Many deep learning models struggle to see the world in which there are objects and the relationships between them. Most models do not understand the entangled relationships between individual objects. Without knowledge of these relationships, a robot designed to assist someone in a kitchen would have difficulty following commands such as “grab the spatula on the left side of the stove and place it on the cutting board.”
In an effort to solve this problem, MIT researchers have developed a model that understands the underlying relationships between objects in a scene. Their model depicts individual relationships one by one and combines these representations to describe the overall scene. This allows the model to generate more accurate images from text descriptions, even when the scene contains multiple objects arranged in different relationships to each other.
This work can be applied in situations where industrial robots need to perform complex, multi-step manipulation tasks, such as stacking items in a warehouse or assembling devices. It also brings the field one step closer to enabling machines that can learn from and interact with their environment more like humans do.
When I look at a table, I cannot tell there is an object in the XYZ location. Our minds do not work that way. In our minds, when we understand a scene, we really understand it based on the relationships between the objects. We think that by building a system that can understand the relationships between objects, we can use that system to more effectively manipulate and change our environments.
– Yilun Du, PhD Computer Science and Artificial Intelligence Laboratory & Co-Lead Author
The framework the researchers developed can generate an image of a scene based on a text description of objects and their relationships, such as ‘A wooden table to the left of a blue stool. A red bench to the right of a blue stool.”
Their system would break these sentences into two smaller pieces describing each individual relationship, then model each part individually. Those pieces are then combined through an optimisation process that generates an image of the scene.
The researchers used a machine learning technique called energy-based models to represent the individual object relationships in a scene description. This technique allows them to use one energy-based model to encode each relational description, then assemble them in a way that infers all objects and relationships.
The system also works in reverse: with an image, it can find text descriptions that correspond to the relationships between objects in the scene. In addition, their model can be used to edit an image by rearranging the objects in the scene to match a new description.
The researchers compared their model with other deep learning methods that were given text descriptions and tasked with generating images showing the associated objects and their relationships. In any case, their model outperformed the baselines.
They also asked people to evaluate whether the images generated matched the original scene description. In the most complex examples, where descriptions included three relationships, 91% of participants concluded that the new model performed better.
While these initial results are encouraging, the researchers would like to see how their model performs on more complex real-world images, with noisy backgrounds and objects blocking each other. They are also interested in eventually incorporating their model into robotic systems, allowing a robot to derive object relationships from videos and then apply this knowledge to manipulate objects in the world.
The Joint Committee Meeting (JCM) on Information and Communications Co-operation between the Government of the Republic of Singapore and the Government of Malaysia had a strong digital theme. Both countries discussed digital transformation efforts and explored areas where bilateral digital cooperation could advance post-pandemic recovery.
Both parties discussed issues relating to enabling trusted data flows between the two countries, and to better connecting the respective innovation and technology ecosystems to support businesses and start-ups. In addition, both are committed to implementing projects to demonstrate the benefits of cooperation in this rapidly developing digital domain to support the recovery of our respective economies.
In addition, the JCM also discussed how media production, distribution and consumption are being disrupted by technologies and online platforms, including growing volumes of information and the rapid spread of falsehoods.
The JCM is a platform of increasing importance, to deepen the bilateral cooperation between Singapore and Malaysia. The pandemic has driven many companies to digitally transform and seize new opportunities. Through the JCM, we have initiated meaningful digital cooperation projects to increase the adoption and interoperability of digital technologies in both countries. Our collaboration will serve as a springboard to enhance connectivity between our businesses and people and to support our recovery from the pandemic.
– Yong Ying-I, Permanent Secretary, Ministry of Communications and Information of Singapore
Malaysia continues to embrace digital technology and develop unique technologies and business models to assist the country in establishing new development engines. With the growth of the digital economy, the Fourth Industrial Revolution, and ever-evolving technology, Malaysia looks forward to exploring potential collaboration in this sector in the future. Malaysia is stepping up efforts to assist Micro, Small and Medium-Sized Enterprises (MSMEs) and business owners in adopting digital technology, and will continue to advance plans to establish an inclusive and progressive digital economy for all.
Malaysia has a sustainable and solid economic foundation, comprehensive business-ready environment and dynamic skilled workforce. As an attractive cost-competitive investment location in the region, she is fast becoming a preferred centre for shared services and leading technology industries. Singaporean companies who are looking to expand into Malaysia should pay attention to the launch of the Future 5 Strategy, and evaluate how their businesses can fit into this plan in order to anchor a foothold into the market.
The five industry sectors that have been identified as key drivers are AgTech, HealthTech, Islamic Digital Economy and FinTech, CleanTech and EduTech. These industries are based on the strategic national industries for digitalisation and have also been mapped to Malaysia’s national priority sectors.
As reported by OpenGov Asia, Singapore and the Republic of Korea (ROK) have also launched negotiations on a new Korea-Singapore Digital Partnership Agreement (KSDPA) last year. The agreement seeks to deepen bilateral cooperation in new emerging digital areas, such as in personal data protection and cross-border data flows, digital identities, fintech, as well as Artificial Intelligence (AI) governance frameworks. It also aims to support and foster greater collaboration between both countries’ SME communities in the digital economy.
Recently, Singapore and ROK have concluded negotiations on the Korea-Singapore Digital Partnership Agreement (KSDPA). The KSDPA will be Singapore’s fourth Digital Economy Agreement (DEA), and the first with an Asian country. The agreement will deepen bilateral cooperation in the digital economy between both countries, by establishing forward-looking digital trade rules and norms to promote interoperability between digital systems. This will enable more seamless cross-border data flows and build a trusted and secure digital environment for our businesses and consumers.
The KSDPA is part of a series of DEAs that Singapore has embarked upon. These agreements are an inter-agency effort led by the Ministry of Trade and Industry, Ministry of Communications and Information, and the Infocomm Media Development Authority, to advance collaboration in the digital economy and enhance digital connectivity.
With the Omicron variant making its way into the local community, the HKSAR Government announced tightening COVID-19 measures to contain the epidemic. The public should stay vigilant to maintain good personal hygiene at all times to strengthen individual defence against the pandemic.
At present, some public facilities such as doorknobs in public toilets and lift buttons have poor cleanliness and can become breeding grounds for viruses and bacteria, thus posing a threat to public health.
An interdisciplinary research team from The Hong Kong Polytechnic University (PolyU) has successfully developed the world’s first “anti-virus 3D printing material” (material) that can kill the COVID-19 virus on surfaces as well as most common viruses and bacteria. The main component of the material is resin, added with anti-viral agents such as cationic compounds, to damage the membrane of the virus and destroy its structure to kill the virus and bacteria.
Dr Kwan Yu Chris LO, Associate Professor of PolyU’s Institute of Textiles and Clothing, who led the research team, said that laboratory tests confirmed the material can kill 70% of the COVID-19 virus and other viruses/bacteria surviving on a surface within two minutes; eliminate over 90% of viruses within 10 minutes, and terminate almost all viruses and bacteria on a surface in 20 minutes.
Dr Lo stated that this material is a resin material with high anti-virus performance. Using 3D printing technology, it can be produced in different forms catering to different needs. It is therefore highly flexible and can be used extensively in public facilities to provide epidemic prevention support to the community.
The team has already applied patent of this technology and application and will use it for commercial purposes in future.
In the past year, with the support of the laboratory of PolyU’s University Research Facility in 3D Printing (U3DP), the research team has collaborated with the Home Affairs Department, the Hong Kong Wetland Park and an environmental organisation to produce recycling bin handles, toilet doorknob covers, lift buttons, braille boards and more, in order to conduct further tests and trials of the effectiveness and durability of the material in killing viruses.
Prof. Chi-wai KAN, a member of the research team and Professor of PolyU’s Institute of Textiles and Clothing stated that even after use for a year, not only is the handle on the recycling bin still in good condition, no COVID-19 virus, Escherichia coli and Staphylococcus aureus are detected on the handle’s surface.
He noted that this proves that the efficacy rate of the material only diminishes gradually after three years of use, and is effective in fighting against viruses and bacteria. Since the material kills viruses via physical means, it can still exert the same effect on mutant viruses.”
Prof. Kan added that because the disinfection components of the material are embedded in the products rather than coated on the surface, daily cleaning with disinfectants such as bleach does not compromise its anti-virus performance.
The research team will also collaborate with the Sham Shui Po District Office to produce doorknob protective covers for over 100 unmanaged “Three-Nil” buildings in the district and install these covers on doors frequently used by residents, so as to reduce the risk of virus transmission in buildings.
The team hopes to apply the material to primary and secondary schools, healthcare facilities, and public transportation systems.
To analyse potentially cancerous lesions in mammography scans, Computer engineers and radiologists at Duke University have developed an Artificial Intelligence (AI) platform to determine if an invasive biopsy is necessary. Unlike its many predecessors, this algorithm is interpretable, meaning it shows physicians exactly how it came to its conclusions.
Rather than allowing the AI to freely develop its own procedures, the researchers trained it to locate and evaluate lesions just like an actual radiologist would be trained. The AI could make for a useful training platform to teach students how to read mammography images. It could also help physicians in sparsely populated regions around the world who do not regularly read mammography scans make better health care decisions.
If a computer is going to help make important medical decisions, physicians need to trust that the AI is basing its conclusions on something that makes sense. We need algorithms that not only work but explain themselves and show examples of what they are basing their conclusions on. That way, whether a physician agrees with the outcome or not, the AI is helping to make better decisions.
– Joseph Lo, Professor of Radiology, Duke University
Engineering AI that reads medical images is a huge industry. Thousands of independent algorithms already exist, and the FDA has approved more than 100 of them for clinical use. Whether reading MRI, CT or mammogram scans, however, very few of them use validation datasets with more than 1000 images or contain demographic information. This dearth of information, coupled with the recent failures of several notable examples, has led many physicians to question the use of AI in high-stakes medical decisions.
The researchers’ idea is to build a system to say that this specific part of a potentially cancerous lesion looks a lot like this other one. Without these explicit details, medical practitioners will lose time and faith in the system if there is no way to understand why it sometimes makes mistakes.
The researchers trained the new AI with 1,136 images taken from 484 patients at Duke University Health System. They first taught the AI to find the suspicious lesions in question and ignore all of the healthy tissue and other irrelevant data. Then they hired radiologists to carefully label the images to teach the AI to focus on the edges of the lesions, where the potential tumours meet healthy surrounding tissue and compare those edges to edges in images with known cancerous and benign outcomes.
This is a unique way to train an AI how to look at medical imagery. Other AIs are not trying to imitate radiologists; they are coming up with their methods for answering the question that is often not helpful or, in some cases, depend on flawed reasoning processes. After training was complete, the researchers put the AI to the test. While it did not outperform human radiologists, it did just as well as other black box computer models. When the new AI is wrong, people working with it will be able to recognise that it is wrong and why it made the mistake.
As reported by OpenGov Asia, a new report showed that Artificial Intelligence (AI) has reached a critical turning point in its evolution. Substantial advances in language processing, computer vision and pattern recognition mean that AI is touching people’s lives daily—from helping people to choose a movie to aid in medical diagnoses.
In terms of AI advances, the panel noted substantial progress across subfields of AI, including speech and language processing, computer vision and other areas. Much of this progress has been driven by advances in machine learning techniques, particularly deep learning systems, which have leapt in recent years from the academic setting to everyday applications.
On behalf of the Australian Government, the Australian Renewable Energy Agency (ARENA) recently announced up to AU$ 40 million in funding to support research and development (R&D) that aims to support the achievement of the Government’s ultra-low-cost solar stretch goal.
The R&D funding round will build on ARENA’s previous R&D investment into solar PV and will seek to support projects that align with ARENA’s Solar 30 30 30 target of 30 per cent module efficiency and 30 cents per installed watt at utility-scale by 2030. ARENA invites applications that can materially reduce the levelised cost of solar PV by 2030 across two streams:
- Stream 1 – Cells and Modules: Building on Australia’s leading track record of R&D and innovation in solar cells and modules
- Stream 2 – Balance of system, operations and maintenance: Seeking to broaden the approach to accelerate innovation that can drive down the upfront and ongoing costs of utility-scale solar PV in the field.
ARENA currently expects to allocate up to $20 million in total funding to each of the two streams.
Ultra-low-cost solar was recently added as a priority technology in the Australian Government’s latest Low Emissions Technology Statement (LETS), which set a stretch goal of $15 per megawatt-hour, roughly a third of today’s cost.
Ultra-low-cost solar will be a key input to scaling up production of low-cost green hydrogen, in support of the LETS hydrogen goal of “H2 under $2”, as well as the key to unlocking other decarbonisation pathways for heavy industry including low emission materials such as green steel and aluminium.
As part of ARENA’s new Investment Plan and in support of ultra-low-cost solar, ARENA has set an ambitious target of ‘Solar 30 30 30’, to improve solar cell efficiency to 30 per cent and reduce the total cost of construction of utility-scale solar farms to 30 cents per watt by 2030.
Since 2012, ARENA through its R&D Programs has committed close to $105 million in grant funding to over 70 projects. In addition to this, ARENA has also supported the Australian Centre for Advanced Photovoltaics (ACAP with AU$ 84 million of funding over 10 years.
The CEO of ARENA stated that the pioneering work of Australian solar researchers will be key to driving cost reductions and improving solar cell efficiency.
He noted that Australia’s solar researchers have been leading the world for decades. Thirty years ago, UNSW researchers invented the PERC silicon solar cell, technology which today is the foundation of more than 80% of the world’s solar panels.
That work continues through ACAP, the universities and CSIRO as well as clean energy start-ups. Just a few months ago a start-up founded by former UNSW students and now based in Sydney, created the world’s most efficient solar cell. This AU$ 40 million R&D funding round will support Australia’s solar researchers and industry to get behind the target of Solar 30 30 30 and drive the innovation that will deliver ultra-low-cost solar, he added.
Ultra-low-cost solar will be a vital component in helping Australia move towards a lower cost, largely renewable electricity system and achieve the goal of net-zero emissions by 2050. Expressions of Interest for the Ultra Low-Cost Solar R&D Funding Round will open in February 2022 with applications due by 5 pm AEST Monday 11 April 2022.
Cross-border e-commerce has become an important driving force for stabilising China’s foreign trade and played a positive role in helping small and medium-sized enterprises hedge against the negative impact of the COVID-19 pandemic.
The import and export volume of China’s cross-border e-commerce totalled Yuan 1.98 trillion (US$ 311.5 billion) in 2021, up 15% year-on-year. E-commerce exports stood at Yuan 1.44 trillion, an increase of 24.5% on a yearly basis. As a new form of foreign trade, cross-border e-commerce has witnessed rapid growth in China by making full use of its advantages in online trading and contactless delivery since the pandemic outbreak.
Digital transformation has emerged as a key pathway to mitigate the impact of the pandemic on traditional trade. More enterprises have attached great importance to cross-border e-commerce as it becomes a vital channel for foreign trade enterprises to open up new markets. Cross-border e-commerce breaks time and geographical barriers and enhances the digital management capacities of enterprises.
Digital tools and digital transformation are the key factors for global micro, small and medium-sized enterprises or MSMEs to survive and thrive in the unpredictable COVID-19 era. Relying on the resiliency of China’s supply chain, a leading Chinese cross-border B2B e-commerce platform has empowered global MSMEs with some capabilities like more data flow, a deeper understanding of customer demand as well as a more tailor-made product portfolio to help them succeed in the challenging business environment.
Relying on the resiliency of China’s supply chain, the platform has empowered global MSMEs with some capabilities like more data flow, a deeper understanding of customer demand as well as a more tailor-made product portfolio to help them succeed in the challenging business environment.
Cross-border e-commerce has become an important channel for China’s foreign trade during the pandemic period and accelerated the innovative development of foreign trade. The outbreak has posed a challenge to logistics and distribution. China’s cross-border e-commerce logistics companies have made efforts to ensure the timely delivery of commodities through charter flights and overseas warehouses.
Shopping via overseas live streaming services could offer detailed information about products to domestic consumers. Such services have gained wide popularity among the younger generation. There is an inevitable trend that more cross-border online retailers will cooperate with live streaming platforms.
New business forms and models, especially cross-border e-commerce, have become a vibrant force driving China’s foreign trade. They also represent an important trend in the development of international trade. China’s cross-border e-commerce has grown by nearly 10 times over the past five years. By both exports and imports, cross-border e-commerce has been expanding much faster than overall foreign trade, and its share in overall foreign trade has gone up significantly.
As reported by OpenGov Asia, Chinese President Xi Jinping pledged to support the development of key technologies while strengthening the regulation of the country’s tech giants as part of his strategy to expand the digital economy. The country needs to boost innovation in core technologies and step up research capabilities to achieve self-sufficiency as soon as possible. China also called for an acceleration in the development of high-speed, secure smart infrastructure that can connect all aspects of the online economy as well as for breakthroughs in key software technologies.
China has identified the digital economy as a key driver for growth over the next few decades and made achieving tech self-sufficiency a top national priority. To support that growth, Beijing has doubled down on funding for strategically important industries such as semiconductors and AI, while rolling out new legislation covering everything from data security to fair competition as part of efforts to bring the country’s once free-wheeling internet giants in line with the national agenda.
When it comes to digital democracy, democracy is the main idea, and digital is just an objective to assist democracy. Around the world, there is the other way of ideas that somehow democracy must give way to the public health measures, to counter disinformation measures. However, technology needs to adapt to the people’s will and the people’s norms, and people’s co-creation and real needs.
In authoritarian uses of technology, the main difficulty would be because of the lack of symmetrical communication. The real-time feedback of what is really going on is hampered. For example, if you can only download, it is more like television. If you can only download but there’s no way to upload, then emerging issues do not tend to get notified in time.
– Audrey Tang, Digital Minister of Taiwan
In Taiwan, the system has been successful in hearing younger people. A lot of the most impactful ideas came from very young people. To shorten the time that a genuinely good idea gets thought by a teenager or young people, and the time that it is understood by the senior people and implemented, is key to moving democracy forward. The younger people, because they are digital natives, they do not think that once every four years is sufficient to upload bandwidth, the latency is too high, they prefer to collaborate on a day-to-day basis.
When the coronavirus began spreading, Taiwan quickly established a mask map system that let people know if they could obtain masks if they went to certain pharmacies. The mask availability map was an idea from the civic technologists, not the government’s idea.
First, they already have a lot of experience building maps of this kind. All sorts of disaster response experiences, including earthquakes, typhoons, gas explosions, occupying of departments, various disasters, were met with this kind of real-time, map-based response by the civic tech people. The second reason is that people are very much willing to participate, because in Taiwan broadband is a human right. So, participating online does not cost any extra connectivity, money, for people.
In Taiwan, when people check-in the public venues, everyone chooses either to scan the QR code and send an SMS to 1922 (Taiwan’s 24-hour communicable disease reporting hotline), which is stored in their telecommunications carrier. But the venue owner learns nothing about their phone number. And the telecom carrier learns nothing about the venue code. de-centralized storage makes sure that nobody’s privacy gets compromised because the telecoms do not know what those digits mean.
There are two main reasons why Taiwan has changed from a very conservative to a democratic society. One is that the public service is really committed to working with the civil society leaders when it comes to gender mainstreaming in the gender equality committee to build the impact assessment, evidence-based projects together. And the civil society leaders always have one more vote than the ministers in the Gender Equality Committee.
The second reason is that the statistics, the dashboard, the gender impact dashboard just keep running. So even after the budgeted project runs its course, the gender impact it created is still being monitored for more than a decade for some projects now. Civil society is not just demonstrating against or protesting against something, it is demonstrating for something, demonstrating something works, and working with the people.
As reported by OpenGov Asia, Taiwan encouraged other nations to consider Taiwan’s example of open digital development and privacy safeguards in countering digital authoritarianism and affirming democratic values. To elaborate on the tools Taiwan has used to foster transparency and public trust, the key is to work not only for the people but with the people.
When it comes to remote learning, students often feel they are struggling alone. Studying in a community can address this, but it is not easy to create a sense of togetherness in a remote learning environment. Hence, the Singapore Institute of Technology (SIT) has introduced an innovative approach to help students find belonging amidst the pandemic.
Every individual has their own approach to learning. A surface learner, for example, mainly cares about achieving a grade or impressing someone, not about educational growth. The university can then provide personalised materials based on their learning profile to help them develop learning approaches suitable for higher education. For example, these could help a surface learner look past material achievements and learn for their personal growth.
Students at SIT have to complete the Freshmen Survey when joining the university to find out what kind of learner they are through a gamified platform named AdventureLEARN. Personalising the recommendation of educational content to a student’s profile and needs helps them to learn more effectively. Students have a limited attention span, so this window of time should be spent focusing on key areas.
– Associate Professor May Lim, Director, Centre for Learning Environment and Assessment Development, SIT
A learning community can be built while personalising online learning. A good example is QUEST – a platform featuring adaptive online courses. The platform helps SIT students get up to speed in core competencies such as Math, Physics, and Chemistry in preparation for university courses.
QUEST helps students through hints and advice indicated in pop-ups while they are answering questions. Learning with real-time feedback is a new method of education, complementing traditional methods such as watching short video lectures. Students are also provided with an online collaborative space to learn from one another. Working together can help reduce procrastination, which is a common challenge in remote learning.
Technology offers a range of other benefits on top of collaborative work and personalised information. Through AdventureLEARN, students earn virtual coins by completing assignments. Students can take part in team challenges as well, in which four to five students work as a group. Together, they can watch videos, collaborate to create learning resources or provide useful tips for well-being to earn more virtual coins.
applying ‘high tech’ alone is insufficient to transform education in this current climate. ‘High touch’ is needed for students to feel connected and supported. SIT believes in building a culture and an ecosystem where academic staff are equipped with skills to coach students effectively. While the e-learning platforms can help students personalise their learning and learn new concepts, the ‘high touch’ from academic staff is vital for supporting students who may be at-risk or struggling.
Analytics from e-learning platforms can help spot students who are struggling. By identifying coachable moments, SIT educators can reach out to such students to coach them on goals such as effective time management, improving group dynamics or better learning approaches. With high tech and high touch, a learning community can be built to facilitate effective learning for the future.
As reported by OpenGov Asia, students enrolling in the Singapore Institute of Technology (SIT) can now sign up for two new courses in Artificial Intelligence (AI) and digital supply chain. The Bachelor of Science in Applied Artificial Intelligence (AAI) and Digital Supply Chain (DSC) being launched in the new academic year are three-year direct honours programmes. AAI emphasises implementing artificial intelligence (AI) within software systems, while DSC focuses on emerging technologies in the digital transformation of the logistics and supply chain sector.