NTT open for ’embryonic research’ of photonics software development

Good software repeats. We know that almost every long-running software platform or tool and every operating system or application goes through a series of logical incremental progressions and upgrades, generally referred to as an iterative process or just ‘iteration’. Even when there are completely new hardware and software development methodologies, programming languages, and automation, there is still an organic evolutionary process involved as software engineers sink their teeth into emerging technologies while exploring new business (or life) deployment scenarios.

NTT took this fundamental truth to heart and explained how its Innovative Optical Wireless Network (IOWN) technology platform is now in a position where the company hopes software engineers, system architects and other technologists will embark on the embryonic research process with it.

Why is this process such a research mission? First of all, because NTT’s IOWN concept is a new type of communication infrastructure that can provide high-speed (low latency) broadband communication and massive computing resources using low-power technologies, including optical technologies. The focus of these optical technologies is the use of photonics. With the long-standing truth of Moore’s Law governing the development of silicon microprocessors now becoming memory, photonics-based processors promise to reimagine the way we create chip ‘wiring’ with light-based technology that can perform high-speed arithmetic calculations.

Why making embryonic prototypes is good

In his role as President and CEO, NTT Research, Inc., Kazuhiro ‘Kazu’ Gomi spoke about the need to address challenges such as unsustainable energy consumption. He has also been vocal about the need to combat data and personal privacy issues and address the rising costs of medical care in an aging world population.

“Amidst so many challenges, invention is necessary – and non-linear thinking is the heart of our enterprise,” said Kazu Gomi, who calls for embryonic experimentation at the heart of what the company says are ‘sophisticated research modalities’ today. Focused on what the company calls ‘basic research’, i.e. technological innovations that will have a profound and long-term impact on the way we live our lives, the call to software developers from NTT is as broad as it is open.

“Building with photonics is an embryonic process, and we want to encourage software engineers at all levels to touch these platforms,” ​​enthused Kazu Gomi. “We have to do things that are a bit disruptive and move away from the notion of linear growth or linear development in the normal sense. Building quantum neural network technology and harnessing the power of photonics is a huge exploration of resources, techniques and approaches – and it’s definitely an industry-wide effort.”

When we talk about ‘basic’ research and development within NTT Research, we are focusing on what we might also call pure (as opposed to applied) research.

“To further clarify, in applied research, an engineering team may typically work to deliver a defined technology product for a Q3 release (for example) with relatively defined functionality and deliverables that match what might be an already existing plan,” Kazu Gomi explained. , in a conversation with journalists and analysts this April 2024.

Since most of the work done at NTT Research’s Silicon Valley facility is of a pure/basic embryonic type, the engineers mesh well with the teams in Tokyo, where there are different pools of resources to tap into. So the question is, how do software developers and other technology engineers feel about working on embryonic-level products, some of which may never make it to market?

Finding the detection factor

“For individuals, it’s all about the ‘finding’ factor and diving into the process of discovering how things work. These are people who thrive on discovery and know that they are part of a team unit that can drive innovation and start producing it at different stages of the evolution of any technology,” said Chris Shaw, director of marketing at NTT Research. “In reality, some people are naturally great at being scientific researchers on esoteric emerging projects – and likewise, some people are inherently suited to work on (in the absence of examples) supply chain manufacturing applications for ERP systems. We know that the team philosophy at NTT Research is to put the right people in the right roles and let them do what they do best.”

Shaw says this level of engineering professional isn’t as concerned about failure. As he noted, the fulfillment and motivation factor for them is all about finding the basic fundamental way things (anything) work. They know that if they find it, someone else (on the production strategy team or elsewhere) will see that spark and maybe turn it into something.

Being a technology that is essentially still in its infancy – and with this discussion focusing on the work of IT engineers in a pure research space – are we at the point where we start talking about app developers, web developers, cloud developers, mobile developers and now to photonics developers?

“That reality is coming,” assured Kazu Gomi of NTT Research. “In terms of validation for these technologies there are many examples of work, in terms of codification (in areas such as certification) we’re not there yet, but that’s okay, this is a journey. The systems emerging from IOWN and the use of photonics will influence our technology use cases at all levels over the next few decades, but at this stage it is important to remember that we are still building the hardware for many of these innovations, i.e. the software services tend to come after the fact , obviously. Photonics today still has an impact on efficiencies that we can apply now, i.e. there is so much communication between microprocessors and memory cells, so light-based engineering is already changing the parameters here, but much more will come.”

Eye AI, NTT’s tsuzumi LLM

As an example to illustrate what NTT says is some of the most progressive work going on at its research facilities in Silicon Valley and Tokyo, the company’s current work with the ‘tsuzumi’ Large Language Model (LLM) is rather esoteric. Now integrating visual reading technology to work like the human eye, NTT’s tsuzumi LLM can now more comprehensively understand and process graphic elements within documents, such as images, graphs, diagrams or icons.

Owned by NTT and light in terms of software code footprint and component ‘parameters’ (factors that govern how an AI model is able to interact with its data and make predictions from it), the company says this visual reading technology supports existing ‘reading ‘ possibilities of tsuzumi in a cheap and energy-efficient way. Currently undergoing enterprise trials to explore prototype use cases, the technology was developed in collaboration with Professor Jun Suzuki of Tohoku University’s Center for Data-Driven Science and Artificial Intelligence.

“LLMs have become capable of handling high-level natural language processing tasks with high accuracy, and multimodal models are beginning to emerge, including those that integrate vision and language,” said Kyosuke Nishida, senior distinguished researcher at NTT. “However, there are still significant challenges in understanding documents or computer screens that contain both textual and visual information, such as graphs and tables. By integrating our visual reading technology with tsuzumi, NTT aims to give ‘eyes’ to AI-powered tools, unlocking new applications and functionalities,” said Nishida.

Light LLM low power

While traditional LLMs require large amounts of energy for training, the smaller size of tsuzumi (the ultra-light version has 600 million parameters and the lightweight version has 7 billion parameters) reduces the energy consumption and costs associated with training, inference and tuning, making it a more viable and cost-effective option for companies. In addition, tsuzumi supports twenty human languages, including English and Japanese, and enables inference on a single GPU or CPU. It is compatible with both visual and audio modalities and can be tailored specifically for specific industries or an organization’s business cases.

This year, we know that NTT researchers envision four initial primary use cases for corporate deployment of tsuzumi and visual reading technology. According to NTT, “This includes customer experience solutions including call center automation; employee experience solutions for tasks involving manual search and reporting, including electronic medical record management in the healthcare industry; value chain transformation for industries including life sciences and manufacturing; and software engineering for systems and IT departments, including development and coding assistance and automation.”

NTT is currently conducting commercial trials of tsuzumi and has consulted with more than 500 companies around the world about potentially introducing the technology into their systems.

Our ephemeral embryonic IT future

Of all the emerging work carried out here, it is not unusual to talk about experimental prototyping when it comes to the development and implementation of artificial intelligence – especially in the generative space of artificial intelligence. Business technology platform companies used to be afraid to say they were in the ‘early experimental phase’ with any technology – perhaps refreshingly, that’s no longer the case.

Because so many technologies at the cutting edge are so embryonic, still in their infancy, and often somewhat transient in changing cloud networks, we need to encourage and embrace experimentation—especially of course in new and radically different fields like photonics. There is light at the end of this tunnel, in every possible sense of the word.