My search
https://www.inria.fr/en
en2MDS
https://www.inria.fr/en/2mds
<p>Our aim is to develop a multi-scale modeling framework for channelopathies, a group of diseases caused by the dysfunction of membrane ion channels or their interacting proteins. These pathologies include Dravet syndrome (DS), a severe form of infantile epilepsy. We will study this class of pathologies at different scales, each with dedicated modeling and experimental tools, in close collaboration with a partner laboratory of&nbsp;the Basque Center for Applied Mathematics and with the Basque Center for Neuroscience.</p>https://www.inria.fr/en/2mdsALPHA-Design
https://www.inria.fr/en/alpha-design
<p>The design of new innovative products generally relies on numerical simulations to predict the physical behaviour of the system (mechanical, thermal, electromagnetic, etc.) and carry out its optimisation. These methods, which are well established, are however costly in terms of computing resources and require significant expertise, which hinders their dissemination in industry.</p><br>
<p>This project aims at exploring an alternative approach, based on neural network learning techniques. Contrary to most approaches in artificial intelligence, the aim is not to approach data but to learn the laws of physics (ordinary differential equations or partial differential equations). The challenge is then to build a neural model including, in a single training, the different physics involved, their couplings, as well as the conditions of optimality of the system.</p><br>
<p>Such an approach would be a powerful lever for democratising multidisciplinary optimal design and the use of digital twins.</p>https://www.inria.fr/en/alpha-designAPOLLON
https://www.inria.fr/en/apollon
<pre class="moz-quote-pre"><span style="font-family: Arial;">The study of ideas of historical texts require an expertise identifying the context, the precise<br /> meaning of the statements and to establish their genealogy. Statistical counts cannot take<br /> into account the nuances induced by the style, the presence or the absence of a word <br />than may change the meaning of a sentence, the context.<br /> <br />Our project aims at automatizing such a task by combining the reinforcement learning <br />technique with human expertise to create the first lexicon of ideas from the corpus of <br />Aristotle's Politics, then to apply such techniques to other corpuses and the to automatic <br />text generation.<br /> <br /> </span></pre>https://www.inria.fr/en/apollonAuDaCITI
https://www.inria.fr/en/audaciti
<p>Recent achievements of AI were led by large computational models and equally large datasets, often requiring intensive human labor to curate and annotate. In some cases however, the AI can interact with its environment and users to autonomously generate its own learning data. In this project, we study a specific setting of a quadruped robot equipped with a gripper arm, allowing the robot to navigate a room and manipulate its objects. Through the use of reinforcement learning, the robot learns to proactively generate a representative dataset of the different object categories in the room, which it is then able to retrieve when queried. The project will be in collaboration with Aalto University, Finland, that will bring its expertise in robotics and computer vision, while Inria Scool will bring its expertise in reinforcement learning. By solving a traditionally supervised learning task in full autonomy, we hope to get insights on how AIs of the future could proactively seek to improve their modeling of the world.</p>https://www.inria.fr/en/audacitiBackbone
https://www.inria.fr/en/backbone
<p>We seek to explicit, and then exploit, potential structural properties of algebraic systems in order to solve them efficiently.</p><br>
<p>Very often, large systems describing real problems are well structured in the sense that the impact of a slight change of one variable is relatively local and do not propagate through out the entire system.</p><br>
<p>We aim to capture automatically such locality and use it appropriately to decompose the original task into a sequence of smaller problems to solve.</p><br>
<p>The challenge is then to find a good sequence that is guaranteed to have a better complexity bound than a generic handling of the problem.</p>https://www.inria.fr/en/backboneBrainGPT
https://www.inria.fr/en/braingpt
<p>In the wake of the emergence of large-scale language models such as ChatGPT, the BrainGPT project is at the forefront of research in Artificial Intelligence and Computational Neuroscience. While these models are remarkably efficient, they do not reflect how our brain processes and learns language. BrainGPT takes up the challenge by focusing on the development of models more faithful to human cognitive functioning, inspired by data from brain activity during listening or reading. The ambition is to create more efficient models, less reliant on intensive computations and massive volumes of data. BrainGPT will open new perspectives on our understanding of language and cognition.</p>https://www.inria.fr/en/braingptDefine
https://www.inria.fr/en/define
<p>ABS develops novel methods to study protein structure and dynamics, using computational geometry/topology and machine learning. LCQB is a leading lab addressing core questions at the heart of modern biology, with a unique synergy between quantitative models and experiments. The goal of DEFINE is to provide a synergy between ABS and LCQB, with a<br /> focus on the prediction of protein functions, at the genome scale and for two specific applications (photosynthesis, DNA repair).</p>https://www.inria.fr/en/defineDEPARTURE
https://www.inria.fr/en/departure
<p>Quantum correlations are a cornerstone of quantum information theory, to understand its foundations (e.g., the 2022 Nobel price on the Bell theorem) as well as its concrete applications (for the &lsquo;device independent&rsquo; certification of quantum devices). And in the context of quantum networks? In collaboration with Jukka Suomela (Aalto University, Finlande), DEPARTURE looks into the limits of quantum distributed computing in quantum networks, in particular the tasks which can be solved more efficiently in the quantum internet.</p>https://www.inria.fr/en/departureDiscotik
https://www.inria.fr/en/discotik
<p><em>Computational morphomechanics</em> is the study of living tissue morphogenesis through the scope of physics-based computational modeling. It has become a forefront tool to study organogenesis, where mechanical stresses play a paramount regulating role. At macroscopic scale, smooth living tissues can be describe as Riemannian manifolds, subject to continuous mechanics. Concomitantly, at the cellular scale, they appear as networks of discrete effectors, where mechanics should be expressed in a combinatorial manner. Current state-of-the-art models, based on &ldquo;classic&rdquo; <em>Finte Element Methods</em>, struggle to efficiently integrate this cellular (discrete) / tissular (continuous) dichotomy.</p><br>
<p>The <strong>Discotik</strong> project aims to alleviate this difficulty through the use <em>Discrete Exterior Calculus</em> to express the laws of mechanics. While classic <em>FEM</em> rely solely on simplicial meshing of manifolds, &rdquo;<em>DEC</em>&rdquo; also exploits their dual structure, composed of cellular complexes. Strikingly, such cellular structures appear naturally in living tissues. We will assess this modeling approach on a specific, circumscribed problem: The morphomechanics of plant epithelia. We expect the &ldquo;<em>DEC</em>&rdquo; framework not only to enable faster computations but also to expose the deep connection between mechanical stress, tissue geometry and the corresponding cellular network topology.</p>https://www.inria.fr/en/discotikExODE
https://www.inria.fr/en/exode
<p>In biology, the vast majority of systems can be modeled as ordinary differential equations (ODEs). Modeling more finely biological objects leads to increase the number of equations. Simulating ever larger systems also leads to increasing the number of equations. Therefore, we observe a large increase in the size of the ODE systems to be solved. A major lock is the limitation of ODE numerical resolution software (ODE solver) to a few thousand equations due to prohibitive calculation time. The AEx ExODE tackles this lock via 1) the introduction of new numerical methods that will take advantage of the mixed precision that mixes several floating number precisions within numerical methods, 2) the adaptation of these new methods for next generation highly hierarchical and heterogeneous computers composed of a large number of CPUs and GPUs. For the past year, a new approach to Deep Learning has been proposed to replace the Recurrent Neural Network (RNN) with ODE systems. The numerical and parallel methods of ExODE will be evaluated and adapted in this framework in order to improve the performance and accuracy of these new approaches.</p>https://www.inria.fr/en/exode