TECH DIGEST- Microsoft CEO Satya Nadella introduced Phi-2, an advanced AI model boasting a staggering 2.7 billion parameters. What sets Phi-2 apart is not just its impressive size but its ability to rival and often surpass AI models up to 25 times its size, marking a paradigm shift in the relationship between model dimensions and performance.

Phi-2’s remarkable capabilities have been demonstrated across various domains, showcasing exceptional proficiency in language comprehension, mathematical problem-solving, programming, and information processing.

What makes Phi-2 truly stand out is its integration into Microsoft Azure AI Studio, providing researchers and developers with a powerful tool for seamlessly incorporating advanced AI into their applications.

A key feature of Phi-2 is its training on meticulously curated data corresponding to varying education levels. This unique approach contributes to the model’s accuracy and efficiency, enabling it to excel in knowledge transfer and interpretation. Microsoft’s emphasis on carefully selecting and structuring training data challenges conventional thinking and positions Phi-2 as a frontrunner in language models.

Phi-2’s compact size, relative to its counterparts, does not compromise its performance. It stands shoulder to shoulder with larger models such as the 7B Mistral and the 13B Llama 2 from Meta, even competing with the massive 70B Llama-2. Notably, Phi-2 surpasses Google’s Gemini Nano, showcasing its potential for mobile technology. With automatic text summarization, advanced proofreading, grammar correction, and contextual intelligent responses, Phi-2 emerges as a game-changer in the mobile AI landscape.

What truly sets Phi-2 apart is its ethical foundation. Microsoft’s experts highlight that Phi-2 achieves its high performance without relying on reinforcement learning methods based on human feedback or instructive optimization methods. This approach minimizes bias and reduces the risk of toxic emissions, positioning Phi-2 as one of the safest and most ethical models in the AI realm.

Phi-2 signals a new era in machine learning, where training quality and efficiency take center stage. Microsoft not only showcases significant strides in AI but also opens doors for global developers and researchers. Phi-2 heralds a future of smarter and safer technologies, redefining the landscape of possibilities in the ever-evolving field of artificial intelligence. As the industry witnesses this remarkable leap, Phi-2 stands as a testament to Microsoft’s commitment to pushing the boundaries of innovation.

The post Artificial Intelligence: Microsoft Unveils Phi-2 appeared first on Tech Digest News.

TECHDIGEST – AI which stands for artificial intelligence refers to systems or automated systems that perform tasks by mimicking human intelligence and can continually refine themselves based on the data they collect.

AI manifests itself in a variety of ways. Christopher Strachey, later director of the University of Oxford’s Programming Research Group, wrote the first successful AI program in 1951. Strachey’s checkers (draughts) program ran on the Ferranti Mark I computer at Manchester University in England.

This program could play a complete game of checkers at a reasonable speed by the summer of 1952. AI is very much about the process and the ability to think faster and analyze the data than it is about any specific format or function.

Although images of high-functioning, human-like robots taking over the world conjure up images of AI, the technology is not intended to replace humans. Its goal is to significantly improve human capabilities and contributions.

READ ALSO: ITU Appoints Pantami To Lead World Summit On Information Society Forum 2022

As it relates to artificial intelligence, there are several types of learning. The most basic method is trial and error.

A simple computer program for solving mate-in-one chess problems, for example, might try moves at random until a mate is found. The program could then save the solution along with the position so that the next time the computer encountered the same position, it could retrieve the solution.

This simple rote learning of individual items and procedures is relatively simple to implement on a computer.

Problem-solving, notably in artificial intelligence, can be defined as a systematic search through a set of possible actions to achieve a given objective or course of action. Problem-solving techniques are classified as either special purpose or general-purpose.

A special-purpose method is tailored to a specific problem and frequently takes advantage of very specific features of the situation in which the problem is embedded.

A general-purpose method, on the other hand, is applicable to a wide range of problems. One general-purpose AI technique is means-end analysis, which is a step-by-step, or incremental, reduction of the gap between the current state and the final goal.

AI programs have solved a wide range of problems. Finding the winning move (or sequence of moves) in a board game, formulating discrete mathematics, and manipulating “virtual objects” in a computer-generated world are some examples.

AI has evolved in lockstep with computer processing power, which appears to be the primary limiting factor. Early AI projects, such as chess and mathematical problem solving, are now considered trivial in comparison to visual pattern recognition, complex decision making, and the use of natural language.

It is a driving force behind the Fourth Industrial Revolution and one of the most important technologies for business, the economy, and society. This necessitates a multifaceted approach and a holistic understanding of AI, encompassing technical, organizational, regulatory, societal, and philosophical aspects.
AI is arguably one of the most powerful technologies in the history of business, the economy, and society. While there are many different types of AI algorithms and AI systems, two that are commonly used in business are those that use predetermined, and possibly human-defined, rules to make predictions, recommendations, and decisions, and those that learn these “rules” (which are in general mathematical functions) from data.

For many years, organizations have used the former type of AI.

The second type of AI system is based on “machine learning.” Machine learning (ML) is on the rise. The heart of the majority of today’s AI systems.

ML is a field that combines computer science and mathematics. Science, math, and statistics concentrate on the creation of algorithms capable of analyzing data that is typically large in volume and complex in nature structure, in order to identify any patterns that may exist to be used to make predictions, decisions or recommendations.

Certain human functions, such as sensing and learning, can be performed by AI:
Sensing: AI allows machines to perceive the outside world by analyzing images, sounds, speech, text, and other data collected via IoT sensors or directly fed into the system.

Learning: AI allows machines to improve their performance over time by learning from past results, successes, and failures.
AI has gradually crept into the mainstream. It now pervades people’s lives, tracking their movements, navigating them, and recommending content to watch, read, and listen to. Access to internet information is mediated by search engines, recommendation systems, and digital assistants.

Examples of common use cases in sales that use AI technologies for analysis, optimization, and prediction include to predict optimal next actions for sales representatives (e.g. when to offer discounts), personalize customer journeys across multiple sales channels, predict demand to support pricing decisions and supply chain optimization.
Similarly, in Human Resources (HR), multiple use cases use natural language processing technologies, including automating screening processes in recruitment and providing targeted appropriate training courses, providing analytic and predictive support to staff compliance monitoring e.g. spotting inappropriate behavior such as bullying in internal emails.

The explosion of data within businesses and societies is a key factor in the recent rise of AI. The rapid growth in data both makes AI possible and necessary. Data of multiple types have proliferated wildly in the past several decades.

In 2020 Prof. Isa Ali Ibrahim Pantami, Minister of Communications and Digital Economy, lauded the immense benefits derivable from the establishment of the National Centre for AI and Robotics (NCAIR) in Nigeria. The Nigeria Federal Government has indicated its willingness to partner with the Nigeria Computer Society on Artificial Intelligence to boost the nation’s economy.

During the maiden edition of the Artificial Intelligence Summit, Pantami, who was represented by the Director, National Centre for Artificial Intelligence and Robotics, Yau Garba, said research on AI and other emerging technologies was ongoing with researchers in various universities across the country. He said, “AI is ushering new ways of planning, traveling, and working. Its innovative use has been changing lives for the better, creating new jobs and augmenting some existing ones while entirely replacing others’’.

Fom Gyem writes from Wuye District Abuja

The post Nigeria’s Steady Growth on Artificial Intelligence appeared first on Tech Digest News.

TECH DIGEST – Aboluwarin Kitan-David, who is the Co-Founder, Future Academy Africa, in this interview speaks on the future of technology as well as reasons why companies and government agencies must adopt it. Ugo Aliogo presents the excerpts:

Can you give an insight into your computer coding and technology startup journey?

I graduated from Federal University of Science and Technology Akure (FUTA) with a degree in Computer Science. But before gaining admission to the University, I had already started coding. I think my coding dream started before the technology startup journey. The journey of coding first started with my father, a missionary and a computer enthusiast, though because of financial constraints we could not afford a computer in the house.Then passion for computers grew stronger and I wanted to learn how they worked, how a clicked button shuts down hardware and so on. So, I started visiting cyber cafes to surf the internet overnight while I was still in secondary school.

But after secondary school, I didn’t get admission into University, because I didn’t pass the Unitary Tertiary Matriculation Examination (UTME) until after four attempts. During those periods of staying at home, I got a form to go to NIIT and I learnt how to code in Java and then I started teaching myself how to code other things. Then I gained admission to study computer science. University provided me with the opportunity to attend a lot of events. Every month, I visited CCHub in Yaba, Lagos. So I got to attend a lot of events both the ones that were necessary and not necessary and I got fascinated about the technology dream. Then I joined the Institute of Software Practitioners of Nigeria (ISPON) and I met the chairman of ISPON, Chris Uwaje, who introduced me to Omobola Johnson who was the former Minister for Communication who encouraged me. In that same meeting, I met the head of Nokia Research Center West Africa (Moses Acquah) too and became an ambassador for the company, running programmes for tertiary institutions in Nigeria, Benin Republic and Ghana. By my third year at the university, I was already running educational programmes and this experience influenced my decision to organise training in web design, web development and basic programming for students in my free time too. I was never interested in school politics, but the drive to do more led me to contest for the position of computer science president (NACOSS) and I won—during my one-term tenure, I ensured that there was a training programme every week.

After I graduated from school, I started learning about the Future of Work technologies: artificial intelligence, data science, UI/UX and machine learning and launched the #futureofworkAfrica initiative with stories to drive people to the reality of what the workplace of the future will look like.

My passion now is to see how to train young people in soft skills that will prepare them for job offers with international companies and I have worked to achieve this under different umbrellas.

What is the future Africa Initiative all about?

There are some technologies known as the future of work technologies. The major leading technologies are artificial intelligence, machine learning, data science, and UIUX. These are technologies that will transform the workplace totally. So I started talking about those technologies online and organising open house conversations with the name, future of work Africa. Later on, I decided to build a structure around it instead of just organising open house conversations and writing stories on social media; which is what we have today as Future Academy Africa(FAA).

At future Academy Africa, what we do is to train Africans in future of work technologies and subsequently get them jobs. Most of the times, I describe it to people as a three-year Doctorate Degree journey. It doesn’t mean you will not get a job in three years, but three years is what I always advise you to put in. In your 8th or 15th months, you might start getting jobs. But don’t let your aim be getting jobs in the first three years because there are jobs and jobs will come. Recently, I secured a job for one of our students at FAA where he is paid N23 million yearly. The future of work academy Africa started off from my campaign of talking about the future of work technologies and using #futureofworkafrica on Instagram. I’m also in conversations with state governments on the need to adopt it and sit on the board of Non-Governmental Organisations (NGOs) that are doing similar training.

Do you think there is increased adoption of Artificial intelligence and machine learning among organisations and government agencies?

I will say no because most of the organisations on the continent still rely on manual labour to execute their tasks. Although industries such as financial institutions are using AI and machine learning, at a slow pace if I may add.

I still think the adoption is slow because there are a lot of youths that are ready to work. For instance in construction, China has machines and robots that can build skyscrapers within a short period, then you ask yourself why aren’t construction companies in Nigeria buying those machines, because there are a lot of youths who want to work and the construction companies prefers to use these youths to work than relying on those machines.

Even construction companies in Africa that says they don’t want to use the future technologies; they will have no choice in the end, but to buy those machines. Now when they buy these machines, the 200 boys that come to the construction site will become jobless. It is behavioural psychological when people are unemployed that is when they will fall back to crime. So the adoption rate of AI and other future technologies by corporations in Nigeria is very slow though they are adopting it.

The fact remains that regarding the future of work, we don’t have a choice; we are going to adopt it. For instance, I consulted for a company and built software for a department in the company. The software replaced the work of 200 staff in that department. So now they are looking for way to lay off those 200 staff off because the software does what they do basically.

One unique feature about future technologies is that they don’t go on break, request for payment or salary increment. They have been designed to deliver on every task given to them. When it comes to making industry decisions, it is machine over humans, but then humans need to step up their game because it is humans that will create these machines. One of our alumni just moved to Canada in December, 2020, with AI skill and he is receiving 80,000 Canadian dollars yearly as starting salary because he is a foreigner, if he was Canadian, he will be paid more. But in another five years, he will be building a machine that we are using, so there are different ways to enter this, you don’t have to code that is why there is UIUX, if you can code there is AI, and to code in AI, you need to understand mathematics very well. For UIUX you don’t need to write any form of code, Data science, you might code or might code depends on what aspect of data science you’re using, so it is just for everybody.

What should state governments do to drive increased adoption of AI and machine learning among youths?

States governments need to give soft landing to the youths. For instance, the Edo state government is doing increased adoption of AI and machine learning in the state, but it is a partner led initiative. The state government collaborated with coven labs and requested them to do certain things for them. The state government was not setting parameters for the coven labs on what to do or not to do. So when you allow a partner to lead the initiative and then you support that is the first step. Then the second step is providing the soft landing for the individuals coming to learn. So if a group of 100 youths are coming to learn in Bauchi State, and these youths are coming every day to learn, we should know that they are humans and they need to eat. My experience with this initiative is when 100 youths come during the opening day and first class, by the second week, we have like 35 persons in class and the rest have dropped out not because they don’t like what they are learning, but because they need to survive. Giving them a stipend to survive on will go a long way, it will make the students more dedicated and everybody more aligned. The work is exciting and the person wants to come everybody, but when if there is nothing to survive on the people will not come. Government should look at giving support and autonomy to partners lead the initiative. The participants should also be given some soft landing.

What is the long term plan for future Academy Africa?

The long-term plan for the academy is to have a university campus that can accommodate 4,000 students at every point in time go through a programme whether they are PhD, University undergraduate or SSCE students. These students will be well trained and become sorted after by companies because of the curriculum and the way things are done.

The future Academy African is running the future of work campaign on the continent with a desire to sensitise people, drive adoption for government and sensitise large corporations to automate some of their processes so that we get Africa to be part of the global trend, even if it is through outsourcing. For instance, government will never come to my school, but while people are in my school, or outside my campus, I want to drive a conversation that will make government to start automating things such as Local Government Identity card. Even if they are talking about AI in the world, even if Africa is not the one creating the framework, they should think of Africa as a place they can get the job done. If someone else is coming up with this initiative, 400 of us are not enough.

What do you think is the future of Africa especially Nigeria, in the development of technology startups; can we compete with some of the technology capital of the world?

We cannot compete with the technology capitals of the world, but we can be unique in our own way. The place where I feel we have strength to compete is outsourcing. In the 1970s, China didn’t try to be like America, there had conversations to carve out a niche, and decided to position herself in such a way that when America wants to create computer and electronic devices, they go to China. So Africa can position herself in the technology talent, and human resources in such a way that if they require technology software engineers and designers in UIUX, AI, Data science and machine learning, they can come to Africa and I think with that we can compete. Think Talent, Think Africa, Think and Outsourcing.

The post Nigeria Has Comparative Advantage In Technology Talents – Kitan David appeared first on Tech Digest News.

TECH DIGEST – MIT researchers have developed a model which can distinguish people with asymptomatic Covid-19 from healthy people, using the sound of their coughs. These differences are impossible to detect with the human ear.

Previously, researchers have trained algorithms to detect conditions such as pneumonia and asthma based on basic phone recordings of coughs.

The MIT team had been developing models to detect signs of Alzheimer’s using these recordings but as the corona virus pandemic emerged, the researchers decided to investigate whether a similar tool could be applicable to Covid-19.

In April, they collected as many recordings of coughs as possible, setting up a website where anyone could submit recordings of forced coughs, along with information about their symptoms and test results.

The researchers collected more than 200,000 forced-cough samples including 2,500 from people confirmed to have Covid-19.

These Covid-19-positive recordings and 2,500 recordings from healthy participants formed a dataset.

The scientists used 4,000 to train the model and the remaining 1,000 to test it.

When fed new recordings of coughs, the model was able to identify people diagnosed with Covid-19 with 98.5 per cent accuracy, and 100 per cent accuracy for people with asymptomatic Covid-19.

The team is partnering with several hospitals to collect a larger and more diverse set of cough recordings to boost the accuracy of the model.

They are also working with a company to incorporate the AI model into a user-friendly pre-screening app, which could provide a free, quick and non-invasive tool for identifying people with asymptomatic Covid-19.

“The effective implementation of this group diagnostic tool could diminish the spread of the pandemic if everyone uses it before going to a classroom, a factory, or a restaurant,” said Professor Brian Subirana, director of the Auto-ID Laboratory.

The post AI Model Detects Covid-19 from Sound of Coughing appeared first on Tech Digest News.