Connect with us

Technology

Linq’s AI Retrieval Model Achieves the Top Spot on the HuggingFace MTEB Leaderboard

Published

on

BOSTON, June 5, 2024 /PRNewswire/ — Linq, a generative AI startup, announced that its large embedding model “Linq-Embed-Mistral” ranked first in the text retrieval evaluation on HuggingFace’s “Massive Text Embedding Benchmark (MTEB)” leaderboard, outpacing competitors like NVIDIA, Salesforce, Google, OpenAI, and Cohere. This evaluation is run by HuggingFace, the world’s largest machine learning platform.

Linq’s embedding model achieved a score of 60.2 points in the text retrieval category, securing the top position. This placed Linq ahead of NVIDIA, which scored 59.4 points, and Voyage AI, which scored 58.3 points. Google’s model followed with a score of 55.7, while OpenAI and Cohere scored 55.4 and 55.0 points, respectively.

The MTEB leaderboard by HuggingFace ranks the performance of embedding models across seven categories, including classification, clustering, pair classification, reranking, retrieval, semantic textual similarity (STS), and summarization. Linq’s embedding model demonstrated excellent performance not only in the text retrieval category but also in other categories, earning an overall rank of third.

The MTEB lists more than 300 embedding models, highlighting the competitive yet manageable landscape of embedding model technology. Linq’s top performance in this specific benchmark underscores its superiority in embedding model technology.

Embedding models are critical in generative AI, particularly for addressing the hallucination problem of large language models (LLMs) by employing retrieval-augmented generation (RAG) technology. RAG allows models to produce reliable outputs by accessing the latest data or internal documents not available within the LLM.

Leading this project, Dr. Junseong Kim stated, “Our research demonstrates that due to the broad topic diversity and challenging difficulty of retrieval data, GPT-generated data is not perfect and requires thorough verification and refinement. Through these processes, we can achieve quality comparable to human-labeled data, ultimately attaining the best retrieval performance based on the MTEB benchmark dataset. This study shows that through elaborate data crafting and filtering using GPT, we can create models optimized for retrieval-augmented generation (RAG) and maximize performance in specific fields.” Additionally, he emphasized, “Not only is refined data crucial, but optimized training methodologies and rapid experimental cycles are also key to maximizing retrieval performance.”

Linq’s Co-founder & CEO, Jacob Choi, emphasized, “Accurate search is crucial for generative AI enterprises’ adoption. We’re proud to have developed the core embedding model to achieve this, and we’ll keep expanding and refining it to ensure precise text searches in specialized fields like finance and legal.” Choi noted that while 2023 saw the rise of B2C use cases for generative AI with the advent of ChatGPT, 2024 will witness the growth of B2B (business-to-business) applications with improved accuracy and security technologies.

[Company Description]

Founded in 2022, Linq (Wecover Platforms Inc) was established by MIT Electrical and Computer Engineering graduate Jacob Choi and MIT Computational Science and Engineering Ph.D. Subeen Pang. In 2021, Choi was named in Forbes’ “30 Under 30” in the science category for his AI neuromorphic computing research. Linq received early investments from KakaoVentures, Smilegate Investment, and Yellowdog in 2022. In 2023, Linq won the Samsung Open Collaboration hosted by Samsung Financial Networks and was selected for MassChallenge Fintech cohort, the largest non-equity accelerator in the U.S., continuing its collaboration with KPMG US.

Contact: Jacob Choi (jacob.choi@getlinq.com)

View original content to download multimedia:https://www.prnewswire.com/apac/news-releases/linqs-ai-retrieval-model-achieves-the-top-spot-on-the-huggingface-mteb-leaderboard-302164494.html

SOURCE Linq (Wecover Platforms Inc)

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Technology

MEMS Drive: Paving the Way for Super-High Image Resolution

Published

on

By

TAIPEI, Dec. 26, 2024 /PRNewswire/ — Fabless chip firm MEMS Drive Hong Kong Ltd manufactures cutting-edge MEMS actuators targeted at mobile imaging in portable electronics. The company uses its proprietary MEMS design and process to allow CMOS sensors to achieve swift and precise ‘SensorShift’ and is the first semiconductor company to implement 5-axis stabilization in mobile cameras.

“With edge computing capabilities steadily advancing, MEMS Drive utilizes precise movement matching algorithms and artificial intelligence [AI] to derive multiple image-enhancing features besides just stabilization, such as super-resolution and improving identification capabilities,” says Colin Kwan, CEO and President of MEMS Drive during an interview with EE Times Asia. “MEMS SensorShift-driven image sensors overturn traditional voice coil motors [VCMs], and could be applied to mobile phones, sports cameras, wearable tech, surveillance, autonomous vehicles, robotic vision and other products, bringing unlimited possibilities to the future of imaging systems.”

It is for these reasons that the company won an Innovation Award and Best Sensor of the Year award for its piezo technology-enabled autofocus actuator and optical image stabilizer (OIS) actuator at EE Awards Asia 2024. Now in its fourth year, EE Awards Asia honors some of the best products, companies, and leaders making a difference every day in Asia’s electronics industry.

The MEMS autofocus (AF) actuator is a piezoelectric multi-morphic thin film actuator that can move at Z-axis and tip-tilt directions to achieve a 3-axis compensation. The MEMS devices are designed to be based on standard silicon fabrication processes, thereby enabling good scalability and ease of handling, testing, and packaging.

According to Colin, this sensor-based technology offers multiple advantages. “It provides 3-axis image adjustments as compared to 1-axis counterparts on lens-based AF systems; it also moves three times faster and 10 times more precise by detecting positions of less than a pixel, compared to 3-5µm on lens-based AF,” he explains. “And finally, by not using a motor to move the lens around, it consumes up to 50 times less power and avoids heating problems that could affect the image’s quality.”

Meanwhile, its OIS, known as SensorShift, is said to be the world’s smallest MEMS OIS actuator. It provides a 5-axis OIS solution to improve clarity and achieve stable and precise image.

“All vibrations in pitch, yaw, roll, translational X, and translational Y can be compensated with a single MEMS chip,” says Colin. “Due to its high precision and fast response in active movement, we can achieve another application called super resolution. By shifting the imager 1 pixel or half a pixel in different directions, capturing the image at each position, and then merging and processing those images, we can get a 4X or even higher-resolution image.”

Colin notes that compared to traditional lens-based OIS, MEMS Drive’s MEMS OIS is a sensor shift solution that provides superior stabilization performance with multiple advantages, including a 10ms fast response, just 4mW of power consumption, sub-pixel precision, and high image resolution in low light and video without blur.

“Powered by our patented electrical conductive flexure [ECF] technology, the MEMS Drive actuator provides consistent, durable performance,” says Colin. “With over 2.5 billion test cycles completed in continuous operation, these innovations are crucial for ensuring the highest standards in automotive applications and safety.”

MEMS Drive’s technology is suitable for smartphones, automotive, wearable technologies such as AR/VR camera systems, sports cameras, smart city applications, and robotic vision.

In this video interview, Colin also talks about his outlook for the semiconductor and electronics manufacturing industry over the next year, upcoming challenges amid the increasingly complex designs and architectures, and the new opportunities in the market.

For more information about MEMS Drive Hong Kong Limited and its innovative solutions, please visit: memsdrive.com
Youtube / Bilibili : MEMSDrive
LinkedIn: MEMS Drive

View original content to download multimedia:https://www.prnewswire.com/news-releases/mems-drive-paving-the-way-for-super-high-image-resolution-302338882.html

SOURCE EE Times Asia

Continue Reading

Technology

Nodepay Raises $7M Total Funding To Power AI Growth with Real-Time Data Infrastructure

Published

on

By

SINGAPORE, Dec. 27, 2024 /PRNewswire/ — Nodepay, a decentralized AI platform transforming unused internet bandwidth into real-time data pipelines for AI training, today announced it has raised a second round of funding, bringing its total to $7 million.

The latest funding round welcomed new strategic investors IDG Capital ($23 Billion AUM), Mythos, Elevate Ventures, IBC, Optic Capital, Funders.VC, Matthew Tan (Etherscan founder) and Yusho Liu (CoinHako Co-founder & CEO) as notable angels. They join an impressive roster of previous backers that includes Animoca Brands, Mirana, OKX Ventures, JUMP Crypto, Tokenbay Capital and more.

Nodepay’s network taps into a global community of users running privacy-protected nodes. By sharing their spare internet bandwidth, these participants earn rewards for creating a real-time data source that improves AI inference with accurate, timely information—an approach known as Retrieval Augmented Generation (RAG).

Darren Nguyen, co-founder of Nodepay commented: “Our mission is to develop solutions that create tangible value for both AI developers and its end users. We give contributors a share in the AI ecosystem they help fundamentally build.”

Nodepay’s infrastructure platform integrates real-time data retrieval, a Web3-focused decentralized answer engine, reinforcement learning for more accurate model output, and gamified human verification. Together, these components combine to create a fair, collaborative, and innovative AI ecosystem.

Eric Le, investment director of IDG Capital, said, “The team at Nodepay is democratizing the AI economy by providing a platform that allows users to share directly in the value they create. We’re proud to support their vision of making AI more accessible and beneficial to all.”

With this funding, Nodepay will continue to commercialize its infrastructure to benefit both its community and partner AI labs. As it prepares to launch on Solana, Nodepay stands ready to lead the next era of decentralized AI development and training.

Already serving over 1.5 million active users worldwide, Nodepay continues to expand its reach, solidifying its role as a leader in the integration of AI and blockchain technology. Users can expect further updates and new announcements through their social channels and official website.

About Nodepay
Nodepay is a decentralized AI platform dedicated to democratizing AI training through real-time data retrieval. By turning idle internet bandwidth into a valuable resource, Nodepay fuels the next generation of AI models and stands at the forefront of AI decentralization.

Photo – https://mma.prnewswire.com/media/2587832/7M_Raised_Funds.jpg

Logo – https://mma.prnewswire.com/media/2587833/1080x1080_Black___Nodepay_2_Logo.jpg

View original content:https://www.prnewswire.co.uk/news-releases/nodepay-raises-7m-total-funding-to-power-ai-growth-with-real-time-data-infrastructure-302339478.html

Continue Reading

Technology

Plume Network Partners with Maseer to Tokenize $200M of Carbon Allowances

Published

on

By

NEW YORK, Dec. 26, 2024 /PRNewswire/ — Plume Network is proud to announce a strategic partnership with Maseer, an Abu Dhabi based tokenization platform, to bring $200M in Carbon Allowances exclusively on-chain to Plume. Built on Plume’s Real-World Asset Finance (RWAfi) ecosystem, Maseer will offer a tokenized solution to one of the fastest-growing alternative asset classes: compliance carbon.

Empowering Climate Action Through Compliance Carbon Tokenization

Compliance carbon has been one of the fastest growing alternative asset classes given increased regulatory and business scrutiny on emissions. The S&P Global Carbon Credit Index, which tracks the most liquid segment of the tradable carbon credit futures markets, has seen a 15.68% annualized return over the past five years. The value of these markets reached nearly one trillion USD in 2023.

The partnership with Plume Network allows Maseer to bring fully collateralized carbon products on-chain, where they will be fully compatible with Web3’s potent DeFi sector. DeFi integration vastly enhances compliance carbon markets with superior liquidity solutions and greater access to a global body of investors, broader market demand, and new yield sources.

“We are excited to partner with Plume to bring carbon allowances on chain. Plume is uniquely positioned to bring this vision to fruition because they are the only chain purpose built for RWAs. They’ve raised the bar with their tokenization engine, infrastructure tooling, and ecosystem network effects. We believe Plume is on the bleeding edge of on-chain adoption of RWAs,” said Bradley Allgood, CEO of Maseer.

“Energy transition is an asset category that we have been increasingly focused on at Plume because of growing demand for climate action, both from a government and corporate sustainability perspective. Volumes for the global carbon credit market are forecasted to grow at a 39% CAGR from 2024 to 2033,” said Teddy Pornprinya, Chief Business Officer and Co-Founder at Plume Network.

What are carbon allowances?

Compliance carbon allowances trade under cap-and-trade programs known as Emissions Trading Systems (ETS). These systems create transparent, liquid markets that are government-mandated and regulated. As of April 1, 2024, approximately 18 percent of global greenhouse gas emissions are covered by emissions trading systems (ETS). Carbon allowances are distinct from project-based carbon offsets and offer a market-based approach to regulating a region’s emissions, with mandatory participation for specified industries. Carbon allowance supply is managed by government agencies and adjusted primarily through an annually declining cap. 

About Plume
Plume is the first fully integrated L1 modular blockchain focused on RWAfi, offering a composable, EVM-compatible environment for onboarding and managing diverse real-world assets. With 180+ projects on its private devnet, Plume provides an end-to-end tokenization engine and a network of financial infrastructure partners, simplifying asset onboarding and enabling seamless DeFi integration for RWAs. Learn more at https://www.plumenetwork.xyz/ or contact press@plumenetwork.xyz

About Maseer
Maseer operates out of Abu Dhabi Global Market (ADGM), the world’s leading Special Economic Zone (SEZ) for digital asset innovation. Maseer is led by Tokenization and Free Zone Veteran Bradley Allgood and is focused on the design of bringing real world assets on chain to be fully interoperable with DeFi. Maseer has developed strategic relationships with Sovereign Nations and Large Enterprises to identify the highest quality real world assets around the world.

View original content to download multimedia:https://www.prnewswire.com/news-releases/plume-network-partners-with-maseer-to-tokenize-200m-of-carbon-allowances-302339461.html

SOURCE Plume Network

Continue Reading

Trending