Blockchain and IT development company Yudiz announced that Chief Financial Officer Zarna Shah will resign on January 31st. Shah said that his resignation is due to personal commitments, and he will also resign from the position of key personnel responsible for determining the disclosure of major events in the securities exchange. Founded in 2011, Yudiz focuses on developing web and mobile applications as well as emerging technologies, including AR/VR, AI/ML, blockchain, and IoT. During the first half of the 2025 fiscal year, the company reported a standalone net profit of 58,00 US dollars, far below last year's 155,433 US dollars. (TechinAsia)

coin This article explores how the convergence of machine learning (ML) and DeFi (Decentralized Finance) is changing the field of Financial Service and improving prediction accuracy. ML powers Decentralized Finance by analyzing, forecasting, and automating financial decisions. However, challenges such as data privacy, model transparency, and regulatory compliance still need to be addressed. Future blogs will focus on the potential and pitfalls of training ML models on Blockchain.

Odaily Planet Daily News AlloraNetwork, a decentralized artificial intelligence network, officially posted on the X platform that it will soon launch the Allora points program, which is mainly aimed at ML model creators and developers who run models on the network, create useful topics, and build applications using Allora's collective intelligence. According to the official profile, AlloraNetwork is mainly supported by institutions such as Polychain, Framework, Blockchain Capital, and others.

Ritual, a decentralized AI computing platform, has announced that it has open-sourced infernet-ml, a toolkit for building machine learning workflows on Infernet.

According to official information, Banyan, a Filecoin ecological project, has reached a strategic cooperation with io.net. The partnership aims to further promote decentralization and affordable alternatives to existing centralized cloud service providers by providing GPU network access to Banyan's enterprise-grade file storage and io.net's AI/ML.

Golden Finance reported that Tom Cat, a listed company, issued an announcement this evening, announcing that its wholly-owned subsidiary, Outfit7 Limited, has signed the "Agreement on investment into alternative investment fund SILICON GARDENS FUND III", that is, on the equity investment fund SILICON GARDENS FUND III (referred to as " SG Fund"), Outfit7 Limited, as an ordinary investor, invested in SG Fund with an amount of 1 million euros, and the investment area mainly includes investing in the equity of regional technology start-ups in neighboring countries such as Slovenia, Croatia, Serbia, etc., covering the software industry, SaaS, market, digitalization, AI/ML, big data, Web3 and other global trend industries.

On December 19, Polygon Labs announced that the ZK oracle network Hyper Oracle is integrating its zkOracle protocol with Polygon CDK to support new applications from fully decentralized stablecoins to AI-powered DApps. The zkOracle protocol provides a programmable oracle solution that facilitates verifiable computation, access to historical on-chain data, and AI inference, enabling developers to achieve trustless automation and on-chain AI/ML.

DeFi and on-chain ML portal HyperOracle has announced that it is integrating with Polygon CDK to provide computing power to Polygon builders, Golden Finance reported. The latest zkOracle integration allows builders to extend EVM capabilities to enable on-chain AI and advanced computing using the Polygon CDK. In addition, HyperOracle is working with Delphinus Lab, the first open-source zkWASM virtual machine, to develop a zkWASM prover.

According to IT Home on December 6, the 2023 ML Insider survey conducted by cnvrg.io, a subsidiary of Intel Corporation that focuses on AI and large language model (LLM) platforms, showed that only 10% of enterprise organizations have integrated generative AI (GenAI) solutions into production in the past year. The survey also shows that about half of companies that have adopted generative AI report improvements in customer experience (58%), efficiency (53%), product capabilities (52%), and significant cost reductions (47%). Despite the obvious benefits, the path to generative AI adoption is not without its challenges. According to the survey, 46% of respondents cited infrastructure as the biggest barrier to developing LLMs into products, along with other barriers including complexity, lack of AI talent, high implementation costs, and compliance issues. Eighty-four percent of respondents said they needed to upskill themselves due to their growing interest in LLM technology, and only 19 percent were confident in their understanding of the mechanics of this technology.

PANews reported on December 1 that according to Nikkei Asia, Vietnamese venture capital firm FEBE Ventures announced the launch of a second fund with a scale of $75 million, which was supported by international family office Otium Capital and established a joint investment partnership with Tekton Ventures. In addition, FEBE Ventures announced that Aditya Pendyala will join as a partner with expertise in artificial intelligence (AI)/machine learning (ML), enterprise software-as-a-service (SaaS), Internet of Things (IoT) and blockchain. At present, FEBE Ventures has made a number of investments in the Web3 field, and its portfolio includes: Atato, a Singapore-based licensed digital asset custody solution provider, and SeMorpheus, a Web3 branded retail platform.

PANews reported on November 26 that according to VentureBeat, Seattle-based distributed computing platform Expanso closed a $7.5 million seed round led by General Catalyst and Hetz Ventures. The funds will be used to accelerate the development of the data processing platform "Bacalhau". Currently, Bacalhau can handle a range of data tasks, from cleaning and processing application logs from the source, running distributed ML training across remote devices, and more.

ARM today announced the Cortex-M52 core, designed for AIoT applications, bringing performance gains in DSP and machine learning (ML) while reducing costs, IT Home reported. ARM expects partners to launch products in 2024. According to reports, Cortex-M52 will promote the popularization of ML functions of MCUs through advantages over the current market price, bringing AI computing power into lower-cost and lower-power devices. The Cortex-M52 delivers up to 5.6x faster ML performance and up to 2.7x faster DSP than previous generations of the Armv8.1-M architecture-based Cortex-M family of products. It is reported that the Cortex-M52 was previously known as the "Xingchen" STAR-MC2 in the Chinese market, and is a collaborative product between ARM and ARM's R&D engineering team.

PANews reported on November 23 that according to Alexablockchain, Hyper Oracle, a ZK oracle network supported by Sequoia China, announced the launch of Optimistic Machine Learning "opML", an open-source solution for large machine learning models , which aims to change the way large machine learning (ML) models run on the Ethereum blockchain, which means that users can run large language models on basic hardware such as laptops, without the need for high-end resources such as GPUs, and with higher performance and flexibility compared to zkML (zero-knowledge machine learning). It is reported that opML supports the current mainstream large-scale machine learning models on the Ethereum mainnet, including complex models like GPT 3.5, and also supports trustless computing, and the Optimism-based verification mechanism can ensure the decentralization and verifiability of computing during AI and machine learning training. According to previous news, Hyper Oracle completed a $3 million Pre-Seed round of financing this year, co-led by dao5 and Sequoia China Seed Fund, with participation from FutureMoney Group and others.

According to BusinessWire on November 20, Indian engineering services company L&T announced a partnership with NVIDIA to develop a software-defined architecture for medical devices focused on endoscopy, thereby improving the image quality and scalability of products. It is reported that the architecture was developed from the ground up and has an image processing pipeline that can be used for noise reduction, edge and contrast enhancement, texture and color enhancement, and speckle correction. In addition, the architecture incorporates AI/ML models developed to detect, identify, and classify polyps. The collaboration leverages the NVIDIA Holoscan and NVIDIA IGX Orin platforms with ultra-low data transfer latency, superior image processing capabilities, and scalability to support a wide range of AI applications.

According to a report by Webmaster's Home on November 13, the field of large language models (LLMs) is growing rapidly, and the demand for professionals such as deep learning, machine learning, and natural language processing (NLP) has surged. According to an analysis by GlobalData, companies across a wide range of industries are posting GPT-4-related jobs, with some of the well-known companies looking for GPT-4 professionals including Microsoft, Citigroup, Merck & Co., Thomson Reuters, and Travelers Insurance Company. Microsoft is looking for a "Skype Senior Software Engineer" for a position dedicated to AI integration in Skype; Citigroup's "Generative AI Full Stack Engineering Lead – Code Control" position is designed to drive and contribute to the technical direction of products and services; The "AI/ML Engineer" position at Merck & Co. is designed to explore and apply the latest large language models, among other things. The emergence of these positions demonstrates the wide application of GPT-4 in various industries, providing a wealth of development opportunities for professionals in different fields.

PANews reported on October 30 that the distributed rendering network Render Network recently said that its proposal based on "RNP-005" was approved with 99.79% support. It is understood that the RNP-005 proposal aims to use Beam as a second computer client to leverage Render Network's GPU provisioning to handle machine learning workloads. Beam is an on-demand GPU cloud service designed to provide GPUs for AI and ML projects that can be used to accelerate the development and deployment of AI without the need to manage infrastructure.

AI startup Giga ML raised $3.6 million in seed funding, led by venture capital firm Nexus Ventures Partners, and angel investors such as Y Combinator and Liquid 2 Venture, 8vdx, and Garry Tan, president and CEO of Y Combinator, reported on October 23. It is reported that Giga ML was founded in 2023 to provide enterprises with customizable large-scale language models (LLM). The company says its X1 Large 32k model is more adaptable and more adaptable than the Ilama2 70B 4k model from open source LLM.

According to a report by Webmaster House on October 12, Apple's AI/ML team and Columbia University's research team developed a multimodal large model "Ferret" that can accurately find traffic lights in images, which performs better than GPT-4V and improves the accuracy of large models in "looking, speaking, answering" tasks. Ferret's key innovation lies in the close combination of spatial understanding of both referring and grounding, enabling the model to understand the semantics of a given region and find the corresponding target at the same time.

According to a report by the Science and Technology Innovation Board Daily on October 6, at a new product launch conference recently, Google officially released the new generation of Android flagship phone Pixel 8/Pro series, equipped with the Tensor G3 chip, which can run more complex ML (machine learning) ) model, adding a number of AI enhancements to the new phone, such as reading web pages to users in different languages and "more natural" voices, and virtual assistants speaking more naturally. Google points out that the Pixel 8 Pro is the first phone to run Google's basic large model directly on the device, which requires 150 times the calculation of the largest ML model on the Pixel 7. At the same time, Google announced the launch of "Assistant with Bard" for Android and iOS devices, which combines the mobile phone's personal assistant function with generative AI. Users can interact with the Bard Assistant through text, voice or images - In other words, it is multimodal.