Audio and machine learning: Gaël Richard’s award-winning project

/0 Comments/in , /by

Gaël Richard, a researcher in Information Processing at Télécom Paris, has been awarded an Advanced Grant from the European Research Council (ERC) for his project entitled HI-Audio. This initiative aims to develop hybrid approaches that combine signal processing with deep machine learning for the purpose of understanding and analyzing sound.

“Artificial intelligence now relies heavily on deep neural networks, which have a major shortcoming: they require very large databases for learning,” says Gaël Richard, a researcher in Information Processing at Télécom Paris. He believes that “using signal models, or physical sound propagation models, in a deep learning algorithm would reduce the amount of data needed for learning while still allowing for the high controllability of the algorithm.” Gaël Richard plans to pursue this breakthrough via his HI-Audio* project, which won an ERC Advanced Grant on April 26, 2022

For example, the integration of physical sound propagation models can improve the characterization and configuration of the types of sound analyzed and help to develop an automatic sound recognition system. “The applications for the methods developed in this project focus on the analysis of music signals and the recognition of sound scenes, which is the identification of the recording’s sound environment (outside, inside, airport) and all the sound sources present,” Gaël Richard explains.

Industrial applications

Learning sound scenes could help autonomous cars identify their surroundings. The algorithm would be able to identify the surrounding sounds using microphones. The vehicle would be able to recognize the sound of a siren and its variations in sound intensity. Autonomous cars would then be able to change lanes to let an ambulance or fire engine pass, without having to “see” it in the detection cameras. The processes developed in the HI-Audio project could be applied to many other areas. The algorithms could be used in predictive maintenance to control the quality of parts in a production line. A car part, such as a bumper, is typically controlled based on the sound resonance generated when a non-destructive impact is applied.

The other key applications for the HI-Audio project are in the field of AI for music, particularly to assist musical creation by developing new interpretable methods for sound synthesis and transformation.

Machine learning and music

“One of the goals of this project is to build a database of music recordings from a wide variety of styles and different cultures,” Gaël Richard explains. “This database, which will be automatically annotated (with precise semantic information), will expand the research to include less studied or less distributed music, especially from audio streaming platforms,” he says. One of the challenges of this project is that of developing algorithms capable of recognizing the words and phrases spoken by the performers, retranscribing the music regardless of its recording location, and contributing new musical transformation capabilities (style transfer, rhythmic transformation, word changes).

“One important aspect of the project will also be the separation of sound sources,” Gaël Richard says. In an audio file, the separation of sources, which in the case of music are each linked to a different instrument, is generally achieved via filtering or “masking”. The idea is to hide all other sources until only the target source remains. One less common approach is to isolate the instrument via sound synthesis. This involves analyzing the music to characterize the sound source to be extracted in order to reproduce it. For Gaël Richard, “the advantage is that, in principle, artifacts from other sources are entirely absent. In addition, the synthesized source can be controlled by a few interpretable parameters, such as the fundamental frequency, which is directly related to the sound’s perceived pitch,” he says. “This type of approach opens up tremendous opportunities for sound manipulation and transformation, with real potential for developing new tools to assist music creation,” says Gaël Richard.

*HI-Audio will start on October 1st, 2022 and will be funded by the ERC Advanced Grant for five years for a total amount of €2.48 million.

Rémy Fauvel

Tatouage des données de santé, health data

Encrypting and watermarking health data to protect it

/0 Comments/in , , /by

As medicine and genetics make increasing use of data science and AI, the question of how to protect this sensitive information is becoming increasingly important to all those involved in health. A team from the LaTIM laboratory is working on these issues, with solutions such as encryption and watermarking. It has just been accredited by Inserm.

The original version of this article has been published on the website of IMT Atlantique

Securing medical data

Securing medical data, preventing it from being misused for commercial or malicious purposes, from being distorted or even destroyed has become a major challenge for both health players and public authorities. This is particularly relevant at a time when progress in medicine (and genetics) is increasingly based on the use of huge quantities of data, particularly with the rise of artificial intelligence. Several recent incidents (cyber-attacks, data leaks, etc.) have highlighted the urgent need to act against this type of risk. The issue also concerns each and every one of us: no one wants their medical information to be accessible to everyone.

“Health data, which is particularly sensitive, can be sold at a higher price than bank data,” points out Gouenou Coatrieux, a teacher-researcher at LaTIM (the Medical Information Processing Laboratory, shared by IMT Atlantique, the University of Western Brittany (UBO) and Inserm), who is working on this subject in conjunction with Brest University Hospital. To enable this data to be shared while also limiting the risks, LaTIM are usnig two techniques: secure computing and watermarking.

Secure computing, which combines a set of cryptographic techniques for distributed computing along with other approaches, ensures confidentiality: the externalized data is coded in such a way that it is possible to continue to perform calculations on it. The research organisation that receives the data – be it a public laboratory or private company – can study it, but doesn’t have access to its initial version, which it cannot reconstruct. They therefore remain protected.

a

Gouenou Coatrieux, teacher-researcher at LaTIM
(Laboratoire de traitement de l’information médicale, common to IMT Atlantique, Université de Bretagne occidentale (UBO) and Inserm

Discreet but effective tattooing

Tattooing involves introducing a minor and imperceptible modification into medical images or data entrusted to a third party. “We simply modify a few pixels on an image, for example to change the colour a little, a subtle change that makes it possible to code a message,” explains Gouenou Coatrieux. We can thus tattoo the identifier of the last person to access the data. This method does not prevent the file from being used, but if a problem occurs, it makes it very easy to identify the person who leaked it. The tattoo thus guarantees traceability. It also creates a form of dissuasion, because users are informed of this device. This technique has long been used to combat digital video piracy. Encryption and tattooing can also be combined: this is called crypto-tattooing.

Initially, LaTIM team was interested in the protection of medical images. A joint laboratory was thus created with Medecom, a Breton company specialising in this field, which produces software dedicated to radiology.

Multiple fields of application

Subsequently, LaTIM extended its field of research to the entire field of cyber-health. This work has led to the filing of several patents. A former doctoral student and engineer from the school has also founded a company, WaToo, specialising in data tagging. A Cyber Health team at LaTIM, the first in this field, has just been accredited by Inserm. This multidisciplinary team includes researchers, research engineers, doctoral students and post-docs, and includes several fields of application: protection of medical images and genetic data, and ‘big data’ in health. In particular, it works on the databases used for AI and deep learning, and on the security of treatments that use AI. “For all these subjects, we need to be in constant contact with health and genetics specialists,” stresses Gouenou Coatrieux, head of the new entity. We also take into account standards in the field such as DICOM, the international standard for medical imaging, and legal issues such as those relating to privacy rights with the application of European RGPD regulations.

The Cyber Health team recently contributed to a project called PrivGen, selected by the Labex (laboratory of excellence) CominLabs. The ongoing work which started with PrivGen aims to identify markers of certain diseases in a secure manner, by comparing the genomes of patients with those of healthy people, and to analyse some of the patients’ genomes. But the volumes of data and the computing power required to analyse them are so large that they have to be shared and taken out of their original information systems and sent to supercomputers. “This data sharing creates an additional risk of leakage or disclosure,” warns the researcher. “PrivGen’s partners are currently working to find a technical solution to secure the treatments, in particular to prevent patient identification”.

Towards the launch of a chaire (French research consortium)

An industrial chaire called Cybaile, dedicated to cybersecurity for trusted artificial intelligence in health, will also be launched next fall. LaTIM will partner with three other organizations: Thales group, Sophia Genetics and the start-up Aiintense, a specialist in neuroscience data. With the support of Inserm, and with the backing of the Regional Council of Brittany, it will focus in particular on securing the learning of AI models in health, in order to help with decision-making – screening, diagnoses, and treatment advice. “If we have a large amount of data, and therefore representations of the disease, we can use AI to detect signs of anomalies and set up decision support systems,” says Gouenou Coatrieux. “In ophthalmology, for example, we rely on a large quantity of images of the back of the eye to identify or detect pathologies and treat them better.“

Nuclear energy: outsourcing issues

/0 Comments/in , /by

Since the end of the 20th century, the practice of outsourcing has increased in France. This phenomenon has included strategic sectors, such as the nuclear power industry. Stéphanie Tillement, a researcher in Sociology at IMT Atlantique, has worked on the relationship between safety and subcontracting in the nuclear industry.

“Since the 1970s, we have witnessed an increase in outsourcing in many industrial sectors, particularly for maintenance activities,” says Stéphanie Tillement, a researcher in Sociology at IMT Atlantique. In the book Contracting and Safety, she and other French and international researchers offer a balanced, open-minded analysis of the practice of subcontracting. The book first addresses the nuclear power context. “We wanted to show the diversity of the relationships that exist between nuclear power operators and subcontractors, and in the links between subcontracting and safety,” says Stéphanie Tillement.

“Contrary to popular belief, the term “subcontracting” does not refer to a uniform reality: subcontracting situations vary in terms of the size of the provider company and the duration of the service provider’s presence on-site, for example,” she says. In addition to cases of “nuclear nomads,” who are often associated with subcontracting in the nuclear industry, some subcontracted staff have been working for years, even decades, at the same site, for the same contracting party. While the so-called nuclear nomads perform ad hoc interventions, which cause some to denounce forms of job insecurity, this is not the case for all external providers.  The working conditions and social interactions between the contracting authority and service provider therefore vary significantly depending on the type of subcontracting.

High-risk occupations

Outsourcing in the nuclear industry and its effects on the safety and security of the facilities and workers has received increased attention in both the political sphere (with the “Pompili” parliamentary committee in 2018) and academia. Annie Thébaud-Mony, the honorary research director of the Inserm Scientific Interest Group on Occupational Cancers demonstrated that “at French nuclear sites, employees of subcontracting companies were exposed to 80% of the collective ionizing radiation dose during maintenance activities,” Stéphanie Tillement says. In other words, subcontracted employees are more exposed to ionizing radiation than others. 

This is linked more to the nature of the outsourced tasks, which are often dangerous because they require intervention in high-risk areas, than it is to the type of protection used or follow-up with subcontracted employees.  In addition, the operators of typical nuclear facilities —such as nuclear reactors or radioactive waste treatment plants— are legally responsible for the safety of their facilities under the terms of the law of June 13, 2006 on transparency and safety in the nuclear sector. This also applies to outsourced activities. In the event of an incident or accident, the operator still remains responsible.

One of the major questions posed by the use of subcontracting is that of the monitoring of activities performed by external service providers. In order to ensure that the tasks are carried out in accordance with safety requirements, the operator is required to monitor subcontracted staff. True supervision by contracting authorities implies that they have maintained their industrial technical mastery of the outsourced activities and have allocated the necessary resources (time, human resources) to this supervision. “A major issue for monitoring pertains to the skills of the person performing the monitoring: if they do not master the technique, there is a risk that the monitoring will be reduced to formal checks without take into account the reality of the activity,” the sociologist explains. In the case of specialty subcontracting, this issue is all the more important since operators hire subcontracted staff who have specific skills which they do not have in-house. 

Complex relationships

In the nuclear sector, one example of specific skills that are both scarce and highly sought after are those of welders, whose role is fundamental in maintaining the safety of the equipment. Their work requires a high level of expertise. In the case of specialty subcontracting, the balance of power can therefore be in favor of the service providers, since the operator is dependent on them. They can therefore negotiate more favorable contracts (with less pressure on costs and deadlines, for example).

“Outsourcing poses a more general problem related to the fragmentation of work and organizations, which is more complex due to the multiple interfaces and interdependencies to be managed,” said Stéphanie Tillage. “We often see that companies that choose to outsource part of their activities are primarily concerned with short-term gain,” she explains. “In doing so, they omit an entire series of hidden long-term costs, including the need for the contracting authority to restructure the internal organization in order to ensure the long-term coordination and monitoring of the activities,” the scientist explains. This restructuring can be costly and require significant training in order to ensure the safety and security of workers in the long-term.

Rémy Fauvel

Stronger 3D prints

/0 Comments/in , /by

3D printing is a manufacturing process used for both consumer and industrial applications in the aeronautics, automotive, rail and medical industries. The Shoryuken project being developed at IMT Nord Europe aims to improve the mechanical performance of the objects printed using plastic and composite materials. To accomplish this, it combines 3D printing with laser welding technology.

In the industrial world, certain parts of cars, trains, airplanes, prostheses and orthoses are now manufactured using 3D printing. This manufacturing method enables the small-scale production of customized, geometrically complex parts using 3D digital models, without requiring expensive, specifically designed molds. This procedure saves time and materials when producing prototypes and products to be marketed. However, 3D printing has its limits, especially in relation to structural composite materials, which are plastic materials reinforced with fibers with a high level of resistance and rigidity.

3D printing processes that use composites with yarn containing cut reinforcing fibers generally produce materials with relatively weak mechanical properties. In order to improve the mechanical performance of the printed parts, manufacturing processes using yarns reinforced with continuous fibers are now in high demand in the industry. These yarns are made of thermoplastics, heat-sensitive materials, and continuous carbon or glass yarns. During the printing process, the yarns are melted in order to add the materials they contain layer by layer. The carbon fibers contained in the thermoplastic yarns do not melt and provide the object with solidity and resistance.  

However, the required level of resistance and rigidity can only be obtained in the direction of the fibers, since they are all positioned on a single printing plane. “The current composite 3D printing technology does not allow for the production of parts containing continuous fibrous reinforcements oriented in all the directions desired in space. This is a disadvantage when there are mechanical constraints in three dimensions,” says André Chateau Akue Asseko, researcher in Materials Science at IMT Nord Europe and winner of the Young Researchers call for projects by the French National Research Agency (ANR).

Hybridization of innovative technologies

This is precisely the technological barrier that the new Shoryuken* project seeks to overcome. To accomplish this, the initiative is studying the pairing of 3D printing with laser welding. This combination makes it possible to print two or more components for the same composite part in different printing directions and then use laser welding to assemble them.

“The difficulty stems from the presence of fibers or porosity, which disrupt the laser beam path due to the heterogeneity, which introduces thermal and optical diffusion phenomena,” the scientist explains. This assembly process therefore requires that the small areas filled with thermoplastics be treated differently during 3D printing. The laser radiation melts the thermoplastic polymer in a targeted manner with the composite material surrounding it. Once they are welded together, the two components become inseparable. This makes it possible to produce objects containing reinforcing fibers positioned in ways that allow them to resist mechanical loads in different directions.

Virtual engineering to optimize production

Modeling and simulation tools integrating multiphysics coupling are being developed to optimize these innovative design and production processes. These tools therefore contain information on the interaction between the laser and materials and their thermal and mechanical behavior. “Virtual engineering makes it possible to define the optimal assembly conditions that will ensure the quality of the welding interface,” says André Chateau Akue Asseko. The software, populated with information on the materials of interest, such as melting points, is used to simulate the behavior of two materials that are welded together in order to prevent spending too much time and materials on 3D printing tests.

The user can therefore adjust the laser parameters in order to conduct an optimal weld right away. “These simulations allows us to identify the optimal temperature and speed ranges for welding,” the researcher explains. The development of this type of tool would allow companies to reduce their development and industrialization costs before production by avoiding potential assembly problems. This would ensure the mechanical performance of the manufactured goods.

Read more on I’MTech: 3D printing, a revolution for the construction industry?

Multisectoral applications

 “For this project, we chose to focus on the health sector by producing a prosthetic arm as a demonstrator,” says the scientist, who is currently in contact with companies specialized in prosthesis design. André Chateau Akue Asseko explains that he initially chose to prioritize this sector for pragmatic reasons. “There is strong demand in this field for customized items, adapted to the users’ morphology. The parts are reasonably sized and compatible with the capabilities of our experimental equipment,” the researcher says.

The Shoryuken project will end in 2026. By that time, the future process and digital tool could convince other industries, such as the rail and automotive sectors, of the benefits of customizing parts and tailoring their functionalization for small and medium-scale production runs. For transportation companies, the significantly lighter weights of the parts designed and produced help to cut down on fuel consumption and thereby reduce carbon emissions, which are a key concern in the current global environmental context.

Rémy Fauvel

* The ANR JCJC SHORYUKEN project on the “Assembly of Hybrid Thermoplastic and Thermosetting Carbon Composite: Customization of Complex Structures” is funded by the French National Research Agency (ANR) as part of the 2021 Generic Call for Proposals (AAPG 2021 – CE10) on “Industry and Factories of the Future: People, Organization, Technology.”