Connect with us

Technology

WiMi Researches Reinforcement Learning-Based Blockchain Federated Learning Framework to Optimize Model Aggregation Strategy and Security

Published

on

BEIJING, Nov. 8, 2024 /PRNewswire/ — WiMi Hologram Cloud Inc. (NASDAQ: WiMi) (“WiMi” or the “Company”), a leading global Hologram Augmented Reality (“AR”) Technology provider, today announced the initiation of exploring the integration of Reinforcement Learning (RL) into the federated learning framework. RL, as a significant branch of machine learning, has become a crucial tool for optimizing the federated learning process due to its decision-making capabilities in complex environments.

Reinforcement Learning is a machine learning approach that enables an intelligent agent to learn optimal strategies through interactions with the environment. In a blockchain-based federated learning framework utilizing reinforcement learning, the reinforcement learning algorithm can dynamically adjust the timing of model aggregation, selection of data participants, and transaction costs. This achieves a balance between information timeliness and data bias, as well as intelligent control over transaction costs, ultimately optimizing the overall learning performance.

In federated learning, there can be significant differences in the datasets of different participants, known as the data bias problem. Additionally, model updates need to be aggregated at the appropriate timing to avoid outdated information affecting overall learning performance. The reinforcement learning algorithm can simulate interactions with the environment to learn when to upload model updates and how to select the most effective models for aggregation under different data distributions. This helps find the optimal balance between information timeliness and data bias. The cost of blockchain transactions, including the consumption of computational resources and network bandwidth, is another important consideration in federated learning. Reinforcement learning can intelligently predict network conditions, resource availability, and transaction priorities to dynamically adjust the frequency and scale of model aggregation. This ensures learning effectiveness while minimizing overall transaction costs. By applying reinforcement learning algorithms to optimize model aggregation strategies, not only does it significantly improve federated learning efficiency and model accuracy, but it also effectively reduces transaction costs.

With the continuous advancement of technology, blockchain-based federated learning frameworks based on reinforcement learning will play a crucial role in various fields such as healthcare, financial services, and the Internet of Things (IoT), promoting the security, efficiency, and widespread adoption of artificial intelligence technology. For example, in the healthcare industry, this framework can facilitate data sharing among hospitals, research institutions, and patients, accelerating the development of disease diagnosis and treatment plans while strictly protecting individual privacy. In the financial services industry, it can assist banks and financial institutions in building more secure and efficient credit assessment and risk management models. In the field of IoT, it enables intelligent collaboration among devices, enhancing the overall network’s responsiveness and intelligence level.

WiMi’s research on the blockchain-based federated learning framework using reinforcement learning represents a significant innovation at the intersection of artificial intelligence, blockchain technology, and reinforcement learning. It provides innovative approaches to address the trust, security, and efficiency issues faced by traditional federated learning. In the future, with further theoretical research and practical applications, the technological potential of blockchain-based federated learning using reinforcement learning will be more fully explored and widely applied in various industry sectors.

About WiMi Hologram Cloud

WiMi Hologram Cloud, Inc. (NASDAQ:WiMi) is a holographic cloud comprehensive technical solution provider that focuses on professional areas including holographic AR automotive HUD software, 3D holographic pulse LiDAR, head-mounted light field holographic equipment, holographic semiconductor, holographic cloud software, holographic car navigation and others. Its services and holographic AR technologies include holographic AR automotive application, 3D holographic pulse LiDAR technology, holographic vision semiconductor technology, holographic software development, holographic AR advertising technology, holographic AR entertainment technology, holographic ARSDK payment, interactive holographic communication and other holographic AR technologies.

Safe Harbor Statements

This press release contains “forward-looking statements” within the Private Securities Litigation Reform Act of 1995. These forward-looking statements can be identified by terminology such as “will,” “expects,” “anticipates,” “future,” “intends,” “plans,” “believes,” “estimates,” and similar statements. Statements that are not historical facts, including statements about the Company’s beliefs and expectations, are forward-looking statements. Among other things, the business outlook and quotations from management in this press release and the Company’s strategic and operational plans contain forward−looking statements. The Company may also make written or oral forward−looking statements in its periodic reports to the US Securities and Exchange Commission (“SEC”) on Forms 20−F and 6−K, in its annual report to shareholders, in press releases, and other written materials, and in oral statements made by its officers, directors or employees to third parties. Forward-looking statements involve inherent risks and uncertainties. Several factors could cause actual results to differ materially from those contained in any forward−looking statement, including but not limited to the following: the Company’s goals and strategies; the Company’s future business development, financial condition, and results of operations; the expected growth of the AR holographic industry; and the Company’s expectations regarding demand for and market acceptance of its products and services.

Further information regarding these and other risks is included in the Company’s annual report on Form 20-F and the current report on Form 6-K and other documents filed with the SEC. All information provided in this press release is as of the date of this press release. The Company does not undertake any obligation to update any forward-looking statement except as required under applicable laws.

 

 

View original content:https://www.prnewswire.com/news-releases/wimi-researches-reinforcement-learning-based-blockchain-federated-learning-framework-to-optimize-model-aggregation-strategy-and-security-302300038.html

SOURCE WiMi Hologram Cloud Inc.

Continue Reading
Click to comment

Leave a Reply

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

Technology

ASUS Accelerates Enterprise AI at Scale with 6th-Gen AMD EPYC Server CPUs

Published

on

By

 ASUS leverages 6th-gen AMD EPYC Server CPUs to deliver scalable, efficient compute for enterprise AI, cloud, virtualization and business-critical workloads

SAN FRANCISCO, July 24, 2026 /PRNewswire/ — ASUS today announced its groundbreaking new server lineup powered by the AMD EPYC™ 9006 processors, engineered to deliver unmatched performance for the most demanding intensive enterprise workloads. This advanced portfolio introduces two highly optimized series with efficiency-optimized AMD EPYC SP8 server CPU, the flagship dual-socket ASUS RS700A/720A for extreme compute density and the single-socket ASUS RS500A/520A for superior space efficiency and deployment flexibility.

Both series integrate full PCIe® 6.0, leading memory support, and high-density E3.S storage, all underpinned by proprietary ASUS innovations for superior thermal management and operational efficiency to meet and exceed the rigorous demands of enterprise AI, virtualization, storage and cloud environments.

“The new ASUS server series, powered by 6th-gen AMD EPYC server CPUs, is engineered to power every enterprise workload with flexible, scalable infrastructure,” Paul Ju, Senior Vice President of ASUS, commented, “This launch marks a significant milestone for ASUS and our clients. The new series empowers businesses with a resilient foundation to achieve unprecedented computing efficiency and accelerating AI innovation with inference.”

ASUS expands 6th-gen AMD EPYC server portfolio with dual optimized series

ASUS has introduced a new server lineup segmented into two distinct series, each precisely engineered to meet diverse enterprise demands.

The flagship RS700A/720A series (dual-socket) delivers extreme compute density, making it ideal for AI inferencing, and complex simulations. It offers exceptional bandwidth with PCIe 6.0, memory leadership via 32 DIMM slots supporting ultrafast MRDIMM, and high-density storage with up to 32 E3.S bays in a compact 2U form factor.

Complementing this is the RS500A/520A series (single-socket), a highly efficient and space-optimized solution with depth under 800mm, perfect for mainstream enterprise workloads and rack-constrained environments. Featuring full PCIe 6.0 capabilities, E3.S storage support, and modular scalability through shared components with the RS700A and RS720A series, it provides uncompromised performance in a streamlined, deployment-friendly design.

ASUS elevates the AMD EPYC platform with cutting-edge proprietary innovations

ASUS has significantly advanced the AMD EPYC 9006 platform with a series of proprietary engineering breakthroughs focused on superior reliability, thermal management, and operational efficiency.

The DC-MHS modular architecture features a zone-partitioned chassis that separates I/O, HPM, fan, and storage modules to accelerate development, reduce capital costs, and enable rapid serviceability. The patented ASUS DIMM.2 Innovation repositions M.2 storage to the cooler DIMM region, eliminating thermal throttling without extra heatsinks and unlocking greater scalability. Thermal Radar 3.0 with PID Control delivers precise real-time fan regulation via advanced algorithms, reducing energy use and maintaining peak performance under heavy enterprise-level workload.

Completing the suite is the optimized tool-less operational-velocity design, which boosts maintenance efficiency, maximizing uptime and lowering TCO and sustaining peak performance even under volatile, high-load AI/HPC workloads.

AVAILABILITY & PRICING

ASUS RS700A/720A series and RS500A/520A series servers will be available soon. Please contact your local ASUS representative for further information.

View original content to download multimedia:https://www.prnewswire.co.uk/news-releases/asus-accelerates-enterprise-ai-at-scale-with-6th-gen-amd-epyc-server-cpus-302833931.html

Continue Reading

Technology

Fractal posts 20% revenue growth and 92% net income growth in Q1 FY27

Published

on

By

Adjusted EBITDA Grows at 35% YoYGross Margin up 29 bps1 to 46%; Adjusted EBITDA Margin up 189 bps to 17%

NEW YORK, July 24, 2026 /PRNewswire/ — Fractal Analytics Ltd (BSE: 544700) (NSE: FRACTAL) announced its consolidated financial results for Q1 FY27, ending June 30, 2026.

In Q1 FY27, the Company reported consolidated operating revenue of INR 9,125 m, a growth of 20% year on year (YoY). Revenue growth was led by the company’s Healthcare and Life Sciences (HLS) industry, which clocked 69% growth YoY. Strong sustained growth in HLS over the last several quarters has resulted in it becoming the second largest industry in the portfolio. Banking, Financial Services and Insurance (BFSI) also performed very well, growing 36% YoY in Q1. Fractal’s largest industry, Consumer Packaged Goods and Retail (CPGR), continued to gather momentum, growing 19% YoY. On the other hand, TMT declined 22% YoY.

Fractal’s focus on deepening customer relationships continues to yield good outcomes. Its clients collectively increased their spending with the company, as reflected in the Net Revenue Retention2 of 117% in Q1. Further, its Net Promoter Score (NPS) during the period stood at 77.

The company reported improved profit margins at all levels. Gross Margin in Q1 was at 46%, while Adjusted EBITDA Margin expanded by 189 bps YoY to 17%. Net Income grew 92% YoY to INR 723 m.

Commenting on the performance, Srikanth Velamakanni, Group CEO and Executive Vice-Chairman, said:

“Enterprises are putting real transformation budgets behind AI now and we’re seeing it directly in the size of the deals coming to us. TMT was the drag on our headline growth this quarter. Excluding TMT, our business grew 35% year on year, which is a better read on the underlying demand we’re seeing.

As data sovereignty becomes a bigger priority for governments and enterprises, and as open-weight models keep improving, clients need a partner who can work across models and infrastructure. We have invested heavily in our people, our research, and our own intellectual property to be that partner.”

1 Basis points = 1/100th of 1%
2 Net Revenue Retention in our Fractal.ai segment measures how effectively we retain and expand revenue from our existing clients over a defined period and is calculated by comparing the current period’s revenue from the clients who existed at the start of the period, with their revenue in the previous period – including the effects of upsells, cross-sells and contractions

About Fractal 

Fractal Analytics Ltd (BSE: 544700) (NSE: FRACTAL) is a globally recognized pure-play enterprise AI company trusted by Fortune 500®-sized enterprises to power decision-making through AI services, solutions, and products, anchored by Cogentiq, its flagship agentic AI platform. With over 6,000 professionals across North America, EMEA, and Asia-Pacific, Fractal partners with business leaders to drive competitive differentiation for their organizations by embedding AI into critical decisions across business functions and industry verticals.

Fractal invests more than 6% of its revenue in AI R&D, supporting foundational AI research, product development, and IP creation that address both immediate client needs and long-term technological advancement. Fractal’s track record includes developing proprietary models and products such as Cogentiq Health – Vaidya.ai and Cogentiq Data Science – PiEvolve, as well as incubating and spinning out Qure.ai, a global healthcare AI leader focused on the rapid identification and management of tuberculosis, lung cancer, and stroke (or critical health conditions). Fractal’s suite of businesses consists of Asper.ai (a Revenue Growth Management product for CPG companies) and Analytics Vidhya (an Ed-tech platform).

For more information, go to www.fractal.ai.

Logo: https://mma.prnewswire.com/media/2931510/5858548/Fractal_Logo.jpg

 

View original content:https://www.prnewswire.com/news-releases/fractal-posts-20-revenue-growth-and-92-net-income-growth-in-q1-fy27-302833932.html

SOURCE Fractal Analytics Limited

Continue Reading

Technology

Xryma Plc : Pre-Listing Liquidity Facility and Price Discovery Process

Published

on

By

NICOSIA, Cyprus, July 24, 2026 /PRNewswire/ — Xryma Plc (“Xryma”)  announces its intention to reapply within the next twelve months for admission to list on Euronext Paris (“Euronext”), with such admission being subject to Euronext’s approval. Before submitting its application, Xryma intends to launch a pre-listing liquidity facility and price discovery process, comprising a private placement to institutional and qualified investors alongside a secondary market offer to Xryma existing shareholders (“shareholders”) wishing to exit prior to listing.  

The admission referred to above that is subject to the approval of Euronext may also be subject to approval by relevant regulatory authorities, and no assurance can be given that approval will be granted or as to the timing of any admission.

The pre-listing liquidity facility and price discovery process is designed to:

Enable shareholders seeking an exit to participate without the need to open an EU brokerage account,Provide a clear and orderly opportunity for existing shareholders to sell all or part of their holdings ahead of any potential admission to trading on Euronext Paris,Enable shareholders to sell all or part of their holdings at the same price at which qualified and institutional investors subscribe for shares in the Company,Establish, through a bookbuild with qualified and institutional investors, a market-validated referenced price for Xryma shares ahead of any potential admission on Euronext Paris (the “Primary Market Placement Price”),Support orderly trading upon potential admission.

Individual shareholder mailouts explaining the details of the pre-listing liquidity facility scheme with instructions and necessary documentation will be conducted during August 2026.

As the Primary Market Placement Price is to be determined by the subsequent bookbuild, shareholders will be given the opportunity to set a floor price which will result in the sale of their shares if the Primary Market Placement Price is higher.  Shareholders will receive the Primary Market Placement Price minus applicable fees.

Shareholders and Investors may be scaled back to match corresponding demand from the other party, with partial fulfilment a possibility if the Company cannot match supply to demand.

Completion of the process is subject to achieving a level of institutional and qualified investor demand that the Board considers appropriate to support an orderly market should Xryma subsequently be admitted to trading on Euronext Paris.

Participation is entirely voluntary. Shareholders who do not wish to sell will simply retain their shares. Shareholders that do not intend to participate should continue to onboard with a Euronext participating broker, or a Euroclear ESES custodian, per previous communications.

The major shareholders, SCP Select All Enterprise (Monaco) and SCP Red 5 Solutions (Monaco) will not participate in the offer and will be subject to lock up arrangements.

Mr Nikogiannis (John) Karantzis, CEO of Xryma Plc comments: “Our shareholders have told us they would value a straightforward way to realise their holdings without the time and cost of opening an EU brokerage account. This process is our response to that feedback. We are structuring the placement to be large enough to establish a credible reference price whilst limiting dilution, with demand directed first towards meeting shareholder sell interest. We look forward to updating the market on the revised timetable in due course.”

Shareholders seeking a more detailed explanation of the pre-listing liquidity facility and price discovery process, should refer to the guide available at https://www.xryma.com/investors

Important Information & Disclaimers

This press release may contain inside information within the meaning of Article 7(1) of Regulation (EU) 596/2014 (Market Abuse Regulation).

This publication is not for publication or distribution or release, directly or indirectly, in or into the United States of America (including its territories and possessions, any state of the United States and the District of Columbia), Canada, Australia, South Africa, Japan or any other jurisdiction where such an announcement would be unlawful. The distribution of this publication may be restricted by law in certain jurisdictions and persons into whose possession this document or other information referred to herein comes should inform themselves about and observe any such restriction. Any failure to comply with these restrictions may constitute a violation of the securities laws of any such jurisdiction. No action has been taken that would permit an offering of the treasury shares or possession or distribution of this publication in any jurisdiction where action for that purpose is required.

This publication does not constitute or form part of an offer for sale or solicitation of an offer to purchase or subscribe for securities in the United States, Canada, Australia, South Africa, Japan or any other jurisdiction and the securities referred to herein have not been registered under the securities laws of any such jurisdiction. Any New Shares (if such are issued) will not be registered under the United States Securities Act of 1933, as amended (the “Securities Act”), or under the securities laws of any State or any other jurisdiction of the United States, and may not be offered or sold, directly or indirectly, in the United States except pursuant to an exemption from, or in a transaction not subject to, the registration requirements of, the Securities Act and in compliance with all applicable securities laws of any State or any other jurisdiction of the United States. No public offering of securities is being made in the United States or in any other jurisdiction.

The information set forth herein must not be distributed in any jurisdiction where such distribution is unlawful, and any recipients are requested to inform themselves about and to observe such restrictions.

The Offering referred to herein by Xryma Plc will only be made in accordance with all applicable corporate and securities laws. Any shares referred to herein will exclusively be offered or sold in reliance on any applicable exemptions from prospectus or registration requirements in any jurisdiction. In member states of the European Economic Area, this publication is only addressed to and directed at persons who are ‘qualified investors’ within the meaning of Article 2(e) of Regulation (EU) 2017/1129 (as amended and including any relevant delegated regulations, the “Prospectus Regulation”) or in any other circumstances falling within exemptions available in the relevant member state under Article 1(4) and/or 1(5) of the Prospectus Regulation. In the United Kingdom, this publication is only addressed to and directed at qualified investors within the meaning of the Prospectus Regulation, as it forms part of domestic law by virtue of the European Union (Withdrawal) Act 2018, as amended (“EUWA”), who are persons (i) who have professional experience in matters relating to investments falling within Article 19(5) (investment professionals) of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2005 (as amended, the “Order”), (ii) falling within article 49(2)(a) to (d) (high net worth companies, incorporated associations, etc.) of the Order, or (iii) to whom it may otherwise be lawfully communicated; any other persons in the United Kingdom should not take any action on the basis of this publication and should not act on or rely on it.

This publication does not constitute a recommendation concerning the prospective Offering. This announcement does not constitute an Offer or invitation to subscribe.

This announcement includes statements that are, or may be deemed to be, ‘forward looking statements’. These forward-looking statements can be identified by the use of forward looking terminology, including the terms ‘believes’, ‘estimates’, ‘anticipates’, ‘expects’, ‘intends’, ‘may’, ‘will’, or ‘should’ or, in each case, their negative or other variations or comparable terminology, or by discussions of strategy, plans, objectives, goals, future events or intentions. By their nature, forward looking statements involve risk and uncertainty because they relate to future events and circumstances which may or may not occur. Many of these factors are beyond the control of the Company. Should one or more of these risks or uncertainties materialise, or should underlying assumptions prove incorrect, actual results and circumstances may vary materially from those described in this announcement as anticipated, believed, estimated or expected.

View original content to download multimedia:https://www.prnewswire.co.uk/news-releases/xryma-plc–pre-listing-liquidity-facility-and-price-discovery-process-302833658.html

Continue Reading

Trending