Connect with us

Technology

Researchers Develop Algorithm for Optimal Decision Making Under Heavy-tailed Noisy Rewards

Published

on

Algorithmic Trading

By Adedapo Adesanya

The exploration algorithms for stochastic multi-armed bandits (MABs)–sequential decision-making problems under uncertain environments–typically assume light-tailed distributions for reward noises.

However, real-world datasets often show heavy-tailed noise. In light of this, researchers from Korea propose an algorithm that can achieve minimax optimality (minimum loss under maximum loss scenario) with minimal prior information.

Superior to existing algorithms, the new algorithm has potential applications in autonomous trading and personalized recommendation systems.

In data science, researchers typically deal with data that contain noisy observations. An important problem explored by data scientists in this context is the problem of sequential decision-making. This is commonly known as a “stochastic multi-armed bandit” or (stochastic MAB).

Here, an intelligent agent sequentially explores and selects actions based on noisy rewards under an uncertain environment. Its goal is to minimize cumulative regret–the difference between the maximum reward and the expected reward of selected actions. A smaller regret implies more efficient decision-making.

Most existing studies on stochastic MABs have performed regret analysis under the assumption that the reward noise follows a light-tailed distribution. However, many real-world datasets, in fact, show a heavy-tailed noise distribution.

These include user behavioural pattern data used for developing personalized recommendation systems, stock price data for automatic transaction development, and sensor data for autonomous driving.

In a recent study, Assistant Professor Kyungjae Lee of Chung-Ang University and Assistant Professor Sungbin Lim of the Ulsan Institute of Science and Technology, both in Korea, addressed this issue. In their theoretical analysis, they proved that the existing algorithms for stochastic MABs were sub-optimal for heavy-tailed rewards.

More specifically, the methods employed in these algorithms–robust upper confidence bound (UCB) and adaptively perturbed exploration (APE) with unbounded perturbation–do not guarantee a minimax (minimization of maximum possible loss) optimality.

“Based on this analysis, minimax optimal robust (MR) UCB and APE methods have been proposed. MR-UCB utilizes a tighter confidence bound of robust mean estimators, and MR-APE is its randomized version. It employs bounded perturbation whose scale follows the modified confidence bound in MR-UCB,” explains Dr Lee, speaking of their work, which was published in the IEEE Transactions on Neural Networks and Learning Systems on 14 September 2022.

The researchers next derived gap-dependent and independent upper bounds of the cumulative regret. For both the proposed methods, the latter value matches the lower bound under the heavy-tailed noise assumption, thereby achieving minimax optimality.

Further, the new methods require minimal prior information and depend only on the maximum order of the bounded moment of rewards. In contrast, the existing algorithms require the upper bound of this moment a priori–information that may not be accessible in many real-world problems.

Having established their theoretical framework, the researchers tested their methods by performing simulations under Pareto and Fréchet noises. They found that MR-UCB consistently outperformed other exploration methods and was more robust with an increase in the number of actions under heavy-tailed noise.

Further, the duo verified their approach for real-world data using a cryptocurrency dataset, showing that MR-UCB and MR-APE were beneficial–minimax optimal regret bounds and minimal prior knowledge–in tackling heavy-tailed synthetic and real-world stochastic MAB problems.

“Being vulnerable to heavy-tailed noise, the existing MAB algorithms show poor performance in modelling stock data. They fail to predict big hikes or sudden drops in stock prices, causing huge losses. In contrast, MR-APE can be used in autonomous trading systems with stable expected returns through stock investment,” comments Dr Lee, discussing the potential applications of the present work.

“Additionally, it can be applied to personalized recommendation systems since behavioural data shows heavy-tailed noise. With better predictions of individual behaviour, it is possible to provide better recommendations than conventional methods, which can maximize the advertising revenue,” he concludes.

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.

Advertisement
Click to comment

Leave a Reply

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

Technology

Google Pumps $37m into Africa to Boost AI Research, Digital Skills, Others

Published

on

Google AI Community Centre

By Aduragbemi Omiyale

About $37 million is being invested in Africa by Google to ensure the continent is not left behind in technology, with $7 million earmarked for Artificial Intelligence (AI) education in Ghana, Nigeria, Kenya, and South Africa to support academic institutions and nonprofits building localized AI curricula, online safety training, and cybersecurity programs.

In addition, two new $1 million grants from Google.org aim to bolster AI research capacity across the continent.

One grant goes to the African Institute for Data Science and Artificial Intelligence (AfriDSAI) at the University of Pretoria to support applied AI research and training. The other supports the Wits Machine Intelligence and Neural Discovery (MIND) Institute in South Africa, which will fund MSc and PhD students to conduct foundational AI research and help shape Africa’s role in the global AI landscape.

Also, the tech giant is providing $25 million for an AI Collaborative for Food Security, which will bring together researchers, and nonprofit organizations to co-develop AI tools for early hunger forecasting, crop resilience, and tailored guidance for smallholder farmers.

The goal is to help make food systems across Africa more adaptive, equitable, and resilient in the face of increasing climate and economic shocks.

Further, the company is providing $3 million for Masakhane Research Foundation to support the development of high-quality datasets, machine translation models, and speech tools that make digital content more accessible to millions of Africans in their native languages.

To further empower innovation, Google is launching a catalytic funding initiative to support AI-driven startups tackling real-world challenges. This platform will combine philanthropic capital, venture investment, and Google’s technical expertise to help more than 100 early-stage ventures scale AI-based solutions in agriculture, healthcare, education, and other vital sectors. Startups will also receive mentorship, access to tools, and technical guidance to support responsible development.

Google has also launched an AI Community Centre in Accra for AI learning, experimentation, and collaboration in Africa. The facility will host training sessions, community events, and workshops focused on responsible AI development.

Its programming will span four pillars: AI literacy, community technology, social impact, and arts and culture — providing a platform for a diverse ecosystem of developers, students, and creators to engage with AI in ways that are grounded in African priorities.

To help meet the rising demand for AI and digital skills, Google is rolling out 100,000 Google Career Certificate scholarships for students in higher learning institutions across Ghana.

These fully funded, self-paced programs will focus on AI Essentials, Prompting Essentials, and other high-growth fields like IT Support, Data Analytics, and Cybersecurity — enabling more learners to access job-ready training and build careers in AI and the digital economy.

“Africa is home to some of the most important and inspiring work in AI today. We are committed to supporting the next wave of innovation through long-term investment, local partnerships, and platforms that help researchers and entrepreneurs build solutions that matter,” the Senior Vice President for Research, Labs, and Technology and Society at Google, Mr James Manyika, said.

Also, the Vice President of Engineering and Research at Google, Mr Yossi Matias, stated, “This new wave of support reflects our belief in the talent, creativity, and ingenuity across the continent. By building with local communities and institutions, we’re supporting solutions that are rooted in Africa’s realities and built for global impact.”

Continue Reading

Technology

How This AI Alert by Airtel is Transforming Mobile Security in Africa

Published

on

AI Alert by Airtel Odeshi

These days, people rely heavily on their mobile phones for talking, texting, banking, social media, and storing important personal information. Because of this, scammers and spammers often target phone users.

Mobile fraud, like fake SMS messages and scams, is becoming more common and putting millions at risk of losing money or having their private details stolen.

Airtel’s new AI-powered Spam Alert Service offers a smart and timely way to fight back, marking a major step forward in protecting mobile users in Africa.

A brief look at mobile fraud and spam

Mobile fraud and spam are problems around the world, but they hit harder in places where mobile phone use is growing fast, and safety measures haven’t caught up.

A 2024 report from GSMA Mobile Economy shows that more than 20% of mobile users globally have experienced some kind of mobile fraud, with spam texts being one of the most common.

In Nigeria, the Nigerian Communications Commission (NCC) has noted a sharp rise in scam messages and fake calls, leading to yearly losses in the hundreds of millions of dollars.

These spam texts often include fake links, harmful ads, or tricks to steal personal details. Many people get caught without knowing, which can lead to stolen bank money, identity theft, or damaged devices.

Older spam blockers only work on certain phones or apps, leaving many people, especially those using basic phones, without protection.

 Why Airtel’s AI Spam Alert Service stands out

Airtel, a top telecom company in Africa, has launched a new and free service called the AI Spam Alert Service. It’s the first of its kind in Africa and aims to protect mobile users from spam text messages as they come in.

What makes this service different is that it doesn’t read or check the actual message content. Instead, it uses advanced artificial intelligence to quickly study the sender’s behaviour using over 250 signs or patterns, all within a fraction of a second.

Some of those parameters, according to Airtel, includes:

  • How frequently the sender changes SIM cards.
  • The volume and frequency of messages sent by the message initiator to different recipients.
  • The geographical spread of the recipient numbers, whether messages are targeted locally or dispersed nationwide.
  • Whether the sender receives replies or only sends messages.
  • A cross-reference of numbers previously reported for spam activity.

The AI completes this analysis in just 2 milliseconds, faster than the blink of an eye, allowing real-time alerts to subscribers as suspected spam messages arrive.

How Airtel’s Spam Alert Service improves mobile safety and trust

Airtel’s new AI-powered spam alert system is set to make a big difference in mobile security across Nigeria and the rest of Africa. Instead of depending on users to block spam themselves, the service works directly through the network to stop suspicious messages before they reach people’s phones.

This kind of technology helps users feel safer and more confident using mobile networks—especially as more people rely on their phones for banking and other money-related services.

A 2025 report by McKinsey Digital shows that many Africans worry about mobile security when using digital financial tools. By reducing the risk of spam and fraud, Airtel is not only protecting its users but also helping to build a safer digital space where more people can take part in the growing mobile economy.

According to Airtel, within two months of its launch, the spam alert service system has identified 9,667,008 messages as potential spam.

Why this is a game changer

By building a service that is first-of-its-kind in Africa, Airtel is leading the way in offering spam alerts, powered by AI directly on its network, for over 150 million subscribers across the continent.

In addition, the service is quick, spotting suspicious sender activity in just milliseconds without reading users’ messages.

Finally, the service is free and requires no app downloads or extra setup. Airtel says the turns on automatically, making it easy for everyone to stay protected, even those using basic phones.

By sending signals to users before problems happen, Airtel boosts trust and encourages more people to safely use mobile money and other digital services.

Scammers are always finding new ways to trick consumers through the ubiquitous mobile phones, so increased demand for security improvements is, understandably, shifting to telecom companies. Airtel’s AI Spam Alert Service is a strong and timely move toward better, smarter protection for users. As more people start using the service, it should help cut down fraud, keep personal information safe, and make mobile use more enjoyable.

As African economies continue to grow more digital, users of telecom services will need more secure and reliable ways to communicate, and Airtel is leading the way towards that safe future.

Continue Reading

Technology

MTN $150m Data Centre Will Unlock Productivity, Drive Diversification—FG

Published

on

mtn data centre

By Aduragbemi Omiyale

The $150 million data centre established in Lagos by MTN Nigeria has been described by the federal government as a gamechanger because of it aligns with its digital economy agenda.

The Minister of Communications, Innovation and Digital Economy, Mr Bosun Tijani, speaking during the unveiling recently in Ikeja, Lagos, said the facility supports the $1 trillion economy the current administration aims to build.

“An investment like this, the one we are here to launch, offers a platform for our young people to thrive. Enterprise-grade infrastructure like this, on our soil, gives startups, developers, and digital creators the ability to build and scale from Nigeria to the world.

“With this facility, MTN is reinforcing its position as Nigeria’s digital backbone. The data centre, named after the late Sifiso Dabengwa, a former CEO of MTN Nigeria and later Group Chief Operating Officer before his passing in September last year is being hailed as Nigeria’s largest prefabricated modular data centre

“It will deliver 4.5 MW in phase 1, with an additional 4.5 MW to be delivered in phase 2, which is expected to be completed soon,” he stated.

Mr Tijani noted that the MTN Data Centre would contribute to growing Nigeria’s economy by “unlocking productivity, hiring enterprise, and driving diversification through technological innovation and inclusion.”

The chief executive of the Nigerian Communications Commission (NCC), Mr Aminu Maida, represented by the Deputy Director for New Media and Information Security Department, Mr Babagaba Digima, praised MTN’s leadership in digital innovation.

“The commission remains committed to creating an environment that supports innovation while ensuring the highest standards of cybersecurity, data protection, and robust internet infrastructure. We will continue to work closely with operators to ensure that the deployment of critical infrastructure meets the high standards our digital economy deserves,” he stated.

Also, the Governor of Lagos State, Mr Babajide Sanwo-Olu, represented by the Secretary to the State Government (SSG), Mrs Bimbola Salu-Hundeyin, said the facility “not only signifies MTN’s unwavering commitment to investing in Nigeria’s digital infrastructure but also reinforces the strategic importance of Lagos as a technology and innovation hub for the nation.”

“As we all know, data is the new oil, and cloud technology is the engine that drives it. With its Tier III facilities, MTN is raising the bar for secure, scalable, and efficient enterprise services, critical enablers for businesses, public services, and national as well as multinational corporations alike,” Mr Sanwo-Olu stated.

Also speaking, the chief executive of MTN Nigeria, Mr Karl Toriola, said, “We are committed to building locally managed, globally competitive digital platforms that will enable businesses to scale faster and engage more people in wide-ranging research and development.

“At MTN, we believe everyone, particularly Nigerians and Africans, deserves the benefits of a modern, connected life. We continue to push boundaries to make the humanly impossible, conceivable, feasible, and ultimately possible.”

Continue Reading

Trending