Discover more from PATENT DROP
PATENT DROP: Microsoft lightens your load
Plus: Okta’s voice 2FA; JPMorgan Chase automates the intern
Happy Monday and welcome to Patent Drop!
Today, we’re digging into Microsoft’s tech to make your VR headset less bulky; Okta’s plan to add your voice into password protection, and JPMorgan Chase’s goal to automate the grunt work away.
Quick programming note: Today’s newsletter features a story by freelance guest writer Samson Amore.
Drum roll, please!
Microsoft’s face reader
Microsoft wants to really understand the look on your face.
The tech firm is seeking to patent a “high-sensitivity” facial tracking system for mixed-reality devices. Microsoft says its system can detect facial movement with less hardware at a lower cost.
Microsoft’s filing is quite technical, but to break it down: Conventional systems track facial movements using circuits which collect sensor data that’s processed by an “analog-to-digital converter.” These converters need to be integrated into different components of the device and take up extra space as they need to be “as large as possible” to get higher-resolution facial tracking.
In contrast, Microsoft’s system integrates those tracking circuits directly into the components, and instead uses what’s called a “DC-AC converter,” thereby reducing the “required circuit area” needed for processing.
Basically, Microsoft is trying to cut down the job of the middleman that is the “expensive and bulky inductor.” This not only reduces power consumption, but can cut manufacturing costs by making the hardware itself simpler, and increase device reliability and safety.
By tracking movements so closely, Microsoft said it can “enhance the immersive MR experience.” But conventional systems to do so require high power-supply levels, sucking up battery quickly in a typical headset.
“Low power and compact form factor are important design issue(s) in such wearable devices, which require portability and long battery life,” Microsoft said in the filing.
As complicated as this sounds, the outcome matters more than the tech itself. Microsoft’s system could lead to a cheaper, lighter, and longer-lasting artificial reality headset that includes advanced facial tracking. And given how “processor intensive” facial tracking can be, making it energy-affordable and efficient could be the key to increased production, said Jake Maymar, VP of Innovation at The Glimpse Group.
And if the tech in this patent works well at capturing facial expressions in an energy-efficient and cost-saving way, Maymar said, Microsoft would have access to “one of the most lucrative metrics you can possibly have” in AR and VR.
“If Microsoft creates this system, it's just easy for the manufacturers to implement it, and they’re licensing that technology and making it affordable, that becomes the standard,” said Maymar. “That's incredibly powerful. You can actually have a clear metric and be able to measure the impact of advertising, the impact of experiences, and the impact of connections with others.”
Several companies have shown interest in reading the look on your face. Meta files patent applications for VR eye and facial expression tracking at a near-constant pace, Snap has sought to patent several similar tools for its AR spectacles, and Sony is interested in tracking your face for its own gaming purposes. Apple, meanwhile, has touted gaze-tracking as a major feature of its ritzy Vision Pro headset.
By securing this patent, Microsoft could control a key element that several of these companies could benefit from, said Maymar. “The thing that I'm watching for is who can capture the actual micro expressions of the face,” said Maymar. “The company that can do that in the cheapest and most efficient way is the one that's going to ultimately win.”
Microsoft offers a line of mixed-reality headsets called the HoloLens, starting at $3,500 for its cheapest model. With this new tech, it may be able to bring the price down. Plus, Maymar said, with the wide scope that Microsoft’s business now has, it could try to sell this tech to everyone — from businesses looking for ways to connect remote employees to gamers looking for immersive experiences.
Okta’s AI-powered voice scanner
Okta is looking at a new way to get rid of passwords.
The identity management company wants to patent a multi-factor authentication system that uses AI to analyze a person’s voice and verify their identity. Okta’s system would “allow voice to be one credential type” among several in a multi-factor authentication system – meaning users likely would need to also input some other kind of password to identify themselves.
Okta’s system uses a neural network voice model to identify a user based on a “small number of sample utterances” spoken. It also can work independent of text, meaning it's less susceptible to relay attacks from multiple devices.
The system can also recognize the same voice even if the user is speaking different phrases or using different tones. According to the patent, Okta does this by training the model to create embedding vectors that capture unique facets of the user’s voice, including pitch, pronunciation or acoustics.
Ari Weil, VP of marketing at cybersecurity firm Cyera, explained why biometric authentication methods are becoming more common: They’re easier to use than a password, which leads to a better user experience, and they’re more trusted by organizations when it comes to fighting fraud.
This last part is especially crucial, as AI deepfakes become more popular and realistic.
AI voices are already so convincing that most humans can’t tell the difference. A recent study out of University College London found that people only correctly identified deepfaked voices 73% of the time. “As the use of machine learning and AI advances… The ability to create and train models using biometrics is becoming an important way to identify that individuals are who they claim to be,” Weil said.
Okta noted the model would account for these fraud attempts. While onboarding a user, the model will store their speech audio data in an embedded vector, then always compare future audio to that data to compute “a degree of similarity” between them.
But biometrics’ uniqueness could also be its downfall. Aubrey Turner, executive advisor for Ping Identity, said that while biometrics can be “harder to steal and reuse, unlike passwords,” if the data is stolen, “you can't really just infinitely swap them out like passwords.”
AI can be both “a gift and a curse” when it comes to biometric security systems like Okta’s, Turner cautioned. Just as AI is being implemented to improve security, “it is also already being used to scale and build more sophisticated phishing attacks, malware and deep fakes.” And in Okta’s case, after two breaches last year, it’s possible some customers’ compromised data could later be used for deepfakes.
JPMorgan’s AI intern
JPMorgan Chase wants to save bankers the time, headaches and intern-wrangling it takes to get the lowdown on other companies.
The financial institution is seeking to patent a system to “automatically discover and extract information” on public and private companies. Using what it calls an “artificial intelligence-based framework” that includes machine learning and natural language processing, the system essentially digs up all publicly available information on a company and aggregates it into one place.
First, a user tells the system what company they’re interested in, and it accesses a well of public data on the company with what it calls a “recursive web crawling operation.” The user can then pick from that well what data in particular they want, and the system displays information via a user interface.
This system can dig a little bit deeper on a company or employee than your average Google or LinkedIn search. Chase notes that this system can scrape for information related to individuals, products and companies as a whole, including number of employees, leadership, assets under management, social media and company policy information, specifically noting that it can find ESG information.
But given that patents often don’t go into the full scope of possible uses, this system can likely extract even more than what Chase noted from the aforementioned “public data sources.”
Frankly, it’s not a super complex system, but Chase said it has the potential to help the institution make financial decisions about companies more easily, as currently, “there are huge challenges for collecting these data manually on a large scale that may approach millions of companies,” it noted.
This filing adds to the abundance of AI tools we’ve seen come out of Chase’s patent activity. The company is working on low- and no-code machine learning tech, an AI-based talent recruiting tool, and a matchmaking algorithm for companies and investors. The company also invested heavily in tech talent last year, and reportedly employs 1,500 engineers and data scientists.
In May, Chase filed for a trademark for a ChatGPT-esque AI investment advisor called IndexGPT. All the patent previews we’ve seen from the company in recent months may give us a sneak peak at how this IndexGPT tool could work.
Chase is definitely bullish on AI. In CEO Jamie Dimon’s annual letter to shareholders, he noted that the company has more than 300 use cases for AI in practice. “AI and the raw material that feeds it, data, will be critical to our company’s future success — the importance of implementing new technologies simply cannot be overstated,” he said in the letter.
Of course, the company could license this tech to other banks, or even industries beyond wall street, such as law, media, or anyone that wants easy access to information. But internally, talent often represents one of the biggest costs for financial firms. Chase could save some significant costs by automating due diligence, research and other tedious grunt work that bogs down Wall Street.
We’re not done yet.
Zoom wants to know how you really felt about that last meeting. The company filed a patent application for a “sentiment scoring” system, which determines how you felt about a meeting by using machine learning to analyze transcripts.
Nvidia wants to get the time right. The company wants to patent a system for “fast clock detection” to find “anomalies” in clock signals.
PayPal wants to read your emojis. The company is seeking to patent a system for “emoji commanded action,” which identifies emojis and determines actions to take based on them.
What else is new?
The iPhone 15 is reportedly set to hit the shelves on September 22, around a week after an Apple event occurring on September 12 or 13, according to Bloomberg.
Tesla CFO Zachary Kirkhorn has stepped down after 13 years at the company. Vaibhav Taneja, the company’s chief accounting officer, has taken on Kirkhorn’s duties in addition to his own.
PayPal is rolling out a stablecoin called PayPal USD, making it the first large fintech to do so. The coin will be issued by Paxos Trust.