AI, the Singularity, and Digital Nature--Yoichi Ochiai's Thoughts Today

Yoichi Ochiai (Adviser to the President of the University of Tsukuba) has become so popular that not a day passes when he is not seen in the media--including TV programs such as variety shows and documentaries, not just media related to technology. Mr. Ochiai discussed a variety of topics over the course of his lecture, ranging from AI, the singularity, education, social issues, and the significance of his research theme of "digital nature." This report highlighted eight topics and took a peek into the mind of this young researcher at the center of attention today.
[Fujitsu Insight 2017 Report on the Special Lecture "Utilization of AI and IoT"]

1. The Modern Age and the Present: Graduating from "Standardization and Unification"

In the modern age, individuals worked hard to "standardize" for the greater mass of people--in other words, to create a human-like society; that was the theme. As a result, we became "standardized." Television and mass media are good examples of how everyone came to share similar sensibilities and thought processes.

However, the present day is different. With the advent of the Internet and widespread use of smartphones, individuals can now do whatever they want. This change toward diversity is crucial. The time has come for us to move out of the phase in which we think within the frameworks of a unified society, and instead to consider updating ourselves to a diversified society.

2. AI: AI Is an "Organism"

I frequently use AI in my lab. When I come up with an idea, I think geometrically whether it can be realized by computer. If it seems possible, I put it into practice.

For example, consider the "bone-like structure" and "dolphin acoustic hologram" (explained later) that I am researching and developing. I optimize the functions that animals acquired over the long course of evolution by using differential geometry and neural networks to reproduce them as technology.

Some people hypothesize that as AI continues to evolve, it will trigger a singularity in which it surpasses humans and achieves a god-like existence. However, I support a different view.

This is because I do not look for AI to have human-like thoughts. On the contrary, I think AI is very animal-like and natural. If we provide it with a certain phenomenon, its reaction comes back like a physical phenomenon. Humans (brains) and today's AI are most likely two completely different things that are not analogous or comparable in the first place.

3. AI: Deep Learning Is "Interconnection"

There is a famous haiku by Basho Matsuo: "The old pond--a frog jumps in--the sound of water." When you hear this poem, what sort of images do you come up with? I think many of you will conjure an image like the following.

Taking it a step further, some of you may also imagine the splash that the frog makes when it jumps into the water--a sound that inexplicably echoes in our hearts. Or, you may feel the sense of spirituality that spreads in the silent air afterwards.

Why do we Japanese conjure this sort of image? Because we hold within us an ancient Buddhist concept, Jiji Muge, which refers to the interconnectedness of all things. This allows us to connect the dots of any given incident in order to arrive at a conclusion drawn from their mutual interactions.

On the other hand, Westerners do not have this concept and therefore cannot "output" it. This interconnectedness is precisely the end-to-end principle that we use for deep learning and reinforcement learning.

The countless neurons (nerve cells) in the brain connect and interact with one another to produce answers from input information. Neural networks of AI and deep learning imitate this function.

What I am getting at is, even without understanding the whole--such as the interconnected sensibilities of Asian people that allow them to imagine an entire scenery from just one word, or the fine workings of an AI's neural network--we can come up with an answer using the end-to-end principle.

And this is the important part: to be prominent, even in the fields of culture and art, like haiku. Training is necessary to illustrate depth from limited information, such as with ink paintings and poems that have verses comprised of a designated number of syllables. The mind must be trained to appreciate art in order to see and understand such subtle contexts. Art theories on the complexities of individuals and models created by artificial neural network training have some similarities.

I believe this kind of perspective will prove a critical Asian strength and topic in the progress of future AI R&D, something that is intuitive to us Japanese.

One more thing. If all phenomena in the world became a block chain and were connected by the end-to-end principle, they would stand up against companies that are fortified by platforms.

I often express this as "BoE" (block chain websing). Platform companies that do not have technology as their principal axis may be replaceable with a local token economy. If an AI were to be built in there using the end-to-end principle, I think the AI and IoT fields would be more than sufficient to win against overseas non-technology IT giants that possess enormous platforms.

In other words, if the model does not require an understanding of the inner workings and there is no need for new technological development, tokens and ICO can be used to create a benefit principle IT service.

4. The Singularity: Bitcoin's Symbolic Existence

Discussing the singularity by limiting it in scope to people and computers misses the mark. Among researchers, the singularity is often used to refer to a singular point. In the sense of the technological singularity, such a phenomenon has already occurred in a variety of fields. For example, the singularity that I define is the singular point of how people work and society's formation since the modern age.

Take, for example, "credit creation." Society's financial capital is increased by the social trust in financial institutions as well as the cycle of deposits and loans made by corporations and individuals. This has allowed capital to be separated from material value within society, multiplied, and handled. However, virtual currency has introduced changes with a speed and on a scale even greater than that. I believe Bitcoin and Ethereum are the singularity of credit creation.

Financial institutions are not the ones producing Bitcoin, which is established based on trust in beneficiaries and miners. Its exchange rate is determined by computer-performed hash processing over a network, and it has no centralized power to begin with. That said, I do not know about current miners.

The entities who acknowledge virtual currencies like Bitcoin as currencies of value are not nations, as used to be the case, but market users. Bitcoin has transcended the framework of nations, which is symbolic of the paradigm (conception) of the modern age, and it can be used anywhere on earth. Just as I mentioned earlier, this is the singularity of moving from a centralized, unified society to a diverse society. This is interesting, both in the sense of trust and that it is unaccompanied by material value.

Personally, I am all for this sort of movement and ecosystem structure. Additionally, if you look at Bitcoin not just as a simple venture target but from the perspective of the singularity, I believe you will begin to see things in a different light. In other words, what we have here is the singular point of a new financial system. Moreover, it is not governed by a centralized power, but rather by computer code. The ecosystem is maintained through the repetition of coding and decoding. Codes and calculations--in the truest sense, this is a market composed of "computational nature" like a physical quantity.

5. Education: "Nurturing Curiosity" Is the Ideal

A short while ago, I held a workshop with WIRED and Dentsu for children averaging 15 years old to teach them about hardware, software, machine learning, and UX (user experience). I was amazed by the learning abilities of today's children and how easily understandable the tutorials on the Internet were. Participants achieved results on a level equivalent to that of a master's degree thesis five years ago.

This was possible because they used sources on the Internet as well as digital fabrication using computers. What can we take away from these results? It is not the observation that "middle school students are smart these days." Rather, it is that by using a computer, we can easily find out how to do something. This made me think that going forward, we should shift the perspective of education from "teaching children about how to do things" to "finding ways to nurture curiosity."

How do we make concepts on a calculator and research results into reality? I believe children should gain as much experience as possible in the field, actually moving their hands and putting their brains to work.

One more thing: until now, education has been divided into fields such as the sciences, humanities, and arts. However, going forward, there will be demand for human resources who have mastered all fields of study regardless of their preferences.

For example, in my research, there is no point if you cannot do art, science, engineering, and design all at once. This is true not only for engineering fields but also for creating works of art.

What sort of complexities or authorship elements do you need for people to pay attention to you? What elements, from a design perspective, can be removed from a blueprint? Meanwhile, on the backend, technical knowledge is necessary to make products work. Moreover, you need management sensibility to monetize the work of art and get customers to buy it.

6. Social Issues: Solving the Aging Population Problem with Technology

The social issues we face today can be solved with technology because many of Japan's problems are rooted in an insufficient workforce, a lack of communication, and insufficient optimization.

The keys here are robots and communication devices. We can use robotics technology to do what humans cannot and solve the lack of manpower caused by the declining birthrate and aging population. Then, people can focus their work-related efforts on problems they can fix. We must skillfully integrate platforms that can lend a human hand with automation.

Care for the elderly, especially tasks that require strength such as transporting and loading wheelchairs, can be handled by stable robots. Making use of the time freed up by robot usage, nursing care businesses can provide added value in the form of services. Staff members should primarily provide care services, but right now, the lack of manpower to conduct operations is forcing companies to use much of their time to deliver and transport people.

If this is done, the result will be increased profits and more spare time due to the reduced workforce. This will enable employees to take side jobs or for employers to raise wages. In other words, it will resolve the stigma that nursing care jobs pay little.

Additionally, the eyeglass-shaped HMD (head-mounted display) that I am developing (more on this later) will allow elderly people to see clearly even if their sight deteriorates--without employing a focus adjustment function.

In another field, technological development is underway to allow muscles that could move but now do not to move again by providing electric stimuli to such muscles. Computers can optimize such calculations. In this sense, the number of healthy elderly people will increase, operating costs will decrease, and the issue of a declining working population will be resolved.

Source: Digital Nature Group

A good example is the "Yadori" communications robot that my lab is developing as part of joint research with the Fujitsu Social Science Lab (Fujitsu SSL).

I believe that in the future there will be great demand for fabrication and automation techniques related to telepresence robots. It may be possible for people at home to become distributed intelligence and perform office reception services and operations without being physically present.

Joint research with Fujitsu SSL

7. Research Themes: What I Am Creating

At the stage of conducting research in my university lab, I seldom think of how what I am creating may be used in the future or what role it will play in society. To be more precise, I try not to think about it.

This is because I believe that the conception of future inventions and technologies should not start from the objective. In other words, I do not limit the scope of my ideas.

On the other hand, when thinking about how to implement the knowledge I have acquired in the lab in actual society, I place a lot of focus on what to emphasize. For this reason, I run a venture business called Pixie Dust Technologies.

Therefore, I also put much effort into joint research projects with companies like Fujitsu. How do we turn the research-level principles of the computer society and digital nature into reality in the form of products for society? Instead of the industrial product approach that has been used up until now, I am thinking about using software to create diversified products.

This "software" is not simply a program that will operate inside a computer. I label it software, but it includes marketing, methods, software, and content. As such, in a sense, I am adding hardware and printed physical materials to what used to be run by software as computer graphics. This is the realization of a digital nature society, which is my research theme. Allow me to introduce a few examples of this.

Coded Skeleton

For example: I created software with a 3D printer which, if an object is designed to change forms using CG, a physical body can be printed in the same manner that it is animated on the computer.

With a spring-like skeletal structure, the direction of movement can be limited in order to make the object move as if it is composed of bones and muscles.

If this invention is utilized in robotics, it will become possible to create robots that can move more like living organisms. The important thing is to create the "software" to design such products.

Echolocation and Printing

Dolphins' eyes and ears have not developed specially--if anything, the truth is the opposite. Their functions are inferior even to those of humans and yet they can swim so effortlessly in the sea.

This is because they use an organ in the forehead called a melon to emit ultrasound waves, which enables them to perceive their surroundings in three dimensions. I am researching whether I can use a 3D printer to create a structure like this melon.

In other words, it is not about making echo phased-array holograms using electrical and electronic controls as it used to be, but rather creating printed holograms with physical substance.

Source: iStock/alexxx1981

Using Silkworms as a 3D Printer

This is another project undertaken by students at my university. When silkworms are raised under specific conditions, they spin and lay silk threads on a surface instead of making round cocoons. In Japan, this characteristic has been utilized in a technique known as "flat cocoon" since the Edo Period.

In the world, there are people who create large architectural structures, but creating human-sized objects is the hardest. This requires geometric calculations. The research my lab is conducting right now aims to get one silkworm to act as a 3D printer head. By gathering thousands of silkworms, we will be able to produce still larger digital fabrications quickly.

Lighting fixture digitally fabricated from silk threads spun by silkworms acting as printer heads (Source: Digital Nature Group)

Retinal Projection Head-mounted Display

I do research to further develop known inventions. Recently, the main development project I have been putting a lot of effort into is a retinal projection HMD (head-mounted display).

With low energy consumption, wide viewing angle retinal projection displays, I suspect we will no longer need portable displays such as smartphones and computers in the future. When contacting friends or doing work, we will no longer have to operate smartphones or computers. Instead, we will be able to simply put on this sort of eyewear device, which will bring up displays and images in the space in front of our eyes.

Additionally, this device's projection technology differs from prior ARHMD. When combined with a camera, it acts as a pair of digital eyeglasses for people with poor eyesight (e.g., old-sightedness, near-sightedness, or far-sightedness). In such cases, the camera installed in the glasses will recognize the space and images to project them directly onto the retina without putting the obtained images through a lens. Therefore, it ignores the focus adjustment function to enable images to appear clearly.

8. Vision: Why I Do Research

Some in society harbor the impression that if computers continue to evolve as they have, AI's capabilities will eventually overtake those of the human brain and we will be ruled by computers and robots. This sort of dystopian future is often seen in the movies, but it is not the future I envision.

If anything, I believe the opposite. The nature brought about by computers will spread, and by deconstructing the modern age paradigm, such as free will, a new form of co-existence will be born. As a result, I envision a future in which we can live lives that are more natural, stress-free, and reasonable than what we experience today.

The reason I do research every day is to make this sort of new Asian computerized nature a reality. All my research has a consistent theme: computerized nature and digital nature. I believe we are literally standing at the turning point of a paradigm shift.

In a digital nature world, the lines separating real things (materials and the environment), virtual things, people, and machines (computers) will become blurred. Robots will become extremely similar to humans, humans will become similar to robots, components will be like collectives, and collectives will be like components.
Phone calls to companies will likely be handled by AI computers that sound and act just like humans, while there may also be humans who speak robotically as if words are inputted into them. Most work performed by humans now will be done by AI, and humans may also supplement robots.

Customer service and autonomous driving too--AI will increasingly permeate our daily lives. In other words, optimization in the broadest sense will occur without being limited in scope by purpose.
When I speak of this, some people may become worried that AI will steal their jobs, but such concerns are unfounded because, as I mentioned earlier regarding care for the elderly, I believe that services and lifestyles can be made more fulfilling by using the time freed up by AI.

In other words, it will become normal for one person to have multiple jobs, and there will be fewer cases of labor time being taken up by low-paying work.

I doubt many people can let go of today's ubiquitous society now that the Internet and IoT have become reality. The same will be true for the future's computerized nature and digital nature.

I predict a sense of integration so great that no one will want to let go of it once they experience it for themselves once. This will be a comfortable future and society--one I look forward to. After all, virtual currencies are convenient and we cannot go back to the age before smartphones. Moreover, autonomous driving, retinal projection, and token economies will further accelerate our society and increase its fluidity.

Such a world is just around the corner. Within a few years, we will begin to see opportunities to make it a reality. The research I am doing is a prerequisite to realize this sort of society. What also must happen is accelerated integration of universities and academic-industrial collaborations in order to nurture human resources who understand and share thoughts like mine. That way, by combining education with practice, we can realize society's grand design through a technological vision.

(Interview and writing by Tadayoshi Sugiyama. Photos by Hirokazu Hasegawa. Editing by Tsuyoshi Kimura.)

Yoichi Ochiai
Advisor to the President of the University of Tsukuba, Assistant Professor of the Faculty of SLIS, Head of the Digital Nature Group, and media artist
Yoichi Ochiai was born in 1987. He completed his studies at the Graduate School of Interdisciplinary Information Studies at the University of Tokyo faster than others and received his PhD (in Interdisciplinary Information Studies). Advisor to the President of the University of Tsukuba, Visiting Professor of Osaka University of Art, and Visiting Professor of Digital Hollywood University. He is also CEO of Pixie Dust Technologies Inc. He has been recognized a number of times both domestically and internationally, including the World Technology Award 2015 from WTN (USA) in 2015, the Prix Ars Electronica from Ars Electronica in 2016, and the STARTS Prize from EU. His major publications include "The Century of Enchantment" and "Survival Strategy for the Super AI Era."