Engineering Patent Applications

𝗘𝗣𝗢 𝗕𝗼𝗮𝗿𝗱 𝗼𝗳 𝗔𝗽𝗽𝗲𝗮𝗹 𝗖𝗹𝗮𝗿𝗶𝗳𝗶𝗲𝘀 𝗗𝗶𝘀𝗰𝗹𝗼𝘀𝘂𝗿𝗲 𝗥𝗲𝗾𝘂𝗶𝗿𝗲𝗺𝗲𝗻𝘁𝘀 𝗳𝗼𝗿 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗣𝗮𝘁𝗲𝗻𝘁𝘀: 𝗔 𝗠𝘂𝘀𝘁-𝗥𝗲𝗮𝗱 𝗳𝗼𝗿 𝗣𝗿𝗮𝗰𝘁𝗶𝘁𝗶𝗼𝗻𝗲𝗿𝘀 The EPO's recent decision in T 1669/21, concerning a method for predicting wear in metallurgical vessels using a "computational model", offers valuable insights into the current EPO approach to examining machine learning patent applications. The case, which was rejected due to insufficient disclosure, provides important lessons for patent practitioners. 𝗪𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝘁𝗵𝗶𝘀 𝗺𝗲𝗮𝗻 𝗳𝗼𝗿 𝗽𝗮𝘁𝗲𝗻𝘁 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗲? The decision underscores the need for meticulous attention to detail when drafting patent applications for machine learning inventions. Specificity is paramount. While the EPO accepts broad claims, these must be supported by a commensurately detailed and enabling disclosure. Here's what the EPO expects - 𝗖𝗹𝗲𝗮𝗿 𝗱𝗲𝗳𝗶𝗻𝗶𝘁𝗶𝗼𝗻 𝗼𝗳 𝘁𝗵𝗲 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗺𝗼𝗱𝗲𝗹: The type of model (e.g., neural network, support vector machine), its architecture, and the specific algorithms used must be explicitly stated. Simply referring to a generic "computational model" is insufficient. 𝗗𝗲𝘁𝗮𝗶𝗹𝗲𝗱 𝗽𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿 𝗺𝗮𝗽𝗽𝗶𝗻𝗴: The application must provide clear guidance on how to select, pre-process, and represent input parameters within the model. This includes specifying how to handle time-varying or multi-dimensional parameters. Examples are crucial for illustrating these steps. 𝗧𝗿𝗮𝗻𝘀𝗽𝗮𝗿𝗲𝗻𝘁 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗽𝗿𝗼𝗰𝗲𝗱𝘂𝗿𝗲𝘀: The description should cover the training data used, the training process, and the criteria for evaluating model performance. It should also address potential challenges such as data scarcity and the prevention of artefacts from random correlations. 𝗪𝗼𝗿𝗸𝗶𝗻𝗴 𝗲𝘅𝗮𝗺𝗽𝗹𝗲𝘀: Where possible, include concrete, workable examples demonstrating the implementation of the invention. This could involve providing sample data, model configurations, and training scripts. 𝗪𝗵𝘆 𝗶𝘀 𝘁𝗵𝗶𝘀 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁? The EPO's focus on specificity stems from the Article 83 EPC requirement for sufficient disclosure. The patent application must enable a skilled person to carry out the invention without undue burden. This is particularly challenging for machine learning inventions, which often involve complex models and data-driven processes. 𝘓𝘪𝘯𝘬𝘴 𝘵𝘰 𝘧𝘶𝘭𝘭 𝘣𝘭𝘰𝘨 𝘱𝘰𝘴𝘵 𝘢𝘯𝘥 𝘵𝘩𝘦 𝘥𝘦𝘤𝘪𝘴𝘪𝘰𝘯 𝘪𝘯 𝘵𝘩𝘦 𝘤𝘰𝘮𝘮𝘦𝘯𝘵𝘴. #EPO #MachineLearning #Patents #SufficiencyOfDisclosure #PatentPractice #CaseLaw

She took out a loan at 25% interest just to afford the patent office fee. What she created still holds the fashion industry together at the seams, more than a century later. Helen Blanchard wasn't supposed to be an inventor. Her family had been wealthy Portland shipping merchants until the financial panic of 1866 wiped them out. After her father died, they lost everything - their family home. She ended up working the floor of a Boston clothing factory. Day after day, the same thread running through every problem. Sewing machines could only sew straight. You couldn't reinforce a buttonhole. You couldn't work with stretchy fabrics. You couldn't seal fabric edges to prevent fraying. Everything that needed strength or flexibility had to be done by hand. Helen had no technical training, no engineering background, no capital. She went about the problem sideways. Helen changed the machine so the needle moved side-to-side as it went forward. ////\ instead of ———— That back-and-forth pattern made possible what straight stitches couldn't: seal edges, reinforce buttonholes, handle stretch fabrics. Everything that had required hand-sewing could now be done by machine. She wasn't a one-stitch wonder. Over the next 42 years, Helen was granted 28 patents. Twenty-two were adopted by large commercial factories - a 78% commercialization rate that remains extraordinary even today. Her machines became factory standard equipment and helped enable the growth of affordable ready-to-wear clothing. The financial success matched the technical achievement. Helen and her sister Louise founded two companies together. By the time she was 50, she'd earned enough to buy back her family's ancestral estate in Portland - the same property they'd lost decades earlier. But Helen recognized something her fellow industrialists didn't. Her labor-saving machines were displacing the women she'd worked alongside on factory floors. In the 1890s, she moved to New York and devoted herself to helping women who had lost work in clothing factories because of her inventions. She used her fortune to support them - not as charity, but as responsibility. AI is displacing workers at scale right now. Helen Blanchard faced the same dilemma 150 years ago - and actually tried to solve it. What can today's tech builders learn from her? #ipidity #patents #sheinvents #fashiontech #seamslegit #stitchesgetriches

Do the names Elias Howe or Isaac Singer ring a bell? They should - they’re two of the key figures behind one of the most revolutionary inventions of the Industrial Age: the sewing machine. While the wheel, telephone, and personal computer often steal the spotlight for transforming daily life, the humble sewing machine is one of history’s true unsung heroes. Before its invention in the 19th century, making clothes was an incredibly time-consuming task. Every garment, from a shirt to a suit, had to be sewn by hand - a labor of love and necessity. Thanks to innovators like Howe (who patented the lockstitch design in 1846) and Singer (who improved and commercialized the machine), sewing quickly became faster, more consistent, and more accessible. This single invention didn’t just change how people made clothes, it changed who could make them. It opened doors for home-based entrepreneurs, fueled the growth of the fashion industry, and quite literally stitched its way into millions of households. Today is National Sewing Machine Day, a moment to appreciate this powerful little machine that’s still helping us turn fabric into function from hemming jeans and crafting curtains to bringing one-of-a-kind designs to life. For some, sewing is a profession. For others, it’s therapy. For many, it’s a treasured tradition passed down through generations. The photo? That’s my grandma’s Singer, a well-loved machine I inherited and cherish. A reminder that great design never goes out of style.

Do you like handwashing dishes? I don’t. Fortunately, because of Josephine Cochrane, I don’t have to; I can use a dishwasher. Cochrane was a socialite who prized her 17th-century heirloom china, which her servants kept chipping (not a problem I have, LOL). She came up with the idea of an automatic dishwasher and shared it with her husband and others, but in an era that did not value women’s time, no one saw any importance in the idea. So she decided to do it herself. Her life took a dramatic turn in 1883 when she became a widow, but she persevered in her endeavor, collaborating in secret with George Butters, a mechanical engineer. This secrecy was necessary due to the societal norms of the time, which frowned upon a young man and a widowed woman being seen together. While existing dishwashing machines relied on abrasive brushes, Cochrane's revolutionary design used water pressure, ensuring a gentle yet effective cleaning process. Initially, her dishwasher was priced at about $100, equivalent to about $3,000 today, making it unaffordable for most households at a time when manual labor was inexpensive. However, Cochrane had a stroke of insight and identified a perfect market in restaurants and hotels, where large volumes of dishes needed washing. At the 1893 Chicago World's Fair, nine of her washers cleaned the dishes in its pavilions and restaurants. Cochrane's dishwasher offered an efficient and gentle solution for dishes, leading to widespread adoption in commercial settings…and, eventually, households like mine. #Illdoitmyself #WednesdayWomen #Inventor #STEM #History #Herstory 

In 1879, Mary Walton, a groundbreaking American inventor, patented an innovative system to reduce train #pollution by funneling smoke through water. Her system aimed to improve air quality in cities affected by the thick smoke produced by trains, which was a major concern in rapidly #industrializing America. Walton's invention utilized water to trap the harmful particles in the smoke, making the air cleaner and reducing environmental damage. This approach was ahead of its time, showcasing her ingenuity and forward-thinking approach to solving urban #pollution problems. Walton didn’t stop with pollution control. After learning of Thomas Edison's failure to reduce noise from elevated #railway tracks in New York City, she took it upon herself to tackle the problem. Edison's method of sound-dampening for railway tracks had proven unsuccessful, but Walton saw it as an opportunity. In response, she developed a new sound-dampening system that was more effective, offering relief to New Yorkers who lived near the noisy and disruptive elevated trains. Her success in this venture further demonstrated her ability to solve complex issues and her critical role in shaping early #environmental #engineering. Walton's contributions to industrial innovation helped shape the way cities dealt with pollution and noise in the late 19th century. While her inventions were not widely celebrated during her time, her work laid the groundwork for later advancements in environmental engineering. Her legacy is an example of how women in science and engineering, though often overlooked, have made significant contributions to technological progress and urban development. Walton's work remains a testament to the power of innovation in improving public health and quality of life. #InnovativeWomen #EnvironmentalEngineering #WittyHistorian

#Section 3(d): efficacy and data A patent application is filed for an invention – to demonstrate technical advancement, to show a workable concept. Data supports working of an invention and upon expiry, enable its replication by the public.  Another function of data is utility – i.e show that invention operates and is useful for its claimed purpose. Data helps to establish this utility. A further function of data is to support enablement requirement -support the claims.  So data provides real-world examples and context for understanding what the invention is intended to accomplish. Should this data be provided in the patent or could it be provided upon demand by the Examiner? Lets examine the second position. If no data is provided in the patent, and one waits for the examiner or opponent to point out – there is an enormous risk – how do you justify plausibility of the invention, the utility, enablement and what do you teach the public after its expiry? Almost amounts to writing one line to obtain a patent. The first position i.e of providing data in the specification invites the next query – how much and when?  While there are no rules on the amount of data, the specification must evince technical advancement through some data. We will deal with enhanced efficacy in the context of section 3(d) in a separate post. Lets examine now data for technical advancement. The Delhi High Court in Astrazeneca vs Intas has held that post-filing data to support technical advancement is admissible provided a “seed of that nature” is planted in the specification. Therefore, it boils down to “basis” in the specification. Which is same approach in EP. And courts have followed this approach in many cases, the latest being Ischemie vs Controller of Patents (Del HC, order dated 22/11/23). In that case, patent was sought for an isomer of a known compound. The grievance was that though there is application included data in the form of in-vitro, in-vivo studies/clinical trials, the same was ignored by the Controller, who eventually was directed to consider the same and matter was remanded.  This case presented an opportunity to examine the difference between data for efficacy and data for technical advancement - the court has made no such distinction. However, there is an important take-away here – the categorical finding that a) efficacy data and explanations should be presented during oral hearings at the patent prosecution stage b) post-filing data can be relied upon to support claims subject to “seed” of that data being found in the original specification. One cannot establish a technical effect for the first time without basis in the specification. I would assume that this applies to data for technical advancement as well as efficacy. Practice tip – ensure sufficient data in specification to support technical advancement or at least a description thereof so that an opportunity to present actual results can be availed post-filing. 

Another PTAB ex parte 101 reversal. I just filed this in a supplemental response because one of our clients has a current 101 rejection, which is rather conclusory, involving steps for training a machine learning model. This decision is useful in 3 ways: (1) reinforces that the USPTO's 101 Example 39 is valid after the issuance of AI Examples 47 to 49, (2) shows that the particularity of the training is not necessarily controlling, and (3) states the genericity of computer elements is not relevant to whether the claim recites an abstract idea related to training a machine learning model. With respect to (1), the decision states: "[A]s Appellant notes, the analysis accompanying Example 39 suggests that the Office’s view is that training a machine learning engine is not a mental process. See SME Examples: Abstract Ideas 9 (hypothetical claim reciting training a neural network “does not recite a mental process because the steps are not practically performed in the human mind”)." With respect to (2), the decision states: "[T]he Examiner further supports its mental-process findings by contending that the claims “are recited very generically, with little to no particularity,” so “there is nothing to suggest that the skilled artisan would not be able to perform the highly generalized identified judicial exception mentally, or using simple pen/paper.” Ans. 3–4. But the degree of specificity in which a limitation is recited has little to do with whether it is a mental process. Nuclear fission may be recited generically in a claim, but that does not make nuclear fission a mental process." With respect to (3), the decision states: "The Examiner further argues that “even if considered that the identified judicial exception is too complex to be practically performed by a human, . . . the claims should not be patent eligible as they pertain to implementing and performing generic computer functions on a generically recited computer.” Ans. at 4. This argument conflates the Step 2B and Step 2A, Prong 2 analyses with the Step 2A, Prong 1 analysis, however. For example, performance of a judicial exception on a generic computer performing generic computer functions is relevant to whether additional elements in the claim integrate the judicial exception into a practical application. See Patent Eligibility Guidance, 84 Fed. Reg. at 55 (stating that additional elements may not integrate a judicial exception into a practical application if the element “uses a computer as a tool to perform an abstract idea”). But the presence in the claim of elements that recite generic computer implementation is not necessarily relevant to whether the claims recited a judicial exception in the first place." Yes, I really do read PTAB decisions every day. Yes, I am an appeal nerd. #patents #patentlaw #patentattorney

📝 Reviewing your first patent application? Before filing a patent application, patent attorneys generally ask their clients to review the draft application. Especially if this will be your first patent, this can be a bit daunting. Where to start? What to look for? Here are some tips: 1. Begin with the claims. They're the most important part as they will define what will be protected. 2. Check the detailed description. Its purpose is to support the claims. 3. Read the background section. It provides context for the invention but has only limited impact. Key steps when reviewing the claims: 1. ✅ Ensure all essential features are included 2. 🚩 Flag potential ambiguities or easy workarounds 3. 🔍 Verify what is claimed covers what you plan to do or sell 4. 🔢 Check for consistent use of terminology and proper claim references Key steps when reviewing the description: 1. 📚 Verify all claimed features are described 2. 💡 Highlight key advantages of your invention 3. 📏 Specify how measurement data is obtained 4. 🔗 Check all terminology is consistent with claims 5. 🖼️ Confirm figures are clear, labeled, and referenced 6. 🔬 Ensure sufficient detail for others in the field to reproduce the invention 7. ⚙️ Include all feasible variations and embodiments unless there are strategic reasons to do otherwise Remember, as an inventor, you know the invention best. Your input is thus crucial in ensuring the application accurately and comprehensively describes the invention. Also take the time you need to thoroughly review the draft. A patent application is a significant investment and once filed, no new information can be added. Good luck reviewing! And if you have any further helpful tips or experiences, please share them in the comments below.

𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝘁𝗶𝗲𝗱 𝘁𝗼 𝗽𝗮𝗿𝘁𝗶𝗰𝘂𝗹𝗮𝗿 𝗺𝗮𝗰𝗵𝗶𝗻𝗲𝘀 𝗼𝗳𝗳𝗲𝗿𝘀 𝗮 𝗰𝗹𝗲𝗮𝗿 𝗽𝗮𝘁𝗵 𝘁𝗼 𝗽𝗮𝘁𝗲𝗻𝘁 𝗲𝗹𝗶𝗴𝗶𝗯𝗶𝗹𝗶𝘁𝘆, 𝗲𝘃𝗲𝗻 𝘄𝗵𝗲𝗻 𝗼𝘁𝗵𝗲𝗿 𝗮𝗽𝗽𝗿𝗼𝗮𝗰𝗵𝗲𝘀 𝗳𝗮𝗶𝗹. When your software patent application faces subject matter eligibility challenges under 35 U.S.C. § 101, a powerful strategy for responding to such rejections involves demonstrating that your software 𝗶𝘀 𝘁𝗶𝗲𝗱 𝘁𝗼 𝗮 𝗽𝗮𝗿𝘁𝗶𝗰𝘂𝗹𝗮𝗿 𝗺𝗮𝗰𝗵𝗶𝗻𝗲. 𝗘𝘅𝘁𝗲𝗿𝗻𝗮𝗹 𝗠𝗮𝗰𝗵𝗶𝗻𝗲𝘀 𝗖𝗼𝗻𝘁𝗿𝗼𝗹𝗹𝗲𝗱 𝗯𝘆 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 Some of the strongest examples of patent-eligible software involve control of external machines: 1. 𝗥𝗼𝗯𝗼𝘁𝗶𝗰 𝘀𝘆𝘀𝘁𝗲𝗺𝘀: Software that controls robots with specific movement patterns or adaptive behaviors based on sensor inputs 2. 𝗩𝗲𝗵𝗶𝗰𝗹𝗲 𝗰𝗼𝗻𝘁𝗿𝗼𝗹 𝘀𝘆𝘀𝘁𝗲𝗺𝘀: Software implementing specific methods for controlling autonomous vehicles, engine management systems, or advanced driver assistance features 3. 𝗠𝗲𝗱𝗶𝗰𝗮𝗹 𝗱𝗲𝘃𝗶𝗰𝗲𝘀: Software controlling the operation of specialized diagnostic or therapeutic equipment, such as MRI machines, radiation therapy systems, or drug delivery devices 𝗣𝗮𝗿𝘁𝗶𝗰𝘂𝗹𝗮𝗿 𝗠𝗮𝗰𝗵𝗶𝗻𝗲𝘀 𝗳𝗼𝗿 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗜𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 Software can also be patent-eligible when implemented on specialized computing hardware, such as: 1. 𝗤𝘂𝗮𝗻𝘁𝘂𝗺 𝗰𝗼𝗺𝗽𝘂𝘁𝗶𝗻𝗴 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲𝘀: Software specifically designed to leverage quantum gates, qubits, and quantum phenomena to solve particular problems 2. 𝗖𝘂𝘀𝘁𝗼𝗺 𝗔𝗦𝗜𝗖𝘀 𝗼𝗿 𝗙𝗣𝗚𝗔𝘀: Software implemented using application-specific integrated circuits or field-programmable gate arrays designed for particular purposes 3. 𝗡𝗲𝘂𝗿𝗮𝗹 𝗽𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 𝘂𝗻𝗶𝘁𝘀: Software utilizing specialized neural network hardware accelerators with specific architectures The key is demonstrating that the machine is not merely an afterthought but integral to the invention's purpose and functionality. As MPEP 2106.05(b) notes, the machine must "play a significant part in permitting the claimed method to be performed, rather than function solely as an obvious mechanism for permitting a solution to be achieved more quickly." 𝗣𝗿𝗮𝗰𝘁𝗶𝗰𝗮𝗹 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗧𝗶𝗽 While the "particular machine" exception won't apply to many software inventions running on conventional computers, it's always worth considering this pathway to eligibility. When it does apply, it can be one of the most straightforward and effective ways to overcome a patent eligibility rejection. In my next post, I'll explore how computer-implemented systems and methods that contain an unconventional arrangement of steps or components can establish patent eligibility. #patents #patenteligibility #IP

Section 101 rejections remain one of the toughest hurdles for software patents. How you approach it can make all the difference. The real question isn’t whether the invention does something useful—it’s whether it solves a technical problem in a technical way. That’s where many applications go wrong. Too often, they frame the idea as a business objective implemented on a computer. The USPTO sees right through that. What examiners want to see is a clear, concrete improvement in computing itself—faster processing, better memory use, more secure data handling, or new interactions between hardware and software. When I work with software clients the goal is always the same: make the invention sound like engineering, not abstraction. That means precise claims, supported by a detailed specification with diagrams, pseudocode, and performance data. If you can show how the invention improves technology, you’ve already done half the work of overcoming Alice. I also encourage clients to anticipate 101 early. Build the eligibility rationale right into the application. Don’t wait for a rejection to start thinking about it. During prosecution, be prepared to cite the USPTO’s own examples, talk to your examiner, and if necessary, use declarations to emphasize what’s unconventional about the invention. Software remains patentable in the U.S.—but only when it’s drafted as a technical solution, not a business plan translated into code. The difference is subtle, but it’s the difference between rejection and allowance.