– Dr Minh Dinho
What are the most anticipated trends in artificial intelligence (AI) and their implications for the Vietnamese labor market? dr. Dinh Ngoc Minh, senior lecturer in AI and software engineering at RMIT Vietnam, shared his insights.
Powerful Language Models
The rapid growth in the use of wearable devices such as smartphones, tablets and laptops has provided businesses with new interfaces to connect with their customers through websites and social media platforms.
According to Dr. Esteban Ortiz-Ospina of Our world in data, there are 2.4 billion Facebook users, while other social media platforms, including YouTube and WhatsApp, also have more than a billion users each. These services increase consumer-generated data and emphasize the need for natural language processing capabilities.
Natural language processing (NLP) is a branch of computer science that focuses on creating, analyzing and interpreting human language to perform tasks such as sentiment classification, machine translation, handwritten character recognition, speech recognition, to name a few. We have seen major breakthroughs in NLP in recent years with the most powerful language models such as GPT-3, which can produce creative fiction, develop computer code and summarize a large corpus of research literature. Yet the full potential of NLP is still being explored.
In the coming years, we will continue to see further progress in NLP, especially in dealing with data in a multilingual context. The effects will be compelling in areas such as finance, services and entertainment, where target audiences are no longer limited to specific regions or languages. NLP will continue to evolve to meet the need to gather consumer opinions, understand customer demand, and personalize work and play options.
dr. Dinh Ngoc Minho
Advanced computer vision technology
AI is rapidly maturing in imaging and computer vision with the recent advent of deep learning techniques, unsupervised learning, and active learning.
Since the days when we were able to classify objects such as letters, animals and household items, computer vision has come a long way and has significant implications for our daily lives. For example, AI solutions can address some of the biggest challenges in the healthcare segment, especially in the field of medical imaging. Typically, the analysis of medical images requires a great deal of time and effort from experienced radiographers and physicians. Today, doctors don’t have to scan through thousands of CXR and CT majors, while preferring to use an automated AI system to help them identify the most important ones so they can speed up their decision-making.
We will continue to see breakthroughs in AI-based computer vision solutions, and the ramifications will extend beyond the realm of medical imaging. Computer vision will showcase successes in autonomous driving, smart manufacturing and smart homes and cities, through a new discipline called ‘computer vision on the edge’. Real-time photos and images will be processed, classified and characterized much closer to where they are generated to support the real-time decision-making process.
Reinforced learning
This is another important domain of AI and machine learning, where AI scientists focus on decision making using reward-based training, rather than optimizing loss functions.
In a nutshell, reinforcement learning models a realistic environment in which agents examine and adjust their behavior to maximize rewards (and/or minimize punishments). This learning model works because it mimics how we learn in real life, where we don’t always make the right decision or perform a completely safe act. While we don’t necessarily perform arbitrary actions and rely entirely on a trial and error process, it is important that we consider the arbitrariness of the real world when training our artificial intelligence agents. As a result, reinforcement learning techniques play a key role in developing autonomous robots, including autonomous vehicle, establishing trading strategies, and validating complex decision-making procedures.
Compared to supervised learning, in which deep learning has led to significant breakthroughs, reinforcement learning has yet to shine. However, as we expect artificial agents to make complex decisions while adhering to long-term goals, reinforcement learning will continue to be one of AI’s most exciting trends.
Generative AI
The emerging trend of generative AI includes machine learning methods, including NLP, that aim to generate realistic and potentially original artifacts by learning features and content from domain-specific data.
Generative AI algorithms can, for example, compose pieces of music, generate fictional literature or create paintings when specific subjects are taught, and therefore have a significant impact on education, creativity and media.
Gartner views generative AI as a strategic AI technology trend for 2022 and will grow in the coming years. Among the potential technologies, generative language models, aimed at generating natural-sounding content, have impactful applications in marketing, customer service, and personalized education. Generative Adversarial Networks (GANs) can be used for fraud detection, synthetic data generation and risk factor modeling, particularly in banking and investment services. Last but not least, generative AI solutions can solve generic problems and adapt to different situations and contexts, thus promising a roadmap for artificial general intelligence (AGI).
Strong demand for AI experts in Vietnam
The Vietnamese government has issued a national strategy for research, development and application of AI up to 2030, with the aim of turning Vietnam into an innovation and AI hub in ASEAN. To achieve that goal, the Ministry of Science and Technology (MoST) recently launched the Vietnam-Australia Artificial Intelligence Cooperation Network (Vietnam-Australia AI Network). According to Vietnam-Australia AI Network, about a million more IT staff is needed in the country by 2030, and the demand for AI talent is expected to increase continuously.
The Vietnamese job market has shown significant demand for top AI talents to respond to the national strategy above. A Vietnam IT Market Report by TopDev indicates that AI and Machine Learning engineers can receive the highest average monthly salary of IT engineers, up to US$3,054, from the 2nd quarter 2021.
With the current trend in the Vietnamese labor market, the highest paying positions in IT require special skills such as data analytics, DevOps, Machine Learning or AI. This trend in the job market shows how important it is for IT professionals to upgrade to AI and Data Science majors to stay competitive.
RMIT University in Vietnam recently launched the Master in Artificial Intelligence program for those wishing to advance their career in AI by exploring practical components of developing AI applications and platforms and understanding the role ethics and social responsibility play in the future of technology.
RMIT has been ranked among the world’s best in AI and Image Processing research by Excellence in Research for Australia (ERA) – Australia’s national research evaluation framework and rated ‘above the world standard’ by the Australian Research Council. Students in Vietnam will participate in one of the most specialized AI programs in Australia and will learn from international experts.
About the author
With a Ph.D. in Computer Science from Monash University, Australia, Dr. Dinh Ngoc Minh’s research expertise lies in computational science, high-performance computing and AI. His expertise has been developed through his previous positions at Monash University, the University of Queensland and the Queensland Cyber Infrastructure Foundation.
dr. Dinh’s recent research projects developed a Deep Neural Network debugger, an optical character recognition (OCR) pipeline for recognizing and transcribing Vietnamese doctor’s handwriting, and a scalable machine-learning pipeline for improving computational modeling techniques.
