tip, if a claim cannot provide an “application” of the abstract concept, it cannot “significantly” provide more than the abstract concept. (Step 2B: NO).
Applicant’s Sample Response to a Section 101 Rejection of Claim 2:
Respond to the rejection by amending your claims in a way that more closely aligns them with Example 47, Claim 1 or Example 47, Claim 3. (Step 2B: NO).
Applicant’s Sample Response to a Section 101 Rejection of Claim 2:
Respond to the rejection with amendments to your claims that more closely align them with Example 47, Claim 1 or Example 47, Claim 3. Also, when you are initially drafting the application, ensure that the how/why of the invention is described to avoid this type of rejection during prosecution.
Key Takeaway for Claim 3
: Example 47, Claim 3 is another helpful example for Applicants to cite when arguing patent eligibility of software-based AI inventions, especially under Step 2A Prong Two. Although the claim recites limitations that can be interpreted as mental processes/mathematical concepts (backpropagation and gradient descent algorithms), as a whole, the claim integrates the judicial exception into a practical application vis-a-vis improving computer functionality or improving a technological field (dropping… malicious network packets/blocking future traffic from the source address).Ineligible Claim 2 and eligible Claim 3 are provided below for reference:Claim 2. A method of using an artificial neural network (ANN) comprising:
(a) receiving, at a computer, continuous training data;
(b) discretizing, by the computer, the continuous training data to generate input data;
(c) training, by the computer, the ANN based on the input data and a selected training algorithm to generate a trained ANN, wherein the selected training algorithm includes a backpropagation algorithm and a gradient descent algorithm;
(d) detecting one or more anomalies in a data set using the trained ANN;
(e) analyzing the one or more detected anomalies using the trained ANN to generate anomaly data; and
(f) outputting the anomaly data from the trained ANN.
Claim 3. A method of using an artificial neural network (ANN) to detect malicious network packets comprising:
(a) training, by a computer, the ANN based on input data and a selected training algorithm to generate a trained ANN, wherein the selected training algorithm includes a backpropagation algorithm and a gradient descent algorithm;
(b) detecting one or more anomalies in network traffic using the trained ANN;
(c) determining at least one detected anomaly is associated with one or more malicious network packets;
(d) detecting a source address associated with the one or more malicious network packets in real time;
(e) dropping the one or more malicious network packets in real time; and
(f) blocking future traffic from the source address.
Background for Claim 3:
Same as Claim 2.
SME Holding for Claim 3:
Claim 3 is eligible.
Step 1: Under MPEP SS 2106.03, the analysis determines whether the claim falls within any statutory category, including processes, machines, manufactures, and compositions of matter. The claim consists of a series steps, and is therefore a process. (Step 1 – YES).
Step 2, Prong One – Under MPEP SS 210604(II), an analysis is performed to determine whether the claim cites a judicial exemption. The claim “recites a judicial exemption” when the judicial is “set forth” or described in the claim. Claim Element a recite mathematical calculations (a gradient descent algorithm and a backpropagation algorithms) to perform the ANN training and encompasses mathematical concepts. Claim Elements b) and c) refer to concepts that can be performed by the human brain, such as “detecting one of more anomalies in network data” and “determining that one detected anomaly has been associated with at least one malicious network packet.” Claim Element d)-f) do NOT refer to mental processes that are practical for the human brain. The human mind cannot detect a source associated with malicious packets of network traffic, drop them in real-time, and block any future traffic, as stated in the claim. Since Claim Elements a, b, and c fall into different groups of abstract ideas (i.e. mathematical concepts and mental process, respectively), the analyses proceeds to Prong Two, under a single abstraction idea. (Step 2, Prong One, YES).
Step 2, Prong Two: Under MPEP, SS 2106.04(d), an analysis is performed to determine whether the entire claim integrates the judicial exemption into a practical application or whether it is “directed” towards the judicialexception. If the claimed invention improves a computer’s functionality or another technology or field, this can be used to determine whether it is integrated into a practical application. To evaluate an improvement made to a computer, or technical field in the specification, and the claim must reflect that improvement. See MPEP SS 210604(d)(1), and 2106.05 (a). Claim Elements d, e, and f, in view of the entire claim, include an improved computer or a technological field. This requires an evaluation of both the specification as well as the claim in order to ensure that the specification contains a technical explanation for the asserted improvements and that the claimed improvement is reflected in the claim. As recited in background, existing systems can alert a network admin to potential problems by using various detection techniques. The disclosed system detects intrusions into the network and takes real-time remedial measures, such as dropping suspicious packets or blocking traffic from suspicious sources. The background section explains further that the disclosed system enhances network security by acting in real time to proactively prevent intrusions. Claim Elements d, e, and f also provide for improved security of the network by using the information from detection to enhance security. This is done by taking proactive steps to remediate the danger. These steps reflect the improvements described in the background. The claim as a unit integrates the judicial exemption into a practical application, such that it is not directed at the judicial (Step 2A: Yes). (Step 2A: NO). The claim is eligible.
Applicant’s Sample Response to a Section 101 Rejection of Claim 3:
Step 1: Under MPEP SS 2106.03, the analysis determines whether the claim falls within any statutory category, including processes, machines, manufactures, and compositions of matter. The claim consists of a series steps, and is therefore a process. (Step 1: Yes).
Step 2, Prong One: According to MPEP SS 2106.04, the analysis determines if the claim recites an exception. The claim “recites a judicial exemption” when the judicial is “described” or “set forth” in the claim. In this case, the Office Action alleges two abstract ideas, including mathematical concepts, and mental processes. The features of claim 3, however, are directed at non-abstract ideas performed on a particular computing device. If the Office finds that the claims are abstract ideas, the analysis moves to Step 2A, prong two. (Step 2, Prong One: YES).
Step 2, Prong Two: According to MPEP SS 2106.04, the analysis determines if the claim as a entire integrates the judicial exemption into a practical application or if the claim is “directed towards” the judicialexception. In this case, Claim Element (d), (e), or (f) include, in the context of the entire claim, an improvement to a technology field or computer. In order to analyze these Claim Elements, it is necessary to evaluate the specification as well as the claim in order for the claim itself and the specification to reflect the improvement asserted. As recited in background, existing systems can alert a network admin to potential problems by using various detection techniques. The disclosed system detects intrusions into the network and takes real-time remedial measures, such as dropping suspicious packets or blocking traffic from suspicious sources. The background section explains further that the disclosed system enhances network security by acting in real-time to proactively prevent intrusions. Claim Elements d, e, and f also provide for improved security of the network by using the information from detection to enhance security. This is done by taking proactive steps to remediate the danger. These steps reflect the improvements described in the background. The claim as a unit integrates the judicial exemption into a practical application, such that it is not directed at the judicial (Step 2A Prong Two: Yes). The claim is eligible for Step 2A Prong Two but the analysis does NOT proceed to Step 2B.
