MDDI 演讲稿 · 2026-04-06
张仁宝部长在AUMOVIO-NTU企业实验室启动仪式上的致辞
张仁宝部长在AUMOVIO-NTU企业实验室启动仪式上的致辞
要点
- • 新加坡推出覆盖先进制造、互联互通、医疗保健与金融服务四大领域的国家AI使命,总理黄循财亲自担任新成立的国家AI理事会主席。
- • AUMOVIO-NTU企业实验室自2019年成立以来,已发表131篇论文、提交104项专利申请,并获得10项专利授权。
- • 实验室开发的AI Pathfinder工具可自动化测试汽车中央仪表盘界面,单次版本发布最多节省1200人时,目前正在一家区域整车厂试点推广。
- • 实验室的AI模糊测试工具经区域整车厂试验证明,能更快检测软件缺陷,并发现了传统测试完全遗漏的关键安全漏洞。
- • 实验室与AWS签署支持函以获取算力资源,并与Origgin签署谅解备忘录,通过初创企业、衍生公司及海外网络推动技术商业化。
- • 实验室计划吸纳逾140名研究人员、工程师与学生,研究方向涵盖AI、网络安全、新型材料与车联网通信。
完整译文(中文)
MDDI 英文原文译文 · 翻译日期: 2026-06-21
林金勇教授,南洋理工大学新加坡(NTU Singapore)副校长(产业)
卢建发先生,AUMOVIO Singapore 总裁兼首席执行官
各位嘉宾
下午好,感谢各位的邀请。
在今年的财政预算案中,黄总理谈到了人工智能,以及超越个别试点项目的必要性。他表示,我们必须在国家层面进行统筹部署,快速推进、扩大规模。正因如此,他亲自主持新成立的国家人工智能理事会,并在四大关键领域启动国家人工智能使命:先进制造、互联互通、医疗卫生和金融服务。
"超越试点"究竟意味着什么?这意味着设定真正难以实现的目标,迫使我们从根本上重新思考工作方式。这并不是说小问题不值得解决——它们当然值得——但我们不能满足于对某项技术的浅层应用,而应认真思考如何实现变革性的运用。
比如将从设计到原型的时间缩短一半,或在总装线上实现近乎零缺陷率。这些都是宏大而富有雄心的目标,值得我们认真对待,绝非靠渐进式改进或局部自动化就能达到。整个工作流程或许需要彻底重新构想。
我们仍在研究各个领域的具体目标形态,但它们有一个共同特点:都是"延伸目标"。这一点至关重要,因为它迫使我们提出不同的问题,引入不同的专业知识,并探索新的工作方式。
这正引出这个企业实验室的重要意义。自2019年成立以来,团队已发表131篇论文,提交104项专利申请,并获得10项专利授权。这些数字本身令人印象深刻,也反映出这是一个积极进取、力求突破的机构。
在汽车领域,最难攻克的问题往往出乎意料。测试工作繁琐、不那么光鲜,却是可靠性得以建立或崩溃的关键环节。作为乘客,你我都知道可靠性至关重要。这个实验室正是在这类问题上下功夫,涵盖现代车辆所需的全方位能力。
此外,停留在纸面上的延伸目标走不了多远。但它能将合适的人聚集到同一张桌子前。
在此次合作之前,AUMOVIO——当时还叫 Continental——正面临一个熟悉的挑战。为复杂汽车系统开发可靠软件耗时费力,且随着产品日趋精密,难度不断攀升。他们清楚人工智能可以提供帮助。
然而,要开发出能在真实汽车条件下稳定运行的解决方案,需要他们自身无法在短期内积累的研究深度。NTU 恰好具备这方面的专业能力。
另一方面,将研究成果转化为能承受真实产业压力的解决方案,需要大学所不具备的实际工业部署经验。但 AUMOVIO 拥有这方面的积累。
NTU 与 AUMOVIO 作为合作伙伴相互补充、相得益彰。除了我此前列举的成果,这一合作还产生了哪些成效?
实验室开发了 AI Pathfinder,可自动化测试车辆中央仪表板界面。在单次仪表板版本发布中,最多节省了1,200个工时,目前正在与一家区域汽车制造商开展试点。
NTU Edwin Teo 教授团队与 AUMOVIO 合作,开发了新型3D打印材料,可模拟车辆部件中的触感和振动反馈,为汽车行业长期面临的难题提供了一套紧凑的解决方案。
实验室在无线通信领域的研究也为国际标准的制定作出了贡献。换言之,源自新加坡的创新正在影响全球车辆之间的通信方式。
或许最具意义的,是这些成果的实现方式。在大多数合作中,业务部门往往在最后才介入,评估研究人员的产出。而在这里,AUMOVIO 的业务部门从一开始就直接参与塑造研究议程。未来负责部署解决方案的人,在解决方案设计阶段便已深度参与。这种更紧密的循环带来了更好的成果;当然,从大学角度来看,这也有助于避免"死亡之谷"——即研究发现和成果最终未能转化为产业所需的可用产品和服务。
我们期待在国家人工智能使命以及"Champions of AI"计划中看到更多此类合作。
在结束发言之前,让我谈谈为何这类合作的意义超越了这个特定的企业实验室本身。
当各组织真正全力攻克现实难题时,往往会发生几件事。
首先,他们向其他人展示了何为可能。许多企业在转型方面裹足不前,并非因为缺乏兴趣,而是不确定能否成功。实验室的 AI fuzzing 工作就是一个很好的例子。通过与区域内一家汽车制造商开展试验,AI fuzzing 工具检测漏洞和安全缺陷的速度远超以往,还发现了传统测试完全遗漏的关键漏洞。当其他企业看到这样的成果,他们对可能性的认知便会随之改变。
其次,此类合作能产生真实的应用案例,吸引更多合作伙伴加入。我很高兴今天见证两项签署——与 AWS 签署的支持意向书,以及与 Origgin 签署的谅解备忘录(MOU)。
AWS 的支持为实验室提供了计算资源,将推动人工智能研究持续前行。
Origgin 将与实验室合作,通过初创企业、分拆公司及海外网络连接,将技术推向市场。
实验室还计划与本地初创企业合作,例如 Squareroot8 Technologies,其量子安全研究与实验室自身的网络安全研究方向高度契合。
为推进各项雄心勃勃的目标,实验室将汇聚逾140名研究人员、工程师和学生,涵盖人工智能、网络安全、新型材料及车载通信等领域。
我衷心祝贺团队全体成员迎来今天的发布,也祝愿大家在未来的岁月中取得更大的成功。非常感谢各位的邀请。
英文原文
MDDI 官网原始记录 · 抓取日期: 2026-06-21
Prof Lam Khin Yong, Vice President (Industry), NTU Singapore
Mr Lo Kien Foh, President and CEO, AUMOVIO Singapore
Distinguished guests
Good afternoon and thank you for inviting me.
At Budget this year, PM Wong spoke about artificial intelligence and the need to go beyond individual pilots. He said we must, instead, organise at a national level and move with speed and scale. That is why he is personally chairing the new National AI Council, and why we are launching the national AI Missions in four key sectors: advanced manufacturing, connectivity, healthcare and financial services.
What does “going beyond pilots” actually mean? Well, it means setting goals that are genuinely hard to achieve, that compel us to fundamentally rethink how the work is done. It doesn’t mean that small problems are not worth solving – they are, but we cannot be content with shallow applications of a technology but should seriously think about transformative uses.
Like halving the time from design to prototype or achieving near-zero defect rates on a final assembly line. These are big ambitious goals that are worthy of our attention, they are not goals that anyone can reach through incremental improvements or a bit more automation here and there. Entire work processes may need to be re-imagined.
We are still working through what these goals look like in each sector. But they share one characteristic: they are stretch goals. That matters, because they force us to ask different questions, bring in different expertise, and find new ways of working.
This brings me to why this corporate lab matters. Since it was established in 2019, the team has produced 131 publications, filed 104 patent applications, and secured 10 patent awards. These numbers are impressive on their own, and also reflect an active organisation that is keen to make progress.
In the automotive sector, the hardest problems are not always where you expect them. Testing, for instance, is painstaking and unglamorous, but this is where reliability is built or broken. You and I know that as passengers, reliability matters. This lab works on problems like that, across the full range of what a modern vehicle needs to be.
Also, a stretch goal on paper does not go very far. However, it can bring the right people to the table.
Before this collaboration, AUMOVIO – then known as Continental – was grappling with a familiar challenge. Building reliable software for complex automotive systems is slow, expensive, and gets harder as products grow more sophisticated. They knew that AI could help.
However, developing something that works reliably under real automotive conditions required research depth they could not themselves build overnight. NTU had exactly that expertise.
On the other hand, translating research into solutions that work under real industry pressures requires real-world industrial deployment experience that a university does not have. But AUMOVIO does.
NTU and AUMOVIO complement each other well as partners. What has this collaboration produced, besides the achievements that I outlined earlier?
The lab built AI Pathfinder, which automates the testing of vehicle central dashboard interfaces. Up to 1,200 man-hours were saved in a single dashboard release and it is now being piloted with a regional automaker.
Professor Edwin Teo’s team at NTU worked with AUMOVIO to develop new 3-D printed materials that replicate touch sensation and vibration feedback in vehicle components, a compact solution to a problem the automotive industry has long struggled with.
And the lab’s wireless communication research has contributed to the international standard. In other words, innovation from Singapore is shaping how vehicles around the world talk to each other.
What is perhaps most significant is how these outcomes were achieved. In most collaborations, business units are brought in at the end to evaluate what researchers produced. Here, AUMOVIO’s business units directly shape the research agenda from the start. The people who will deploy these solutions are involved as these solutions are being designed. The tighter loop produced better results, and of course, from the university’s perspective, this also helps to avoid the valley of death where research findings and discoveries do not eventually make their way into usable products and services demanded by the industry.
We hope to see more of such collaborations in our National AI Missions and in our Champions of AI programme.
So, let me close with why these collaborations matter beyond this particular corporate lab.
When organisations push genuinely hard at real problems, a few things tend to happen.
First, they show others what is possible. Many companies are holding back on transformation, not because they are uninterested, but because they are not sure it can be done. The lab's AI fuzzing work is a good example. Through a trial with an automaker in the region, the AI fuzzing tool detected bugs and vulnerabilities much faster than before and uncovered critical vulnerabilities that traditional testing had missed entirely. When other companies see results like that, it changes what they think is possible.
Second, such collaborations generate real use cases that bring partners along. I am pleased to witness two signings today - a Letter of Support with AWS, and an MOU with Origgin.
The support of AWS gives the lab access to computing resources that will keep AI research moving.
Origgin will work with the lab to bring technologies to market through startups, spin-offs, and connections to overseas networks.
The lab is also looking to work with local startups like Squareroot8 Technologies, whose quantum security work aligns with its own cybersecurity research.
To pursue its ambitions, the lab will engage over 140 researchers, engineers and students across AI, cybersecurity, novel materials, and vehicular communications.
I congratulate everyone in the team on today’s launch, and wish you more success in the years ahead. Thank you very much for inviting me.