Information and Intelligence
Grand challenge
How can we efficiently transport the exponentially growing amount of information that connects our society over the internet, while also providing intelligent insights in real-time on massive data streams, like those that come from new microcomb-based instruments?
Background
Our society increasingly relies on internet connectivity, which drives up data rates in our networks every year. We need new technologies to keep this critical infrastructure running.
In parallel, we are constantly trying to better understand our changing world through sensing and measurement, however this generates massive amounts of raw data.
This presents new challenges in efficiently and rapidly processing this data to gain intelligent insights, that will enable us to get good use out of all this information that we generate.
What generates this data, and how do we use it
We live in a world saturated by information, but continually demand more. Gaining intelligent insights from this information has been a focus from the first internet search engines, and continues with new tools like ChatGPT (and other machine-learning – or “AI” – examples).
What current role does the science of light – called photonics – play?
- It serves as the backbone of our internet, enabling video calls to keep in touch with loved ones, flexible work arrangements, and generally supporting our connected society.
- It acts as the connective tissue within data centres, expanding the rate at which they process information
- It senses the world around us, including observing earthquakes and weather patterns, measuring traffic in our cities, assessing the health of civil infrastructure, and monitoring our borders and underwater environments.
The challenge
As the amount of data we send through our current fibre optic infrastructure doubles every three years, the strain on this “backbone of the internet” increases.
To gain intelligence from this information via tools like ChatGPT, the total rates of processing and shuffling around data inside data centres is orders of magnitude larger. To sustain this growth, our team believes that we need new approaches to better send and use information in optical systems.
The challenges with current approaches are:
- The size and power of fibre optic transmission systems to achieve the required data rates, through using multiple independent devices, is too high to expand much further
- Sensing systems are centralised, as the data processing devices are too energy hungry to work on site, leading to unnecessary data transmission and delays in gaining needed intelligence.
- Photonic and electronic technologies have complementary strengths, but are not effectively integrated to create solutions that leverage both the:
- Vast bandwidth and high precision of photonics
- Information processing flexibility of electronics
Our vision
To enhance both the capacity of internet’s bandwidth, and to reduce the amount of raw data travelling through our networks, microcombs may hold the answer.
Microcombs are tiny devices that live on a fingernail-sized chip, and create a rainbow of infrared light allowing data to be transmitted on many frequencies of light at the same time.
These have already proven to achieve record-breaking internet transmission and processing, however, have even more potential to provide:
- Record-breaking internet speeds on our existing infrastructure, while shrinking size and power consumption.
- More efficient photonic techniques to support digital electronics in analysing mammoth data sets, reducing the power needed to support emerging “AI” applications.
- Real-time processing of the massive bandwidth of signals coming from microcomb-enabled technologies, giving us insight into these measurements so that we can focus on the features we need to.
How does our team fit into the broader Centre?
Our team works closely with:
- Microcomb Science and Technology: The insights into fundamental physics of our microcomb platforms will help us to integrate our lab-bench components onto a chip the size of a fingernail.
- Spectroscopy and Microscopy: By adapting our expertise in analogue optical signal processing to take in image data, we will be able to find key features in spectroscopic imaging to help end-users gain insights fast.
- Sensing and Measurement: Sensing technologies using the parallelism of optical microcombs will generate vast bandwidths of data. Techniques in optical accelerators for machine learning will help us synthesise data from these sensors to provide intelligence in real time.
- Astrocombs: Microcombs in orbit provide a promising new platform to test our understanding of the Universe, but also provide a basis for expanding our internet into space – supporting exploration, and providing new ways to connect with people back on Earth.