chip – Progress in Research

Photonic chips for low-power neural networks

Posted on 15/05/202306/08/2024 by Erik Franco

A study by the Politecnico di Milano and Stanford University, published in the journal Science, shows that it is possible to create extremely efficient neural networks using photonic chips.

Neural networks are distributed computing structures inspired by the structure of a biological brain and aim to achieve cognitive performance comparable to that of humans. They are used in many areas, such as speech and image recognition and synthesis, autonomous driving and augmented reality systems, bioinformatics, genetic and molecular sequencing, and high-performance computing technologies.

Neural networks are trained with a large amount of known information, on the basis of which they become able to adapt their behaviour, working autonomously. However, their training is an extremely energy-intensive process.

Researchers from the Politecnico’s Photonic Devices Lab and Polifab, the university’s micro- and nano-technology centre, in collaboration with researchers from Stanford University, have sought a solution and developed a silicon microchip just a few square millimetres in size with an integrated photonic accelerator that allows calculations to be performed very quickly – in less than a billionth of a second – and efficiently. Thanks to this photonic chip, neural network operations take place with considerable energy savings.

In addition to neural networks, it will be possible to use this device as a computing unit for multiple applications where high computational efficiency is required, e.g., for graphics accelerators, mathematical coprocessors, data mining, cryptography and quantum computers.

The study online

Optical wireless: the new frontier for communication

Posted on 08/07/202206/08/2024 by Erik Franco

In the field of cable transmission, the advent of optical fibres represented an epochal technological leap, allowing light to be used to transfer enormous amounts of data, and they now form the basic infrastructure of the Internet and global telecommunications systems.

For wireless communications too, it is expected that optical connections will soon represent the new frontier. Similarly to what happens in optical fibres, even in free space, light can travel in the form of beams having different shapes, called “modes”, and each of these modes can carry a flow of information. Generating, manipulating and receiving more modes therefore means transmitting more information. The problem is that free space is a much more hostile, variable and unpredictable environment for light than an optical fibre. Obstacles, atmospheric agents or more simply the wind encountered along the way, can alter the shape of the light beams, mix them and make them at first sight unrecognisable and unusable.

A study by the Politecnico di Milano, conducted together with Stanford University, the Scuola Superiore Sant’Anna in Pisa and the University of Glasgow and published in the prestigious journal Light: Science & Applications, has found a way to separate and distinguish optical beams even if they are superimposed and the form in which they arrive at their destination is drastically changed and unknown.

This operation is made possible by a programmable photonic processor built on a silicon chip of just 5 mm². The processor created is able to receive all the optical beams through a multitude of microscopic optical antennas integrated on the chip, to manipulate them through a network of integrated interferometers and to separate them on distinct optical fibres, eliminating mutual interference. This device allows information quantities of over 5,000 Ghz to be managed, at least 100 times greater than current high-capacity wireless systems.

The activity is funded by the European Horizon 2020 Superpixels project, which aims to create next-generation sensor and imaging systems by exploiting the on-chip manipulation of light signals.

The studio is authored among the others by Francesco Morichetti, head of the Photonic Devices Lab and Andrea Melloni, director of Polifab, the Politecnico di Milano centre for micro and nanotechnologies.

Read the study online

DIANA: studying drugs for brain by miniaturized platform

Posted on 22/06/202206/08/2024 by Erik Franco

The Politecnico di Milano has developed an innovative technological device, for industrial use, aimed at the study of new drugs for the treatment of brain disorders such as Alzheimer’s or Parkinson’s disease. This is the main result of the European project DIANA (Organ-on-a-chip Drug screenIng device to tArget braiN diseAse), funded in 2019 by the Proof-of Concept call of the ERC (European Research Council).

DIANA brought together universities and companies in a consortium between the Politecnico di Milano and the innovative SME Neuro-Zone srl, specialized in discovery activities to support the development of drugs in the field of neurological and neurodegenerative diseases. The project was enhanced by the involvement of Diego Albani, researcher in neuroscience at the Istituto di Ricerche Farmacologiche Mario Negri IRCCS in Milan, an expert in innovative pharmacological approaches for neurodegenerative diseases.

The Chip4DBrain platform developed by DIANA is based on a state-of-the-art technology known as” organ-on-a-chip “that allows you to reproduce complex organ functions on systems the size of a microscope slide.
Carmen Giordano, Professor of Bioengineering at the Politecnico di Milano.

This is a further step towards the development of evolved in vitro models, which can reproduce some of the key characteristics of biological systems, such as the three-dimensionality or the simultaneous presence of different types of cells, just as it is in our brain, to evaluate the potential of a new drug to cross the blood brain barrier and effectively target the brain.

This innovative miniaturized platform is able to integrate in a single in vitro system the blood-brain barrier, which protects our brain from the aggression of molecules and external agents, and a model of brain tissue.

Chip4D Brain has also allowed the implementation of cellular models of the blood brain and brain barrier, already in use at Neuro-Zone, making them closer to the biological profile of a patient thanks to the use of commercial human stem cells.

In an international scenario where the restrictions or ethical assessments towards the use of animal models also in the field of neuroscience are very complex, predictive and advanced in vitro models are increasingly urgent.

The mission that DIANA has faced is highly topical: in the coming decades, brain diseases such as Alzheimer’s disease or Parkinson’s disease will have a significant increase, but unfortunately the development of effective drugs requires a ten-year process, investments of billions of euros to facing a failure rate, which for Alzheimer’s disease alone is close to 95%.

DIANA project factsheet

The project DIANA has received funding from the European Research Council (ERC) under the EU’s Horizon 2020 research and innovation programme, under Grant Agreement No. 899431.