Connect with us

Technology

Facebook Unveils AI Model Capable of Translating 100 Languages for Users

Published

on

facebook south africa

By Adedapo Adesanya

Facebook has unveiled a software based on machine learning which the company said is the first to be able to translate from any of 100 languages without relying on English.

The open-source artificial intelligence software was created to help the massive social network deliver content better in 160 languages to its more than two billion users around the world.

In a blog post, Facebook Research Assistant, Ms Angela Fan, noted that, “This milestone is a culmination of years of Facebook AI’s foundational work in machine translation.”

She said the new model is more accurate than other systems because it does not rely on English as an intermediary translation step.

It’s called M2M-100 and was trained with a data set of 7.5 billion sentence pairs from 100 different languages and the team built the universal model with 15 billion parameters.

“When translating, say, Chinese to French, most English-centric multilingual models train on Chinese to English and English to French, because English training data is the most widely available.

“Our model directly trains on Chinese to French data to better preserve meaning. It outperforms English-centric systems by 10 points on the widely used BLEU (bilingual evaluation understudy) metric for evaluating machine translations,” Ms Fan said.

Facebook said it already handles an average of 20 billion translations every day on its news feed, and that it hopes the new system will deliver better results.

“Breaking language barriers through machine translation is one of the most important ways to bring people together, provide authoritative information on Covid-19, and keep them safe from harmful content,” she said.

Facebook researchers used specific criteria for language selection, which includes those from different families with geographic diversity and those that are widely used.

However, they avoided directions that were statistically rare, including Icelandic-Nepali or Sinhala-Javanese, it noted.

Then the team organized languages into 14 groups based on linguistic classification, geography and cultural similarities.

“People living in countries with languages of the same family tend to communicate more often and would benefit from high-quality translations. For instance, one group would include languages spoken in India, like Bengali, Hindi, Marathi, Nepali, Tamil, and Urdu,“ she said.

Although Facebook provides an option to “Translate” on posts, the firm says the new AI model is 90 per cent more accurate, as it does not need to first translate text to English and then to the desired language.

Adedapo Adesanya is a journalist, polymath, and connoisseur of everything art. When he is not writing, he has his nose buried in one of the many books or articles he has bookmarked or simply listening to good music with a bottle of beer or wine. He supports the greatest club in the world, Manchester United F.C.

Click to comment

Leave a Reply

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

Technology

Zoho Launches Nathu La Server

Published

on

Zoho Nathu La Server

By Modupe Gbadeyanka

A designed-in-house server known as Nathu La has been launched by a global technology company, Zoho Corporation.

Nathu La is engineered with hardware-rooted security at every layer of the stack. Its indigenous IP-driven approach reduces dependency on external entities for security audits, firmware updates, and licensing continuity.

The solution aligns with open-source software principles and reflects Zoho’s broader commitment to building sustainable, secure, and scalable digital infrastructure. It also supports the growing global focus on digital sovereignty, local innovation ecosystems, and high-performance computing capabilities.

The platform was introduced by the company as part of a pivotal step in its journey towards building its full technology stack, from the hardware layer to software applications.

With Nathu La, Zoho has achieved equivalent performance with 12-18 per cent lower power consumption and 20-30 per cent lower total cost of ownership (TCO), thereby reducing inference costs.

The Nathu La server, comprising Intel® Xeon® 6 processors, was developed collaboratively with Intel, leveraging their enablement capabilities and technical expertise.

The design philosophy behind Nathu La is rooted in the Open Compute Project (OCP), emphasising modularity, thermal efficiency, and ease of maintenance. This enables Zoho’s data centres to significantly reduce total cost of ownership and power consumption.

Zoho plans to host its applications on the Nathu La server platform, enabling the company to optimise the full software-hardware stack for its specific workloads, reduce costs, improve performance, and strengthen data governance for its global customers. This will also help bring down inference costs for Zoho’s AI usage.

The Nathu La server motherboard and chassis platform is the result of five years of R&D across hardware, firmware, and systems management. Based on Intel® Xeon® 6 Processors, the server is designed to optimise performance for virtualisation (VM), High Performance Computing (HPC), AI inference, and storage applications. This results in improved performance of Zoho applications for end users.

The server features customised power delivery subsystems, an in-house DC-SCM (Data Centre Secure Control Module) design, and modular chassis options compatible with diverse end-user environments, offering flexibility across deployment types.

All modular components – including the DC-SCM and NIC (Network Interface Card) – were designed in-house by Zoho’s hardware engineering team and assembled through electronics manufacturing partners, enabling tighter integration and quality control across the platform. Over five patents have been filed covering advanced thermal management and cost-optimised server architecture designs.

“Zoho Corporation has invested in building its own technology stack from the ground up over the last three decades. The Nathu La server launch is in line with that goal.

“With our strategy of using contextual, right-sized models, running on our own platform, on our own servers, in our own data centres, we are compounding the benefits accrued from owning and operating our entire technology stack. This ensures that our solutions are more sustainable and accessible for businesses.

“These long-term R&D investments we are making at every layer of the stack are aimed at delivering customer value,” the Country Head for Zoho Nigeria, Mr Kehinde Ogundare, stated.

In 2020, Zoho established a small R&D team in Nagpur, a Tier 2 town in India, focused on projects such as server design and systems engineering.

Members of the Nathu La R&D team include hires from SETU – short for Students’ Engagement for Transformative Upskilling – an initiative designed to build a pipeline of industry-ready engineers, with a focus on advanced learning in Electronics System Design and Manufacturing (ESDM).

Continue Reading

Technology

MTN Fintech Targets Credit Market With Direct Lending Plans

Published

on

mtn data centre

By Adedapo Adesanya

The financial technology arm of MTN is mulling a direct shift into lending after bringing on its parent company, MTN Group, as a major investor to help cushion against losses that have plagued the business.

According to MTN Group Fintech chief executive, Mr Serigne Dioum, the company wants to move beyond helping customers access loans through partners.

He said in markets where regulators allow it, MTN wants to lend directly and use its own balance sheet.

“We’ve expanded access to credit for more people, but we also want to move further up the lending value chain,” Mr Dioum told investors at the company’s capital markets day.

“Where appropriate, we will seek licences that allow us not only to facilitate loans but also to lend directly to customers and deploy our own balance sheet.”

This development is expected to create a shift in its current fintech model which provides financial services, including deposits, payments, transfers and digital wallets to individuals and small businesses via digital and mobile‑based platforms.

The company has applied for Payment Solution Service Provider and Payment Terminal Service Provider licences through MoMo PSB, its Nigerian fintech subsidiary. If approved, the licences would allow MTN to handle more payment processing, build merchant payment tools, deploy and manage POS terminals, and reduce its dependence on third-party processors.

Despite the opportunities present in the credit market, direct lending could give MTN a larger share of revenue, but it would also expose the company to credit risk, regulation and tougher competition with banks and digital lenders.

Mr Dioum said only about 4 per cent to 5 per cent of adults have access to formal credit across the African continent. In Nigeria, the funding problem is especially severe.

A 2025 report by the National Credit Guarantee Company said nearly 80 per cent of Nigerian MSMEs lack access to formal credit, while Stears has estimated the country’s MSME financing gap at about $236 billion.

For traders, small shop owners, transport operators and households, access to small loans can determine whether they restock inventory, pay suppliers, cover emergencies or expand a business.

In April, MTN Nigeria announced that its parent firm, based in South Africa, would acquire a 60 per cent stake in MoMo Payment Service Bank Limited (MoMo PSB) and Y’ello Digital Financial Services (YDFS) Limited.

The fintech units are currently loss-making, and this move will help MTN Nigeria to reduce financial risk and share future losses and investment burden. However, it will still keep a significant minority stake (40 per cent).

Continue Reading

Technology

Meta Expands Business Agent to Instagram, WhatsApp, Messenger

Published

on

Meta Business Agent

By Aduragbemi Omiyale

The reach of the Meta Business Agent is being expanded to Instagram and other platforms of the social media giant.

Meta Business Agent is an artificial intelligence (AI) that allows business owners to attend to customers’ needs with ease.

Customers expect instant responses, but no team can be everywhere at once. This innovation handles such without hassles.

It helps businesses to answer questions specific to the business, makes product recommendations from the catalogue, books appointments, qualifies incoming leads, and closes sales.

More than one million businesses are already using a Meta Business Agent on WhatsApp and Messenger to respond to customers around the clock.

“We’re now expanding our Business Agent to businesses big and small globally, so within minutes you can have yours up and running, responding in your customer’s local language using your tone,” Meta said in a statement.

“We’re also expanding these agents to Instagram since businesses connect with their customers there, too. Businesses can activate their Business Agent here. Getting started with the Business Agent is free. In the coming months, businesses will access the agent through our paid subscription offerings, with options for businesses of every size,” it added.

Meta also stated that it is making it simpler for people to discover businesses powered by a Meta Business Agent directly on WhatsApp. It noted that starting soon, people will be able to find businesses by typing their name in the Search bar, or by sharing their phone number or contact card in chats with friends and family. This way, when more customers reach out, they get a quick, helpful response.

Continue Reading

Trending