독서토론모임 시드니시나브로
5 in 5
5 innovations that will change our lives within 5 years – IBM Research
1. AI-powered robot microscopes will help clean up the worlds water supplies – 인공지능 구동 로봇 현미경
물 부족(Water shortage) 현상은 심각한 문제이며 향후 2025년까지 전 세계 인구 1/4에게 영향을 미칠 수 있음.
플랑크톤(plankton)은 물의 변화나 화학물질 오염 등 모든 경우에 대한 핵심적인 단서를 제공할 수 있는 생물이다. 따라서 프랑크톤에 대한 연구를 통해 바다, 호수, 강 등의 상태에 대한 실시간 정보를 얻을 수 있게 된다.
하지만 이 생물에 대해 현재까지 알려진 것이 별로 없으며 샘플을 연구실로 보내야만 연구를 할 수 있다는 한계도 있다.
IBM은 이 연구 분야를 발전시키기 위해 자연 서식지에서 플랑크톤을 관찰할 수 있는 소형 AI구동 로봇 현미경을 개발하고 있다. AI기술을 통해 로봇이 수집한 정보를 실시간으로 연구자에게 보내는 것은 물론 연구자의 비정상적인 행동에 경고를 보낼 수도 있다.
2. Crypto-Anchors and blockchains to fight Counterfeiter – 암호 앵커
현재 금융 분야에서 활용되고 있는 블록체인 기술을 물류에 적용하는 것도 각광받을 것으로 예상된다.
공급망은 여러 나라에 걸쳐 있고 여러 당사자들이 상품을 중심으로 복잡하게 연계돼 있다. 블록체인 기술을 사용할 경우 공급망 내 디지털 거래의 안전성과 신뢰성이 매우 높아질 것으로 기대된다.
이에 따라 IBM의 연구자들은 디지털 지문 형태의 암호 앵커(crypto-anchors)라는 이름의 좁쌀보다 작은 디지털 지문에 대해 연구하고 있다. 암호 앵커는 제품에 내장될 수 있고 제품의 신뢰성을 증명하기 위해 블록체인과 연결될 수 있다.
3. Lattice cryptography will discourage even quantum-powered hackers – 격자 암호화
양자 컴퓨터의 시대가 오면 현재의 암호화 프로토콜(규칙)은 제 역할을 하지 못할 가능성이 높다. 양자 컴퓨터는 수백만가지 경우의 수를 이용해 암호 시스템을 무력화할 수 있기 때문이다.
IBM은 ‘격자(lattice)’암호화 기술을 제안한다. 이 기술은 격자로 알려진 복잡한 수학 문제 내에 암호를 숨겨 양자 컴퓨터도 해킹이 어렵게 한다.
또 현재의 기술상으로는 사용자가 암호를 입력하고 데이터에 접근할 때 잠재적인 해커에게 노출될 수 있다. 새로운 기술을 활용하면 사용자가 해커에게 노출될 위험을 없애면서 보안 데이터에 접근할 수 있다.
Lattice란 – Lattice는 “격자”란 용어로 번역 및 기술되고 있음. 기하학적으로 기술하면, n차원 공간 R에서 점(point)들이 규칙적 주기적 반복적으로 그리드(격자무늬)배열로 배치되어 있는 것을 의미. 이때 점을 Lattice Point(격자점)라고 함.
즉 Lattice는 “Lattice Point”들의 집찹이라고 할 수 있음.
Lattice Point들은 특정한 패턴(간격 및 각도)으로 무한히 반복된는데, 특정한 패턴은 Basis Vector에 의해서 결정되어 짐. 수학에서는 이것을 아래와 같은 수식으로 기술함.
“n차원 공간 R에 속하는 basis vector 값들(b1, b2, … bn)과 모든 정수들의 선형(linear) 조합(combination) rkatemf의 집합”
즉, 모든 Lattice Point는 Basis Vector 값들의 선형 결합으로 표시될 수 있음. 여기서 basis가 되는 vector값은 다양하게 구성될 수 있으므로, 다양한 형태의 Lattice가 존재함.
2차원 공간에서 그림으로 나타내며 아래와 같음. b1과 b2는 Lattice의 basis(기반)를 구성하는 벡타(Vector)값.
4. AI bias will explode, but only unbiased AI will survive – 선입관 없는 인공지능
AI시스템은 자신이 훈련을 받는데 활용된 데이터에 크게 의존한다. 데이터에 편차가 있는 경우 AI시스템과 실제 세계와의 관련성이 낮아지게 된다.
IBM은 AI시스템이 독립적으로 공정함을 판정할 수 있는 시스템을 만들어 이같은 문제를 해결하려고 한다. 이 기술을 활용하면 인간이 더 나은 의사결정을 할 수 있도록 하는 신뢰성 있는 안내자가 될 수 있다.
5. Quantum computing will move from the research labs into the real world – 양자 컴퓨터
양자 컴퓨터는 양자여각을 이용해 자료를 처리하는 컴퓨터다. 전통적인 컴퓨터가 0 또는 1로 구성된 이분법으로 연산한다면 양자 컴퓨터는 두 가지가 중첩될 수 있는 ‘큐비트’(Qubit)라는 양자적 상태의 조합으로 연산을 처리한다.
이 때문에 양자 컴퓨터는 대규모 계산과 특정 문제 해결 능력에 있어 한계를 보이는 오늘날의 수퍼 컴퓨터를 압도할 것으로 예상된다.
해리엇 그린 IBM 아시아태평양 지역 최고경영자(CEO)는 CNBC와의 인터뷰에서 “IBM은 지난 5년 동안 새로운 기술에 380억 달러를 투입했다”며 “이런 투자는 매우 현실적이며 기업을 변화시키는데 도움이 될 것”이라고 말함.
현재 산업계는 양자 컴퓨터를 어떻게 사용할 수 있을지 발견해 가는 단계에 있다. IBM의 연구부서는 앞으로 5년 내에 업계가 양자 컴퓨터에 적용할 수 있는 응용 프로그램을 개발할 수 있을 것으로 보고 있음.
References
1. Bernard Marr, Forbes “5 Big Technology Innovations of 2018: IBM Reveals Amazing Developments that will Impact All of Us”
2. 중앙일보 “IBM이 뽑은 미래 이끌 5대 혁신기술”
3. IT World “낟알 크기 컴퓨터로 모조품 잡는다. 초소형 블록체인 컴퓨터 ‘크립토앵커’”
4. AEP 코리아 “Lattice Cryptography”
김광덕 교수(빅토리아대학, 시드니시나브로 회원)
…………………………………
참고 자료
5 Big Technology Innovations Of 2018: IBM Reveals Amazing Developments That Will Impact All Of Us
Today IBM is announcing its ‘Five-for-Five’ report, highlighting five amazing technological advances that will have a real impact on lives in the next half decade.
From AI-powered micro-cameras monitoring the world’s oceans to the graduation of quantum computers from the lab into the real world, here’s a summary of these game-changing breakthroughs.
1. AI-powered robot microscopes will help clean up the world’s water supplies
Water shortage is a problem that could affect up to a quarter of the world’s population by 2025. The behavior of microscopic plankton can give vital clues on everything from chemical pollution levels to temperature change.
Autonomous, robotic cameras developed by IBM and powered by AI have the potential to monitor this behavior in more detail than has been possible before. Data from the cameras can be analyzed to give real-time insights into factors affecting water quality and life in our lakes and oceans.
Ahead of today’s release, Jeff Welser, vice president and lab director at IBM Research told me “So with internet of things (IoT) we talk about putting sensors everywhere – and this is an example of just how far we can take this, when we combine it with AI.
“We know people are going to have all kinds of problems with clean water in the future, and we know there are micro-organisms in water, that if we can get them to tell us what’s happening that would be a really great way to understand any potential problems.”
Making the devices as low-powered as possible is essential, in order to be able to deploy them at scale. To this end, they don’t contain lenses or focus mechanisms or other complicated mechanical parts, but simply track shadows and movements through light sensors.
“We can get a lot of information from that”, Welser says. “Are the microscopic organisms moving around as they should be? There’s a lot of interesting science on what that behavior means.”
2. Crypto-Anchors and blockchains to fight Counterfeiter
Nobody likes knockoffs – blockchain and crypto-anchors will help to crack down on counterfeiting as well as ensure security in the food supply chain
With $600 billion a year lost to the global economy through fraud and counterfeiting, blockchain offers the potential to ensure the provenance of everything from food to diamonds and life-saving medicines.
In a global economy, goods pass through many different sets of hands between their point of production and the end consumer. This leaves them open to tampering and theft problems which blockchain technology could help to eliminate.
In order to work, however, there needs to be a tamper-proof link between the physical products and the digital records on the blockchain. This is where crypto-anchors come in – microscopic codes or identifiers which can serve as “digital fingerprints” to ensure security at every stage of the journey.
“The challenge here is that the blockchain can record all the transactions but somewhere you’ve got to link the transactions to the actual physical object itself – so that you know the banana that got scanned is the actual banana that got to you,” Welser tells me.
“What crypto anchors do is they basically embed tiny codes, like microscopic QR codes, in a way that makes it so that if you tried to replace it with a similar one, you could tell it had been tampered with. When you put those codes onto the blockchain, the supply chain is then protected.”
3. Lattice cryptography will discourage even quantum-powered hackers
Complex algebraic structures called lattices will become a valuable tool in the age of quantum computers. With more and more sensitive data being collected and stored online, security measures will need to keep pace with the growing capability of hackers, as virtually unlimited amounts of computing power become cheaper and more available.
Until now ever-more complex cryptography – from 64-bit encryption to 128 bit and 256 bit – has been the standard response to the increasing amount of CPU power available to hackers. As quantum computing becomes mainstream, this will no longer be enough.
“The reality is there’s constantly a battle on with cybersecurity, we need to make sure we continue to have cryptography and encryption that can keep the bad guys out, and all of that relies on the fact that the maths is so hard to do that trying to solve it with a computer takes an unreasonable amount of time,” Welser says.
“We have to make sure that as computers get faster, we can continue to keep ahead of them. In particular, this is a concern with the quantum computers that are coming up.”
Lattice cryptography involves encoding data within high-dimensional algebraic structures which even theoretical million-qubit quantum computers will find tough to crack. It also opens up the possibility of Fully Homomorphic Encryption (FHE), which will enable computers to operate on data while it is still in an encrypted state – eliminating the security flaw inherent in existing systems whereby data has to be decrypted (and thus made vulnerable to hackers) in order to be processed. This could, for example, mean credit reference systems which can make credit scoring decisions without personal data ever being exposed.
4. AI bias will explode, but only unbiased AI will survive
The most sophisticated AI systems are only as good as the data they are trained on, and if that data has been collected in a biased or compromised way, then results are unlikely to fit with the real world which we are attempting to model.
Devising new ways to monitor for bias, and eliminate it at the source, are keys to creating AI software which accurately reflects reality, rather than the biased human view of reality that AI promises to help us transcend.
Welser tells me “One of the hopes of AI is that it will help us make decisions in less biased ways, because AI won’t have human biases. So if you’re making decisions on mortgages, or who should get bail, or who you should recruit, all of those things have biases built into them.
“AI systems would hopefully be able to make those decisions with less bias, but the challenge is that the AI gets trained on data, and if that data has a bias then your AI will be biased.
“We spend a lot of time right now working on how the systems we’re training aren’t inadvertently learning bias, and also protecting them from players who might be trying to teach them bias that we don’t want them to have.”
AI systems trained in this way to provide an unbiased, objective model of the world are likely to be the most successful. This will help us to tackle moral and ethical problems which will be encountered by any industries or fields of research attempting to use AI to tackle social issues or make decisions that will affect human lives.
5. Quantum computing will move from the research labs into the real world
Half a decade from now, quantum computing will be an essential element of any computer engineering degree, IBM researchers are today predicting. Rather than a technology shrouded in mystery, it will be fundamentally understood and a practical tool in use solving problems in many disciplines and industries.
“We’re doing a lot with quantum computers,” Welser says. “We have a 15 qubit system on the cloud which anyone can go and use and we’re seeing a lot of interesting things.
There’s over 100,000 hits on it now which are people going and writing programs on it.
“But it’s still a toy at the moment – a researcher’s playground. I think that in the next five years we will have systems that are both large enough and have low enough error rates that we will see some really interesting things that have real value.
“The most likely area is things like quantum chemistry – simulations of things like molecules or chemical bonds.
“Right now we use very large high-performance computing systems but even with those, once we go beyond simulating a few molecules or atoms it becomes very difficult because there are too many variables. And of course, as we are working at the sub-atomic level those are quantum variables. They can be simulated very directly with a quantum computer.”
An understanding of quantum computing will be essential for those looking for careers in any scientific field, and students will leave university with hands-on experience of running practical experiments on quantum powered machines. And just as most engineers or scientists could today outline what is meant by the computing term “bit”, in five years’ time, the term “qubit” will be widely understood. _ Bernard Marr (Contributor)