Discover more from PATENT DROP
PATENT DROP: Baidu powers up
Plus: Goldman breaks down data; PayPal gets to know you
Happy Thursday and welcome to Patent Drop!
Today, we’re digging into Baidu’s EV charging goals; Goldman Sachs’ plan to predict the future for traders; and PayPal’s patent to understand chatbot users better ... that’s quite similar to a recent Nvidia filing.
Before we dive into today’s tech, we have a favor to ask. Our goal here at Patent Drop is pretty simple — we want to provide unique perspective into the future of tech and the novel strategies reshaping the landscape. To enable us to provide the most razor sharp analysis we can, and serve you, our reader — we'd so appreciate if you could answer just a few quick questions. In just 5 minutes you can shape the future of Patent Drop. Thank you so much in advance for your help, we really appreciate it.
Anyways, let’s go.
#1. Baidu’s battery saver
Electric vehicles may not be the first thing that comes to mind when you think of Baidu, but the Chinese search giant wants to make sure you stay charged.
The company is seeking to patent a system for “recommending a charging station” and determining the wait time at that station. Baidu’s system isn’t terribly complicated, essentially using your location data to recommend the most optimal station.
Using a pre-trained recommendation AI model, Baidu’s system takes into account all surrounding stations, your estimated arrival time at any given station, and each station’s “charge pile” — the wait-time estimate calculated by the number of vehicles ahead of you with reservations and the estimated time it’ll take them to charge.
With that information, it determines the most ideal station in terms of distance, wait times and arrival times. Baidu claims its system cuts down “unnecessary queuing,” so that the “time cost of vehicle charging is decreased, and traveling experience of the owner of the electric vehicle is optimized.”
China dominates the electric vehicle market. EVs accounted for a quarter of all cars sold in China in 2022, Bloomberg reported this week, and clean car sales in the country — which includes plug-in hybrids — hit 5.6 million. HSBC expects EV penetration to reach 90% in the country by 2030. (EVs made up around 5.7% of all new US car sales in 2022).
Solid charging infrastructure plays a major role in EV popularity, said Matt McCaffree, VP of utility market development at clean energy parking and mobility company Flash. According to Bloomberg’s report, China had more than 6 million EV chargers at the end of May, the most of any country globally. Plus, charging standards are uniform, so anyone can plug in at any station.
Though demand for EVs in China outpaced that of all vehicles, the country has the “biggest growth market for vehicles, period, ” McCaffree said. “This is the perfect environment and the ideal set of circumstances for expansion.”
Companies like BYD, Tesla and SAIC-GM-Wuling lead in the Chinese auto market. While Baidu has been plugging away at EVs for a while, touting both self-driving and conversational AI capabilities in its cars, the company is a relatively small fish in China’s EV pond. By offering a service that works in tandem with EVs, Baidu may be seeking to capitalize on the country’s rapidly growing market and boost consumer confidence in EVs, said McCaffree.
“It's a matter of taking advantage of growth in the market and growth in popularity, and recognizing the need for drivers to have visibility into the charging network and what's available,” said McCaffree. “There is a recognized need to create consumer confidence in charging.”
However, it’s unclear whether Baidu will actually be able to obtain this patent, said McCaffree. Similar tech already exists on the market, he noted. Baidu may have a harder time proving that its tech is novel and unique from others in order to claim it as it’s own.
#2. Goldman’s data breakdown
Some financial decisions can be hard. Goldman Sachs wants to lessen some of the guesswork for traders.
The financial institution is seeking to patent an “information-dense user interface” to visualize asset prices and transaction parameters. To put it simply, Goldman’s tech essentially aggregates everything a trader needs to know in one place to help predict stock prices, as “existing systems make it very difficult for a trader to understand the universe of scenarios or place those scenarios in historical context.”
Goldman’s user interface comes with a host of different displays, including:
A polling tool to predict asset prices;
Historical data based on past predictions;
An asset ordering system that provides “intra-trade updates;”
And an interface that allows users to define the “parameters,” such as spot price, timeframe or expiration date, for a transaction based on their predicted prices.
This interface also shows how often an individual user’s price prediction is correct, and uses machine learning to figure out how likely they are to be correct in the future, assigning “kudos” to the user that predicts the most correctly. “Higher kudos corresponds to more accurate historical predictions and vice versa,” Goldman noted.
The company said that in conventional systems, relevant information is spread across several spreadsheets, apps and databases that a trader must flip through, “if it is available at all.” Using those systems requires a high level of skill and experience, Goldman said, and “has a high propensity for human errors resulting from the trader not correctly drawing correlations between disparate pieces of data in the spreadsheets.”
Patent tech like this is par for the course for Goldman: Most of the firm’s patent filings relate to financial data and how to manage it. Plus, given that this tech could likely help Goldman continue to make money faster, there's certainly incentive to implement it among their own workflow.
With the skyrocketing demand for financial data, creating systems that help traders meaningfully organize it is itself a lucrative business. Along with using a system like this internally, Goldman could license a system like this to other firms. But if the system could give the firm an advantage over competitors, keeping it for itself may be a better option, said Dean Kim, head of equity research at William O’Neil.
Internally, tech like this could provide a major benefit, making trading more efficient, helping make better decisions and limiting the chance for human error, said Kim. “There's clients behind every trade,” he said. “So if a trader at Goldman makes a mistake, then at the end of the day, they're going to be on the hook in terms of making the client whole.”
Tech like this can also save on labor costs, said Kim. Since conventional trading systems require more expertise and time to operate, a simpler, AI-backed system could create opportunities for less-skilled traders to join the big leagues.
Trading human labor for AI is a controversial issue that’s sprung up among the tech’s critics. But Goldman faced bleak earnings yesterday morning, and has several rounds of layoffs from the past year under its belt. The company may see AI-based tech as, at least, a helping hand.
“The number one cost for Goldman Sachs is people,” said Kim. “If you can reduce the number of people that are trading, all of that is going to fall to the bottom line.”
#3. A Tale of Two Chatbots
PayPal wants to help its chatbots think critically.
The company wants to patent a system for “predicting and providing automated online chat assistance.” Basically, PayPal’s system uses two AI models to determine what a user is trying to communicate in an online customer service chat scenario.
Because different people use different language, phrases or short-form abbreviations, an AI model determining intent can be challenging, PayPal noted. Misspellings of words and typographical errors also tend to trip up AI models. This is why PayPal employs two models, each of which are trained to analyze different things.
The first model may be originally trained on a “generic corpus” for training a natural language processing model, PayPal noted. The second model, meanwhile, is trained using a “bidirectional encoder representations from transformations,” or BERT model, which uses deep learning to analyze phrases as a whole, rather than individual words, to derive context.
“The second model may be able to process a phrase correctly even though the phrase is not a complete sentence, or includes inaccurate and/or missing words,” PayPal said in its filing.
In practice, a user's chat message is sent to a prediction machine learning model to determine the user’s intent. If the first prediction model is unable to figure it out, the message is sent to a second prediction model to understand the intent. The results of this are then used to respond to the user. After the fact, the interaction is used to re-train the first prediction model to better understand intents the first time around.
If it feels like déjà vu, it is: Nvidia is trying to patent tech that’s used for practically the same purpose. Nvidia’s patent details a system for “determining intents and responses” in a conversational AI scenario, which relies on multiple AI models to decipher meaning from passages of text.
While both Nvidia and PayPal are leveraging multi-model approaches to understand a user’s intent, they are going about it in different ways, said Micah Drayton, partner and chair of the technology practice group at Caldwell Intellectual Property Law.
PayPal’s system is essentially using its second model as a backup — If the system isn’t confident in the output of the first model, it’ll scrap that, and pass the job of determining intent onto the second one. Nvidia, meanwhile, is using one AI model to break down the initial meaning of a passage of text, and the second model to break down any “sub-intents,” or additional context. This means that the output of both AI models is taken into consideration in its response.
“They're taking two slightly different approaches, but both of their approaches are basically doing the same thing, with specifically figuring out that intent,” said Drayton.
Each companies’ patents likely display just a single part of a larger system, which also likely differ in various ways, Drayton noted. Plus, when a company is patenting software, it’s less about what they’re trying to solve, and more about how they’re trying to solve it, he said.
“It’s theoretically possible that a patent examiner would say that (Nvidia’s first claim) was very similar to concepts displayed in the PayPal application. The way it's written it's quite broad,” said Drayton. “But there is a difference between these inventions.”
Patent law aside, the tech firms would also likely utilize these tools in different ways, too. While the tech Nvidia’s patent would probably be licensed and sold to other businesses, PayPal would likely use its tech in-house to power customer service for its financial apps.
We’re not done yet.
Zoom wants to give you a camera-off option. The company is seeking to patent a method for “avatar generation” in a video platform, creating an avatar that mimes your facial expressions over video.
Microsoft wants to guide you through new places. The company is seeking to patent a system for “geospatially referenced building floor plan data,” which essentially maps out floor plans in buildings.
Salesforce doesn’t want you to get Slack FOMO. The company wants to patent a system for “sharing channels” on a communication platform, essentially allowing users to add someone to several channels at once to eliminate users needing to “hunt for the channels” themselves.
What else is new?
Google is reportedly testing an AI tool that can write news articles. The tool, internally called “Genesis,” has been pitched to The New York Times, The Washington Post and News Corp.
Amazon will add its pay-by-palm technology to all Whole Foods by the end of the year. The biometric scanning system, called Amazon One, was first introduced in 2020.
TSMC reported a revenue decline of 10%, its first profit drop in four years, as macroeconomic headwinds “dampened the end market demand.”