[ad_1]
Analysis outcomes have potential for dual-use functions, for leverage by each the trade & defence
SINGAPORE – Media OutReach – 26 July 2023 – AI Singapore (AISG) introduced at the moment that it’s going to award as much as S$20 million, by means of the Sturdy AI Grand Problem and the AI for Supplies Discovery Grand Problem, to help 5 multidisciplinary groups which is able to deal with challenges associated to the growing use of AI in rising functions.
AISG launched each Grand Challenges in December 2022 in collaboration with the Future Methods and Expertise Directorate (FSTD), MINDEF Singapore, and DSO Nationwide Laboratories (DSO).
Every Problem is a first-of-its-kind and represents a serious milestone in analysis collaboration between the academia and defence ecosystems. The analysis outcomes could have the potential for dual-use functions and might be leveraged by each the trade and defence to profit Singapore.
The grant awards have been introduced by Mr Heng Chee How, Senior Minister of State for Defence on the Joint Grant Awards Ceremony.
In his speech, Mr Heng stated, “The development of Synthetic Intelligence (AI) will open up new potentialities and MINDEF/SAF is poised to harness its energy to construct necessary capabilities for our future. MINDEF’s strategy is to collaborate extensively with native and world companions from analysis institutes, and institutes of upper studying and trade. MINDEF’s Defence Expertise Group, which designs and develops technological capabilities for the SAF, will play a key function on this effort.
“Transferring ahead, we may also be increasing our partnerships with trade, to speed up progress in creating AI applied sciences and functions. I encourage the varied communities to come back collectively to co-ideate and co-innovate. I want to invite all researchers and practitioners with daring concepts to achieve out to us. Collectively, let’s collaborate, innovate and construct an thrilling and higher future, he added.”
Professor Mohan Kankanhalli, Deputy Government Chairman for AISG, stated that the present improvement efforts and outcomes might be translated into higher and safer applied sciences for on a regular basis dwelling.
“The issues which can be being addressed have dual-use functions and might be leveraged by each trade and defence ecosystems to profit Singapore. We’re excited to work with our valued collaborators, FSTD and DSO, as they’ve deep experience in each grand challenges to advise, information and empower our analysis groups to develop novel AI options that may successfully deal with necessary issues confronted by Singapore and the world. We imagine that there’s potential for these applied sciences to be scaled globally sooner or later”.
Sturdy AI and AI for Supplies Discovery Grand Challenges
AISG’s Grand Problem programme takes an outcome-driven strategy to unravel necessary issues confronted by Singapore and the world by means of multidisciplinary collaboration and translation of AI-based options from analysis to deployment.
Researchers from native Institutes of Greater Studying (IHLs) and Analysis Institutes (RIs), together with their collaborators from the general public and/or personal sectors, have been invited to suggest ground-breaking analysis concepts that undertake AI methods to handle each Grand Challenges’ statements.
Every problem assertion was collectively outlined by AISG, DSO and FSTD.
Sturdy AI Grand Problem – “How can we design sturdy CV programs for AVs that may get better a minimum of 80% of their unique accuracy after bodily testing-time adversarial assaults?”
Pc imaginative and prescient (CV) programs primarily based on AI/ML are susceptible to adversarial assaults which might compromise their security and safety, thus limiting their efficiency and adoption in vital environments. Within the area of autonomous automobiles (AV), these might end in delays, accidents and doubtlessly, even demise.
This Grand Problem goals to encourage the event of revolutionary approaches that enhance the robustness of those CV programs, to speed up the adoption of CV programs in vital environments.
AI for Supplies Discovery Grand Problem – “How can AI speed up inverse materials design to find superior supplies which can be 50% lighter whereas retaining and/or enhancing their useful properties?”
Technological breakthroughs and the event of next-generation programs equivalent to light-weight transportation automobiles and energy-efficient buildings are sometimes underpinned by the invention and improvement of recent and superior supplies. Nevertheless, this discovery course of might be prohibitively lengthy and costly. As well as, standard strategies might be restricted by a scientist’s private expertise, data and even biases, leading to solely incremental discoveries.
This Grand Problem goals to advance the state-of-the-art functions of AI in supplies discovery, to develop options that may assist scientists speed up materials discovery course of and overcome the constraints of standard strategies. This could in flip allow them to maintain tempo with the growing demand for brand new and superior supplies for next-gen programs.
AISG acquired 16 (7 Sturdy AI; 9 AI for Supplies Discovery) high quality analysis proposals from multidisciplinary groups in response to each grant calls. The groups comprise lead Principal Investigators (PIs) and Co-PIs from the Nationwide College of Singapore (NUS), Nanyang Technological College (NTU), Company for Science, Expertise and Analysis (A*STAR), Singapore Administration College (SMU), Singapore College of Expertise and Design (SUTD) and Singapore Institute of Expertise (SIT). In addition they contain trade companions and/or collaborators from abroad universities.
An Analysis Committee comprising members from numerous stakeholders within the AI ecosystem chosen 5 initiatives to be awarded Stage 1 funding:
Sturdy AI Grand Problem
- Holistic Transferring Goal Defence for Autonomous Driving Notion (by Affiliate Professor Rui Tan, NTU)
– This undertaking will develop a Transferring Goal Defence (MTD) strategy that creates and deploys AI mechanisms which can be numerous and quickly altering to restrict the publicity of vulnerabilities and alternatives for assaults, enhance the complexity and price for attackers. - Growth of Steady Sturdy and Safe Clever Methods for Autonomous Automobiles (by Professor Shuzhi Sam Ge, NUS)
– This undertaking will harness the collective strengths of machine studying, steady management and generative AI to develop a strong and clever system that may regulate its neural community’s behaviour and traits in real- time to counter adversarial assaults in advanced environments. - In direction of Constructing Unified Autonomous Automobile Scene Illustration for Bodily AV Adversarial Assaults and Visible Robustness Enhancement (by Professor Yang Liu, NTU)
– This undertaking will develop novel multi-view and multi-modal defences primarily based on a unified framework that includes a novel AV scene illustration that can be used to generate real looking scenes with numerous circumstances and assault vectors.
The initiatives are anticipated to ship suites of multi-modal assault and defence methods that may be deployed in the actual world to extend the robustness of AI fashions utilized in multi-sensor CV programs.
AI for Supplies Discovery Grand Problem
- MATAI: An AI-Powered Generalist Materials Discovery Platform (by Professor Bo An, NTU)
– This undertaking will develop an AI-powered generalist platform with three key parts: a holistic materials database, a foundational materials property predictor and a generalist multi-property materials designer in a big search house. - Design Past What You Know: Materials Knowledgeable Differential Generative AI (MIDGAI) for Gentle-Weight Excessive-Entropy Alloys and Multi-functional Composites (by Professor Ivor Tsang, A*STAR)
– This undertaking will develop a generative AI framework for supplies design that mixes with discriminative fashions to reinforce robustness and overcome knowledge shortage. By incorporating area data, the AI mannequin will allow a extra controllable, directed, and environment friendly exploration of fabric design areas by means of superior optimisation algorithms.
The initiatives are anticipated to ship AI options that can be utilized by AI and supplies researchers to (i) uncover novel supplies primarily based on efficiency wants, (ii) effectively discover past the boundaries of standard strategies, (iii) exploit new and present scientific data and (iv) scale back useful resource and time value of fabric discovery cycles.
In Stage 1, the groups can be awarded as much as S$4M every to work on their undertaking for the subsequent three years. Funding help for Stage 2 can be decided as a part of the ultimate analysis close to the tip of Stage 1. Chosen groups might be funded as much as S$5M over 2 years the place the groups could have the chance to work with businesses to adapt and scale up the developed AI methods for dual-use functions.
Hashtag: #AISingapore
The issuer is solely accountable for the content material of this announcement.
About AI Singapore
AI Singapore (AISG) is a nationwide programme launched by the Nationwide Analysis Basis (NRF) to catalyse, synergise and increase Singapore’s synthetic intelligence (AI) capabilities to energy our future, digital economic system.
AISG may also convey collectively all Singapore-based analysis establishments and the colourful ecosystem of AI start-ups and corporations creating AI merchandise, to carry out use-inspired analysis, develop the data, create the instruments, and develop the expertise to energy Singapore’s AI efforts.
AISG is pushed by a government-wide partnership comprising NRF, Good Nation and Digital Authorities Workplace (SNDGO), Infocomm Media Growth Authority (IMDA), Financial Growth Board (EDB), Enterprise Singapore (ESG), amongst others.
[ad_2]
Source link
