Fondecyt 2021: 28 projects sponsored by FCFM were selected by the fund

Fondecyt 2021: 28 projects sponsored by FCFM were selected by the fund

Those selected include 4 projects presented by Center for Mathematical Modeling (CMM) researchers Aris Daniilidis, Helena Montenegro, Michal Kowalczyk and Felipe Tobar.

The Fondecyt 2021 competition of the Chilean National Agency for Research and Development (ANID) granted funding for 28 projects of FCFM researchers, which will be developed over the next 3 to 4 years. The projects are from the departments of Electrical Engineering, Mechanical Engineering, Astronomy, Geology, Physics, Mining, Computer Science and Geophysics, as well as the Center for Mathematical Modeling (CMM) and the Advanced Mining Technology Center (AMTC).

“Energy harvesting from active matter”, is the name of the proposal led by María Luisa Cordero, academic of the Department of Physics (DFI) and researcher of the Active Matter Physics Millennium Nucleus. The project was born precisely in the Nucleus, where they have already studied bacteria capable of swimming, which have been encapsulated inside small droplets (tenths or hundredths of microns in diameter), managing to observe how these, when swimming, agitate the fluid inside the droplet and make it move. “We have an experiment in which the bacteria make the droplet travel in any direction, randomly, so the goal is to rectify the movement of the droplet and get it to move in one direction. Basically, what we are doing is making an engine that is made of these bacteria,” he says.

Another of their experiments uses magnetotactic bacteria (to which movement can be directed with a magnet) in which they demonstrated that the droplet rotates, but what they want is to use that rotation to make the droplet roll, another type of bacterial motor. “The goal is to generate a controllable motor, what we have now is a random motor and we want to direct the movement of the droplet,” he says. Achieving this could allow it to be applied at the biochemical or biomedical level. “Bacteria can often be used to generate proteins, we could put them in the droplet and then take it where it can be used,” he adds.

In the Department of Industrial Engineering (DII), Denis Sauré will develop the project “Exploration vs Exploitation in Bilevel Optimization”, which seeks to extend a classical methodological framework for sequential decision making under parametric uncertainty (multi-armed bandits) to scenarios where the rewards obtained by implementing a decision come from the rational response of an opponent. “An example is attack-defense models, which are used to make infrastructure defense decisions before potential terrorist attacks, or to decide how to break into drug trafficking networks,” he explains.

“The main challenge in this scenario is that the decision maker has no control over the information he receives back when implementing a decision, and rather it goes through the rationale of the opponent. In this project we incorporate the stochastic response of the opponent; we seek to establish fundamental performance boundaries, along with optimal policies,” he says.

Nancy Hitschfeld, from the Department of Computer Science (DCC), will work on the design and implementation of new algorithms to generate spatial discretizations (meshes) that can be used to model and solve problems in science and engineering. “Some applications where there is a need for such meshes to model increasingly complex problems, minimizing the number of cells, are the simulation of fracture mechanics, the representation of topography for both surface and subsurface hydrological modeling and modeling of rock and porous media, among others,” she says. In order for the generated meshes to be useful, professors from the Mechanical Engineering and Civil Engineering departments collaborate in the project, as well as co-investigators from the computing area of the Austral and Federico Santa María universities.

Important research for the new Institute of Data and Artificial Intelligence

Six projects selected in this competition are related to Artificial Intelligence (AI), showing the critical mass that is already working in the area and that will be the basis of the next Institute of Data and Artificial Intelligence, which will be implemented this year in the faculty.

The proposal of Felipe Tobar, researcher at the Center for Mathematical Modeling (Advances on generative models for statistical machine learning: theory and practice) focuses on the development of models for machine learning from a statistical point of view. Machine learning is the way to build structures or models for AI. “It is important to develop models, in particular statistical models, for machine learning, because AI is increasingly taking a more preponderant role in our daily activity; it is in our phone, in our car, in the way credits are determined, etc., it is so present in our life that one has to pay special attention to the models we use in AI, which statistical point of view is able to represent uncertainty when one occupies an AI model. The representation of uncertainty, from my point of view, is crucial to apply different models in human activity”, the researcher indicates.

Among the objectives of the project is to obtain a unified vision of statistical methods for machine learning where different methods that already exist can be used to improve others, detecting synergies between them. But also to be able to apply them to three areas relevant to society: health, social good and climate.

“In an era where in the country we are apparently super interested in developing AI methods, it is super positive that a proposal that is at the core of the conceptual development of AI is awarded. AI is not only about buying software, more powerful computers or hiring the services of consulting firms that are going to make an AI to improve our production, it is about developing tools that are in the depths of AI theory, and the fact that a project of this type has been awarded is a good sign for the development of this area in the country”, he says.

The Department of Electrical Engineering (DIE) researchers Jorge Silva (Topics on information theory applied to machine learning and decision: theory, methods, and applications), Néstor Becerra (Robust speech processing for human-robot interaction) and Marcos Orchard (A hidden-Markov-model-based failure prognostic framework for real-time prognostic decision making), in addition to the postdoctoral Fondecyt of Kerlyns Martínez (Stochastic models of mean-field interacting system in mathematical physics and machine learning: theoretical problems and applications) led by Joaquín Fontbona (DIM), and Antonio Miranda’s (“Understanding the spatial configuration of forest fire resistant landscapes using artificial intelligence”) with Andrés Weintraub (DII), which are also related to data and AI.

More frontier science

Simón Casassus, an academic from the Department of Astronomy (DAS), will apply his original methodology to study the formation of planets around stars in formation. The observation of these planetary systems in formation will be carried out mainly with data from the ALMA radio observatory, trying to detect planets in birth immersed within the disk that forms the star (pre-stellar disk), which is then considered a protoplanetary disk. “It is very difficult to detect the presence of planets there and in this four-year project, I am proposing an alternative that uses algorithms that I am developing and seeks the indirect detection of protoplanets through the footprint they leave in the planetary disk in which they are immersed,” explains the astronomer.

The necessary data will be obtained as soon as the ALMA observatory, closed due to the pandemic, is reopened. “The observing projects have already been selected and there are two of these projects led by me and others by collaborators that are on the program and will be executed during 2021 if the observatory is reopened,” he assures.

Angelo Castruccio, academic of the Department of Geology, will investigate the role of volcanic building dimensions in arc volcanic activity patterns (“The role of volcanic building dimensions in arc volcanic activity patterns”), using the Parinacota volcanoes in northern Chile, and Lonquimay and Llaima in the south as the object of study. “The project tries to explain the relationship between the construction of a volcanic edifice (the mountain or elevation that constitutes the volcano) and its size, with the type of eruptions that are generated, in terms of erupted volume, explosiveness, frequency, etc.,” says the researcher. The Parinacota, Lonquimay and Llaima volcanoes were chosen because they have suffered events of destruction or collapse and rebuilding, “which allows us to study how the eruptive style changes as the volcano grows,” he says.

In the Department of Geophysics (DGF), Eduardo Contreras-Reyes will work on the analysis of the structure of the Carnegie (Ecuador), Malpelo (Panama) and Taltal (Chile) submarine ridges -Structure of the Carnegie, Malpelo and Taltal ridges along the oceanic Nazca and Cocos plates: implications for hotspot magmatism and lithospheric-. “These are quite large mountains, comparable to the Andes Mountains, which collide with the continent slowly, taking millions of years, and the subduction under the South American continent, in this case, causes the continent to erode, or can produce tremors due to friction. My project is to study the structure of these submarine ridges to better understand how the collision with the continent will be”, he explains.

Using geophysical techniques similar to tomography, a kind of magnetic resonance is performed on a seismological scale, to study the internal structure, in this way it is easier to carry out studies and interpretations of how the friction with the continent is. “There are several mountains that have been studied in the world, these data from Panama and Ecuador have been available for 20 years and in Taltal, a German ship will have access to these data this year. These mountain ranges are important from a scientific point of view, because those in Panama and Ecuador are wider than those in northern Chile, and by being able to study all three we can make this comparison to better understand the processes of volcanism,” he adds.

Meanwhile, the project of the director of the Department of Mechanical Engineering (DIMEC), Viviana Meruane (Optimal design of ultralight sandwich panels with cellular truss cores and large phononic band gaps), will focus on the optimal design of ultralight sandwich panels, which consist of two thin sheets adhered to a thick plate of lightweight material (core). “The concept of sandwich structure is a common principle in nature; for example, tree branches and animal bones are examples of sandwich materials with foam-like cores. Because of their high stiffness and strength with minimal weight, sandwich panels are attractive for use in applications where light weight is critical,” he explains.

This project proposes to develop a methodology to obtain sandwich panel core topologies that optimize strength-to-weight ratio in conjunction with vibration isolation. “This creates numerous opportunities to use such panels in applications such as satellites, spacecraft, aircraft, ships and automobiles,” says the researcher, who is also a member of the Metamaterials Millennium Nucleus.

Finally, the objective of the proposal (Advances in visible light communication systems applied to underground mining environments) by César Azurdia, an academic from the Department of Electrical Engineering (DIE), is to model, analyze and experimentally validate the communication channel in schemes based on visible light (VLC) in subway mining environments. “It is also expected to evaluate the impact of the communication channel in various ICT applications required in subway mining scenarios,” he explains. The methodology to be used to validate the hypotheses and achieve the proposed objectives is based on developing mathematical models, validation through computer simulations and experimental validations.

List of selected researchers:

1 Cesar Azurdia (DIE)
2 Nestor Becerra (DIE)
3 Roger Bustamante (DIMEC)
4 Simón Cassasus (DAS)
5 Angelo Castruccio (DGL)
6 Marcel Clerc (DFI)
7 Eduardo Contreras Reyes (DGF)
8 Maria Luisa Cordero (DFI)
9 Aris Daniilidis (CMM)
10 Daniel Díaz (DGF)
11 Xavier Emery (DIMIN)
12 Juan Fernando Escobar (DII)
13 Humberto Estay (AMTC)
14 Luis Foa (DFI)
15 René Garreaud (DGF)
16 Nancy Hitschfeld (DCC)
17 Christian Ihle (DIMIN)
18 Michal Kowalczyk (CMM)
19 Viviana Meruane (DIMEC)
20 Marcos Orchard (DIE)
21 Rodrigo Palma B (DIE)
22 Gonzalo Palma (DFI)
23 Rodrigo Pascual (DIMEC)
24 Sebastián Riquelme (DGF)
25 Denis Sauré (DII)
26 Jorge Silva (DIE)
27 Felipe Tobar (CMM)
28 Helena Montenegro (CMM)

Source: Subdirección de Comunicaciones FCFM – U. de Chile (in spanish)

Related news: Doce investigadores postdoctorales se adjudicaron Fondecyt 2021 (in spanish)

Posted on Jan 29, 2021 in News