PATENT DROP: Nvidia reads between the lines
Discover more from PATENT DROP
PATENT DROP: Nvidia reads between the lines
Plus: Microsoft’s carbon capture; Intel lends an ear
Happy Thursday, and welcome back to Patent Drop! We hope you had a restful holiday and are ready to dive into some patents.
Today, we’re checking out a filing from Nvidia to help AI chatbots understand you a little bit better; a filing from Microsoft for carbon capture tech, and Intel’s plans for more affordable Bluetooth hearing aids.
Let's lift the curtain.
#1. Nvidia gets chatty
Nvidia wants its AI chatbots to be a little more understanding.
The company wants to patent a method for “determining intents and responses” with machine learning in conversational AI. Here’s how it works: the first machine learning model takes a passage of text, such as a query from a chatbot user, and boils down the first initial intent or meaning of the text. (FYI, an intent could be just about anything, including asking it to book a reservation, schedule an event or create a shopping list.)
That data then goes through another machine learning model to break down what it calls “sub-intents,” or additional meaning or context. This process is rinsed and repeated “in order to determine a final intent associated with the text,” creating a branching “tree-like structure” to associate intents with one another.
Deciphering intent behind text can be a difficult task for a machine learning model. If you train a model on a set of specific intents, as you would in an “intent-slot-based model,” you need a ton of training. For example, if you use this kind of model for a customer-service bot, it may only be able to answer a very narrow set of questions based on the number of intents it’s trained on.
Alternatively, you could use what’s called a “zero-shot model,” or one that essentially categorizes the intent of text based on a smaller amount of pre-trained categories. While this model requires less training to answer more questions, the result is often lower accuracy.
Nvidia’s system, however, circumvents both these issues by passing the data through machine learning models multiple times, until a final intent is uncovered.
Nvidia is without a doubt a heavyweight in the AI space. Much of its grip, however, is in hardware: The company controls 80% of the market for AI chips and GPUs, and is miles ahead of competitors like Intel and AMD.
But Nvidia isn’t new to the software or chatbot space, either. The company has long been working on conversational AI tools with a focus for industries like financial services, customer service and telecommunications.
In April, the company also announced new software called NeMo Guardrails, which claims to address the issue of large language models hallucinating incorrect answers. The company also has sought several patents for chatbot-related tech, including one that can utilize both open and closed domains to answer questions more accurately.
OpenAI, Google or Microsoft are often the first names that come to mind when you think of an AI chatbot. While these companies lead in gaining consumer attention, Nvidia has targeted the enterprise market, Romeo Alvarez, director and research analyst at William O'Neil. Since Nvidia is already the “gold standard” when it comes to hardware, the fact that it’s able to offer a neatly wrapped package for a variety of different enterprises to embed AI into their operations gives it a leg up, Alvarez noted.
Making that AI more intuitive, as this patent aims to do, will only draw in more business, Alvarez added.
“When it comes to more enterprise-oriented technology, where they don't have to compete with OpenAI, they do have the advantage because they have the complete stack of software and hardware,” said Alvarez. “It's a combination of the fact that they are working on both that gives them the advantage.”
#2. Microsoft sucks it up
Microsoft may be adding another piece to its green agenda.
The company filed a patent application for a “carbon capture system.” Microsoft’s system works specifically in conjunction with data centers, as they generate a great deal of exhaust heat from their cooling systems.
“Using waste heat to release captured carbon dioxide may help to reduce the carbon footprint (of) commercial operations,” Microsoft said. “For example, using the waste heat of a computing device or a plurality of computing devices may help to reduce the carbon footprint of a data center.”
To break it down, Microsoft’s tech consists of two “carbon capture plates,” the first of which collects carbon dioxide from “a flow of ambient air,” or outside, environmental air. The second plate releases that carbon dioxide once heat is applied from the exhaust vent of a data center’s cooling system. That carbon dioxide is emptied into a dedicated “release chamber" with a dedicated carbon storage tank, where it is compressed.
Microsoft said that this entire system is operated by a machine learning model, which monitors things like air temperature, computing load of the data center and amount of carbon captured, and “adjusts one or more operating parameter(s) to improve performance of the carbon capture system.”
Applying machine learning in this way could help reduce the costs of carbon capture, the company noted.
Microsoft has long been interested in carbon removal as part of its plan to reduce its carbon footprint. The company aims to go carbon-negative by 2030, and remove the equivalent of its historical emissions by 2050. In 2022, the company made a small dent in its emissions, cutting its output by 0.5% while keeping business growing. Melanie Nakagawa, the company’s chief sustainability officer, said in a LinkedIn post that “While this can be counted as progress, it’s not happening fast enough.”
Microsoft pairing carbon removal tech with its data centers adds up, too, said Dr. Dan Stein, founder and director of climate giving consultancy Giving Green. For one, data centers are major energy hogs, taking up 1,000 kWh per square meter, or 10 times that of the average American home. “A major part of large tech companies' carbon footprints is power consumption from data centers,” said Stein. With the 10 largest data center operators in the world reportedly owning a total of more than 1,250 facilities, if patented, Microsoft could potentially market this method to companies to help them cut down their own massive footprints (while making some cash in the process).
Microsoft also noted in its latest sustainability report that a lot of the carbon emissions created by its data centers come from “hard-to-abate” sources of materials like steel and concrete to construct the facilities themselves. Given data center-reliant cloud technology is a massive part of Microsoft’s business, finding a way to offset that could reduce its footprint even further.
As it stands, large-scale and long-lasting carbon capture and removal projects have yet to reach fruition, Stein said. These projects currently exist on a small scale, Stein noted, and Microsoft has also invested in a host of carbon removal companies as a part of its $1 billion Climate Innovation Fund. But the cost to actually remove that carbon from the atmosphere remains prohibitive, he said, costing roughly $500 per ton to remove, compared to between $15 and $50 to purchase a carbon credit.
“The big thing is the cost – everyone’s trying to make it cheaper,” said Stein. “It's easy to be pessimistic … but there's so many different possible solutions, you just need one of them to be really scalable to work.”
Companies like Microsoft throwing their weight behind carbon removal projects could lead to that solution, breaking them out of the tunnel vision of what Stein calls the “net zero paradigm” and creating something that benefits the world at large.
SPONSORED BY THE DONUT
Be Informed…Without Tearing Your Hair Out
An all too common dilemma: You want to be a good citizen who knows what’s going on in the world. But the news – biased, sensational, and packed with inscrutable jargon – just raises your blood pressure.
That’s where The DONUT comes in. They deliver news that’s unbiased, quick, engaging, easy to understand, and 100% free.
The DONUT believes that your time is the most precious resource you have. They respect that precious time by keeping you up to date – not to mention equipped with the insights you need to carry your next dinner party conversation – without draining your time or causing your anxiety to spike.
See for yourself and sign up for The DONUT today.
#3. Intel’s assistive hearing initiative
Intel wants to lend a helping hand … or a helping ear.
The company filed a patent application for a Bluetooth-capable “hearing aid system” that can be made more affordable than a conventional hearing aid. Rather than using a traditional hearing aid with Bluetooth capabilities, this system offers the alternative of using “lower cost earbuds as a terminal hearing device.”
These earbuds specifically work when connected to computers, tablets, smartphones and other devices. When connected, these earbuds essentially outsource the audio processing work to the device, allowing it to take over “a remarkable portion of the computational effort and audio adaptation” so the earbuds themselves don’t have to. This, in turn, allows the earbuds to be made at a lower cost, as they themselves don’t need to be highly capable of audio processing.
Intel noted in its patent filing that Bluetooth-enabled hearing aids can range from $3,000 to $5,000 each, and are “inaccessible to the majority of the global population experiencing degrees of hearing loss.” Meanwhile, the earbud system that Intel lays out could cost $200 or less, the company says.
The company also said its system aims to circumvent the communication barriers caused by “online communication and other audio-based computing tasks,” and amplified by remote work and school environments.
Along with benefitting the hard-of-hearing community, Intel said that this tech improves audio quality for the general population, upgrading things like speech recognition and clarity, noise canceling and feedback suppression.
This isn’t the first time we’ve seen Intel take an interest in hearing aids. In mid-May, the company announced several assistive tech-related projects, including plans to partner with hearing aid companies to enable direct Bluetooth connection between the aids and Intel Evo PCs. While Intel has a medical solutions sector, it’s far from the company’s bread and butter. Getting into hearing aids – and tying them to Intel devices specifically – could help the company in more ways than one.
In an interview with Forbes about the initiative, Darryl Adams, Intel’s director of accessibility, said that the company is “well-positioned to work with our industry partners to design future technologies that are more accessible to more people.”
“We are on a journey to have all user experience teams at Intel adopt inclusive design and research practices, along with building out the operational support for these inclusive practices,” Adams said in the interview.
While this patent claims to offer a cheaper alternative to the traditional hearing aid, the plan could come with a few hitches. For one, this tech is reliant on a Bluetooth connection, meaning that it works until it doesn’t. If the device it’s connected to malfunctions, or if that Bluetooth connection is lost, it could risk interruption for the user relying on it to hear.
Plus, the need to count on a smartphone, tablet or laptop for so-called low-cost hearing support can itself be prohibitive, as those devices themselves cost consumers a lot of money and are only getting more expensive.
Extra Drops
If you’re not full, here’s a few more bites before you go.
Google wants to make Airplane Mode less annoying. The company is seeking to patent a method for “activating a connected flight mode” which turns on “ before or during a flight to provide portable computing devices with granular levels of connectivity.”
Mastercard wants to let you pay with a mask on. The company wants to patent a method for payment authentication “while wearing a face covering” by embedding a “wireless transmitter chip associated with a unique identifier” into the user’s mask.
Block wants to get its hands on cold hard cash. The company is seeking to patent a “cash transaction machine” which allows users to withdraw cash from an ATM or a merchant’s point-of-sale using their phone as a virtual card.
What else is new?
AI wins could launch Microsoft into a market cap of $3 trillion, according to a recent model from Morgan Stanley analysts.
Twitter clones are on the rise: Meta’s Threads app hit 10 million sign ups within the first seven hours (despite issues launching in the EU due to privacy concerns); and Bluesky raised $8 million in seed funding.
SoftBank is backing Japanese robotics startup Telexistence, leading the company’s recent $170 million Series B funding round.
Have any comments, tips or suggestions? Drop us a line! Email at admin@patentdrop.xyz or shoot us a DM on Twitter @patentdrop.
 | A guest post by
|