Connect with us


AI armed with multiple senses could gain more flexible intelligence



In late 2012, AI scientists first figured out how to get neural networks to “see.” They proved that software designed to loosely mimic the human brain could dramatically improve existing computer-vision systems. The field has since learned how to get neural networks to imitate the way we reason, hear, speak, and write.

But while AI has grown remarkably human-like—even superhuman—at achieving a specific task, it still doesn’t capture the flexibility of the human brain. We can learn skills in one context and apply them to another. By contrast, though DeepMind’s game-playing algorithm AlphaGo can beat the world’s best Go masters, it can’t extend that strategy beyond the board. Deep-learning algorithms, in other words, are masters at picking up patterns, but they cannot understand and adapt to a changing world.

Researchers have many hypotheses about how this problem might be overcome, but one in particular has gained traction. Children learn about the world by sensing and talking about it. The combination seems key. As kids begin to associate words with sights, sounds, and other sensory information, they are able to describe more and more complicated phenomena and dynamics, tease apart what is causal from what reflects only correlation, and construct a sophisticated model of the world. That model then helps them navigate unfamiliar environments and put new knowledge and experiences in context.

AI systems, on the other hand, are built to do only one of these things at a time. Computer-vision and audio-recognition algorithms can sense things but cannot use language to describe them. A natural-­language model can manipulate words, but the words are detached from any sensory reality. If senses and language were combined to give an AI a more human-like way to gather and process new information, could it finally develop something like an understanding of the world?

The hope is that these “multimodal” systems, with access to both the sensory and linguistic “modes” of human intelligence, should give rise to a more robust kind of AI that can adapt more easily to new situations or problems. Such algorithms could then help us tackle more complex problems, or be ported into robots that can communicate and collaborate with us in our daily life.

New advances in language-­processing algorithms like OpenAI’s GPT-3 have helped. Researchers now understand how to replicate language manipulation well enough to make combining it with sensing capabilities more potentially fruitful. To start with, they are using the very first sensing capability the field achieved: computer vision. The results are simple bimodal models, or visual-language AI.

In the past year, there have been several exciting results in this area. In September, researchers at the Allen Institute for Artificial Intelligence, AI2, created a model that can generate an image from a text caption, demonstrating the algorithm’s ability to associate words with visual information. In November, researchers at the University of North Carolina, Chapel Hill, developed a method that incorporates images into existing language models, which boosted the models’ reading comprehension.

OpenAI then used these ideas to extend GPT-3. At the start of 2021, the lab released two visual-language models. One links the objects in an image to the words that describe them in a caption. The other generates images based on a combination of the concepts it has learned. You can prompt it, for example, to produce “a painting of a capybara sitting in a field at sunrise.” Though it may have never seen this before, it can mix and match what it knows of paintings, capybaras, fields, and sunrises to dream up dozens of examples.

Achieving more flexible intelligence wouldn’t just unlock new AI applications: it would make them safer, too.

More sophisticated multimodal systems will also make possible more advanced robotic assistants (think robot butlers, not just Alexa). The current generation of AI-powered robots primarily use visual data to navigate and interact with their surroundings. That’s good for completing simple tasks in constrained environments, like fulfilling orders in a warehouse. But labs like AI2 are working to add language and incorporate more sensory inputs, like audio and tactile data, so the machines can understand commands and perform more complex operations, like opening a door when someone is knocking.

In the long run, multimodal breakthroughs could help overcome some of AI’s biggest limitations. Experts argue, for example, that its inability to understand the world is also why it can easily fail or be tricked. (An image can be altered in a way that’s imperceptible to humans but makes an AI identify it as something completely different.) Achieving more flexible intelligence wouldn’t just unlock new AI applications: it would make them safer, too. Algorithms that screen résumés wouldn’t treat irrelevant characteristics like gender and race as signs of ability. Self-driving cars wouldn’t lose their bearings in unfamiliar surroundings and crash in the dark or in snowy weather. Multimodal systems might become the first AIs we can really trust with our lives.

Lyron Foster is a Hawaii based African American Musician, Author, Actor, Blogger, Filmmaker, Philanthropist and Multinational Serial Tech Entrepreneur.

Continue Reading


Vietnamese electric motorbike startup Dat Bike raises $2.6M led by Jungle Ventures



Son Nguyen, founder and chief executive officer of Dat Bike on one of the startup's motorbikes

Son Nguyen, founder and chief executive officer of Dat Bike

Dat Bike, a Vietnamese startup with ambitions to become the top electric motorbike company in Southeast Asia, has raised $2.6 million in pre-Series A funding led by Jungle Ventures. Made in Vietnam with mostly domestic parts, Dat Bike’s selling point is its ability to compete with gas motorbikes in terms of pricing and performance. Its new funding is the first time Jungle Ventures has invested in the mobility sector and included participation from Wavemaker Partners, Hustle Fund and iSeed Ventures.

Founder and chief executive officer Son Nguyen began learning how to build bikes from scrap parts while working as a software engineer in Silicon Valley. In 2018, he moved back to Vietnam and launched Dat Bike. More than 80% of households in Indonesia, Malaysia, Thailand and Vietnam own two-wheeled vehicles, but the majority are fueled by gas. Nguyen told TechCrunch that many people want to switch to electric motorbikes, but a major obstacle is performance.

Nguyen said that Dat Bike offers three times the performance (5 kW versus 1.5 kW) and 2 times the range (100 km versus 50 km) of most electric motorbikes in the market, at the same price point. The company’s flagship motorbike, called Weaver, was created to compete against gas motorbikes. It seats two people, which Nguyen noted is an important selling point in Southeast Asian countries, and has a 5000W motor that accelerates from 0 to 50 km per hour in three seconds. The Weaver can be fully charged at a standard electric outlet in about three hours, and reach up to 100 km on one charge (the motorbike’s next iteration will go up to 200 km on one charge).

Dat Bike’s opened its first physical store in Ho Chi Minh City last December. Nguyen said the company “has shipped a few hundred motorbikes so far and still have a backlog of orders.” He added that it saw a 35% month-over-month growth in new orders after the Ho Chi Minh City store opened.

At 39.9 million dong, or about $1,700 USD, Weaver’s pricing is also comparable to the median price of gas motorbikes. Dat Bike partners with banks and financial institutions to offer consumers twelve-month payment plans with no interest.

“These guys are competing with each other to put the emerging middle class of Vietnam on the digital financial market for the first time ever and as a result, we get a very favorable rate,” he said.

While Vietnam’s government hasn’t implemented subsidies for electric motorbikes yet, the Ministry of Transportation has proposed new regulations mandating electric infrastructure at parking lots and bike stations, which Nguyen said will increase the adoption of electric vehicles. Other Vietnamese companies making electric two-wheeled vehicles include VinFast and PEGA.

One of Dat Bike’s advantages is that its bikes are developed in house, with locally-sourced parts. Nguyen said the benefits of manufacturing in Vietnam, instead of sourcing from China and other countries, include streamlined logistics and a more efficient supply chain, since most of Dat Bike’s suppliers are also domestic.

“There are also huge tax advantages for being local, as import tax for bikes is 45% and for bike parts ranging from 15% to 30%,” said Nguyen. “Trade within Southeast Asia is tariff-free though, which means that we have a competitive advantage to expand to the region, compare to foreign imported bikes.”

Dat Bike plans to expand by building its supply chain in Southeast Asia over the next two to three years, with the help of investors like Jungle Ventures.

In a statement, Jungle Ventures founding partner Amit Anand said, “The $25 billion two-wheeler industry in Southeast Asia in particular is ripe for reaping benefits of new developments in electric vehicles and automation. We believe that Dat Bike will lead this charge and create a new benchmark not just in the region but potentially globally for what the next generation of two-wheeler electric vehicles will look and perform like.”

Continue Reading


Binance Labs leads $1.6M seed round in DeFi startup MOUND, the developer of Pancake Bunny



Decentralized finance startup MOUND, known for its yield farming aggregator Pancake Bunny, has raised $1.6 million in seed funding led by Binance Labs. Other participants included IDEO CoLab, SparkLabs Korea and Handshake co-founder Andrew Lee.

Built on Binance Smart Chain, a blockchain for developing high-performance DeFi apps, MOUND says Pancake Bunny now has over 30,000 daily average users, and has accumulated more than $2.1 billion in total value locked (TVL) since its launch in December 2020.

The new funding will be used to expand Pancake Bunny and develop new products. MOUND recently launched Smart Vaults and plans to unveil Cross-Chain Collateralization in about a month, bringing the startup closer to its goal of covering a wide range of DeFi use cases, including farming, lending and swapping.

Smart Vaults are for farming single asset yields on leveraged lending products. It also automatically checks if the cost of leveraging may be more than anticipated returns and can actively lend assets for MOUND’s cross-chain farming.

Cross-Chain Collateralization is cross-chain yield farming that lets users keep original assets on their native blockchain instead of relying on a bridge token. The user’s original assets serve as collateral when the Bunny protocol borrows assets on the Binance Smart Chain for yield farming. This allows users to keep assets on native blockchains while giving them liquidity to generate returns on the Binance Smart Chain.

In statement, Wei Zhou, Binance chief financial officer, and head of Binance Labs and M&A’s, said “Pancake Bunny’s growth and MOUND’s commitent to execution are impressive. Team MOUND’s expertise in live product design and servie was a key factor in our decision to invest. We look forward to expanding the horizons of Defi together with MOUND.”

Continue Reading


Battery Resourcers raises $20M to commercialize its recycling-plus-manufacturing operations



As a greater share of the transportation market becomes electrified, companies have started to grapple with how to dispose of the thousands of tons of used electric vehicle batteries that are expected to come off the roads by the end of the decade.

Battery Resourcers proposes a seemingly simple solution: recycle them. But the company doesn’t stop there. It’s engineered a “closed loop” process to turn that recycled material into nickel-manganese-cobalt cathodes to sell back to battery manufacturers. It is also developing a process to recover and purify graphite, a material used in anodes, to battery-grade.

Battery Resourcers’ business model has attracted another round of investor attention, this time with a $20 million Series B equity round led by Orbia Ventures, with injections from At One Ventures, TDK Ventures, TRUMPF Venture, Doral Energy-Tech Ventures and InMotion Ventures. Battery Resourcers CEO Mike O’Kronley declined to disclose the company’s new valuation.

The cathode and anode, along with the electrolyzer, are major components of battery architecture, and O’Kronley told TechCrunch it is this recycling-plus-manufacturing process that distinguishes the company from other recyclers.

“When we say that we’re on the verge of revolutionizing this industry, what we are doing is we are making the cathode active material — we’re not just recovering the metals that are in the battery, which a lot of other recyclers are doing,” he said. “We’re recovering those materials, and formulating brand new cathode active material, and also recovering and purifying the graphite active material. So those two active materials will be sold to a battery manufacturer and go right back into the new battery.”

“Other recycling companies, they’re focused on recovering just the metals that are in [batteries]: there’s copper, there’s aluminum, there’s nickel, there’s cobalt. They’re focused on recovering those metals and selling them back as commodities into whatever industry needs those metals,” he added. “And they may or may not go back into a battery.”

The company says its approach could reduce the battery industry’s reliance on mined metals — a reliance that’s only anticipated to grow in the coming decades. A study published last December found that demand for cobalt could increase by a factor of 17 and nickel by a factor of 28, depending on the size of EV uptake and advances in battery chemistries.

Thus far, the company’s been operating a demonstration-scale facility in Worcester, Massachusetts, and has expanded into a facility in Novi, Michigan, where it does analytical testing and material characterization. Between the two sites, the company can make around 15 tons of cathode materials a year. This latest funding round will help facilitate the development of a commercial-scale facility, which Battery Resourcers said in a statement will boost its capacity to process 10,000 tons of batteries per year, or batteries from around 20,000 EVs.

Another major piece of its proprietary recycling process is the ability to take in both old and new EV batteries, process them and formulate the newest kind of cathodes used in today’s batteries. “So they can take in 10-year-old batteries from a Chevy Volt and reformulate the metals to make the high-Ni cathode active materials in use today,” a company spokesman explained to TechCrunch.

Battery Resourcers is already receiving inquiries from automakers and consumer electronics companies, O’Kronley said, though he did not provide additional details. But InMotion Ventures, the venture capital arm of Jaguar Land Rover, said in a statement its participation in the round as a “significant investment.”

“[Battery Resourcers’] proprietary end-to-end recycling process supports Jaguar Land Rover’s journey to become a net zero carbon business by 2039,” InMotion managing director Sebastian Peck said.

Battery Resourcers was founded in 2015 after being spun out from Massachusetts’ Worcester Polytechnic Institute. The company has previously received support from the National Science Foundation and the U.S. Advanced Battery Consortium, a collaboration between General Motors, Ford Motor Company and Fiat Chrysler Automobiles.

Continue Reading