From spending P7.7 million for two artificial intelligence research and development projects in the last 10 years, the Department of Science and Technology (DOST), through its Philippine Council for Industry, Energy and Emerging Technology Research and Development (PCIEERD), is “trying to catch up” on the country’s AI R&D investment by bankrolling the implementation of nine “mission-driven” AI R&D projects amounting to almost P316 million.
“DOST started investing in its AI program [seven years ago], but even then, we have been trying to catch up, and within our available resources we are able to catch up,” said Science Secretary Fortunato T. de la Peña at the online launching of the country’s 10-year AI R&D program called AI Pinas on April 8.
De la Peña said the AI Pinas, or the DOST-PCIEERD Artificial Intelligence Program, represents the AI R&D roadmap from 2019 to 2029.
The five targets of the 10-year AI R&D framework include the universal and affordable access to AI infrastructure; upskilling the workforce; science-based solutions for socio-economic opportunities; unprecedented innovation for the industry sector; and enhanced policy support and stakeholder engagement.
The AI R&D grants provided by the DOST-PCIEERD totaled P328 million from 2018 to the present, according to Executive Director Dr. Enrico C. Paringit.
“To jumpstart the AI Pinas program, we supported 13 mission-driven projects in seven higher-education institutions and research and development institutions,” Paringit said at the same launching.
Paringit added that, to date, “almost P340 million had been poured out and we look forward to reaching a critical mass of researchers developing AI capabilities, applications, and services and the rapid delivery of responsive AI-based solutions.”
“To turn this roadmap into reality,” Paringit said the DOST-PCIEERD is allocating at least P70 million to support AI-related researches in its call for R&D proposals for 2023.
“PCIEERD is investing P40 million and will ask the DOST at least P30 million to support some of these AI-related researches as part of our core or platform technologies or the applications or sectoral support or demonstrations,” Paringit added.
Investing harder and faster on AI R&D
AI had been among the bottom five of the 21 sectors, ranking either 20th or 21st for R&D budget allocation since 2011.
“We actually started quite slow so we are acting on that momentum, but we want to improve this standing and we are really investing hard [on AI R&D] in a fast-paced manner,” Paringit explained.
At the virtual launch of AI Pinas, the DOST-PCIEERD presented the nine new AI R&D projects and their leaders and heads of the implementing institutions, which include the DOST-Advanced Technology Science Institute (ASTI), University of the Philippines Mindanao (UPMin), De La Salle University (DLSU), University of the Philippines Los Baños (UPLB) and Caraga State University (CarSU).
“I am very proud of the variety of these projects in scope and scale, anywhere from environmental protection to agriculture to education,” Paringit said.
The five educational institutions each received a P3-million Dell Poweredge T640, a high-performance computing machine designed for handling intensive data processing activities. They will then also serve as the AI hub in their respective regions.
Paringit called on researchers to develop mission-driven programs to help the industry and find breakthroughs that will make use of the computing resources that are now available in these institutions.
“For R&D proposals, we have actually identified a few of them, including AI robotics solutions for emerging needs, enhancing disaster risk-reduction response, and improving infrastructure management,” Paringit said.
DOST-PCIEERD will accept proposals for “game-changing innovations” from public and private universities and colleges, research and development institutes, R&D consortia, nonprofit laboratories, and other public or private non-profit science and technology institutions in the country from May 3 to June 3, 2021 through the DOST Project Management Information system portal www://dpmis.dost.gov.ph.
Pushing for mission-driven AI projects
First is the Autonomous Societally Inspired Mission Oriented Vehicles (Asimov) program of the DOST-ASTI and UPMin.
Composed of two component projects, the Asimov program will develop AI-enhanced, mission-driven robots working autonomously or with humans to help address the needs of society. Its initial phase calls for developing and innovating on key functional modules of intelligent mobile robots: sensing, actuation, control, navigation and communications.
The UPMin and DOST-ASTI will also spearhead the Philippine Sky Artificial Intelligence Program (SkAI-Pinas) with the Automated Labeling Machine-Large-Scale Initiative as its main research component.
SkAI-Pinas will bridge the gap between the availability of massive remote sensing data in the country and address the lack of a sustainable technology-based framework to facilitate their widespread processing systematically and effectively.
Fifth is the DLSU’s Intelligent Structural Health Monitoring via Mesh of Tremor Sensors (meSHM) project that will develop a low-cost, wireless structural health monitoring system with visualization for buildings, bridges or metro rail systems utilizing the Internet of Things (IoT) technology and mesh networks.
The project will pave the way for a more complete data collection and analysis for upgraded studies and policies on disaster preparedness involving vertical and horizontal structures in the country.
The DLSU will also implement the Development of Multi-lingual Chatbot for Health Monitoring of Public-School Children Project, a system that can interpret audio input and converse with in Filipino and Visayan languages.
The project will develop speech and natural language processing models that can provide appropriate and intelligent responses in the form of questions or suggestions.
The information gathered will then be extracted to update the health database of the students. Health analytics and visualization of these data will be provided for decision-making.
The seventh project, also under the UPMin, addresses the global threat of anthropogenic marine debris using a towed camera system for marine litter monitoring.
Developing a simple, cost-effective technology to monitor and quantify the marine litter in shallow coastal areas will help protect the environment and reduce marine pollution.
The UPMin will base its technology on existing towed optical camera array system for deep-sea monitoring and redesign and improve this by adding sensors and cameras.
It will also have a built-in image processing and deep learning or machine learning capability to identify marine litter, compute and map the area covered by the litter and build models for predicting scenarios.
Meanwhile, to help identify and classify materials and their capacity and performance, the UPLB will develop an automated software that accepts values from a standard Impedance Spectrometer and uses a machine learning algorithm to identify electrical, mass and temperature parameters by looking into the time series plot and plot library.
This also involves properly fitting a spectrum with sufficient parameters that minimizes common errors in existing numerical fittings.
The program will be integrated with a simple interface where the user can just input the values or parameters. The academe and industries involving electronics, semiconductors, food, medicine, and agriculture will benefit from this project.
Lastly, using an IoT sensor network and deep learning, CarSU will design and develop an intelligent traffic control and management system, which can monitor traffic in an area by using various devices to measure such parameters as flow, density, volume, headway, waiting time, throughput and even pollution.
The system’s base station will be established and equipped with intelligent behavior and direct policy search capabilities using reinforcement learning to automatically and efficiently manage traffic and avoid congestion.
