- semfis - Mailing lists@ IFISC

Talk on Tuesday, Nov 22, 2022 at 11:00
by semfis＠ifisc.uib-csic.es 18 Nov '22

18 Nov '22

Talk on Tuesday Nov 22, 2022 at 11:00 Place: IFISC Seminar Room Title: Benchmarking the Role of Particle Statistics in Quantum Reservoir Computing Speaker: Guillem Llodrà, IFISC Abstract: MdM Quantum Meeting Quantum reservoir computing is a neuro-inspired machine learning approach harnessing the rich dynamics of quantum systems to solve temporal tasks. &nbsp;It has gathered attention for its suitability for NISQ devices, for easy and fast trainability, and for potential quantum advantage. Although several types of systems have been proposed as quantum reservoirs, &nbsp;differences arising from particle statistics have not been established yet. In this work, we assess and compare the ability of bosons, fermions, and qubits to store information from past inputs by measuring linear and nonlinear memory capacity. While, in general, the performance of the system improves with the Hilbert space size, we show that also the information spreading capability is a key factor. For the simplest reservoir Hamiltonian choice, and for each boson limited to at most one excitation, fermions provide the best reservoir due to their intrinsic nonlocal properties. On the other hand, a tailored input injection strategy allows the exploitation of the abundance of degrees of freedom of the Hilbert space for bosonic quantum reservoir computing and enhances the computational power compared to both qubits and fermions. The talk will be broadcasted at: <a href="https://uibuniversitat.zoom.us/j/93820093409" style="word-wrap: break-word;" target="_blank">https://uibuniversitat.zoom.us/j/93820093409</a> Url event:https://ifisc.uib-csic.es/en/events/seminars/benchmarking-the-role-of-particle-statistics-in-qu/

1 0

IFISC Seminar on Wednesday, Nov 23, 2022 at 14:30
by semfis＠ifisc.uib-csic.es 18 Nov '22

18 Nov '22

IFISC Seminar on Wednesday Nov 23, 2022 at 14:30 Place: Zoom Seminar Title: Complexity Economics of Global Change Speaker: Roger Cremades, Wageningen University, Wageningen, The Netherlands Abstract: Ranging from system dynamics models to network science, the evolution of the intersection between complex system science and the human and economic dimensions of global change achieved important milestones during the last 10 years. A summary of contributions to this interdisciplinary intersection will be provided, showing how is it evolving towards the Complexity Economics of Global Change. While keeping the overall perspective on the sustainable use of scarce natural resources like water, land and energy, and on how climate change alters their balances, the range of scales involved will be explored. First, on human behaviour and its psychological and biophysical context. Second, on cities as social tipping elements and grounds for information cascades. Third, on river basins as systems under urbanisation and global change. And finally fourth, on national and global scale economic models capturing different phenomena previously analysed with limited neoclassical economic s. Zoom link:&nbsp;&nbsp;https://zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09 Url event:https://ifisc.uib-csic.es/en/events/seminars/complexity-economics-of-global-change/

1 0

Reminder: IFISC Seminar, TODAY - Wednesday, Nov 16, 2022 at 14:30
by semfis＠ifisc.uib-csic.es 16 Nov '22

16 Nov '22

IFISC Seminar on Wednesday Nov 16, 2022 at 14:30 Place: IFISC Seminar Room Title: A Lagrangian view on the seascape: theory and applications Speaker: Enrico Ser-Giacomi, MIT - IFISC Abstract: Abstract: The Lagrangian perspective on flow systems consists in adopting a frame of reference that is solidal with fluid particles. In the last decades, such framework has demonstrated a great efficacy in addressing fluid dynamics problems as well as in providing theoretical connections with Dynamical Systems and Network Theory. In this presentation I will focus on the oceanic environment and I will show how Lagrangian approaches can be used at different spatio-temporal scales to tackle complementary scientific objectives. At the micro-scale, one can study the advection, deformation and dilution of a single Lagrangian particle and link such processes with the evolution of bio-chemical fields in it contained. At the macro-scale, one can instead characterize the spatial patterns generated by the tangle of several Lagrangian trajectories under the effect of ocean turbulence. Following this paradigm, I will present recent theoretical developments as well as applications to Climate and Marine Ecology. Zoom coordinates: https://us06web.zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09&… Url event:https://ifisc.uib-csic.es/en/events/seminars/a-lagrangian-view-on-the-seascape-theory-and-appli/

1 0

IFISC Seminar on Wednesday, Nov 16, 2022 at 14:30
by semfis＠ifisc.uib-csic.es 11 Nov '22

11 Nov '22

IFISC Seminar on Wednesday Nov 16, 2022 at 14:30 Place: IFISC Seminar Room Title: A Lagrangian view on the seascape: theory and applications Speaker: Enrico Ser-Giacomi, MIT - IFISC Abstract: Abstract: The Lagrangian perspective on flow systems consists in adopting a frame of reference that is solidal with fluid particles. In the last decades, such framework has demonstrated a great efficacy in addressing fluid dynamics problems as well as in providing theoretical connections with Dynamical Systems and Network Theory. In this presentation I will focus on the oceanic environment and I will show how Lagrangian approaches can be used at different spatio-temporal scales to tackle complementary scientific objectives. At the micro-scale, one can study the advection, deformation and dilution of a single Lagrangian particle and link such processes with the evolution of bio-chemical fields in it contained. At the macro-scale, one can instead characterize the spatial patterns generated by the tangle of several Lagrangian trajectories under the effect of ocean turbulence. Following this paradigm, I will present recent theoretical developments as well as applications to Climate and Marine Ecology. Zoom coordinates: https://us06web.zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09&… Url event:https://ifisc.uib-csic.es/en/events/seminars/a-lagrangian-view-on-the-seascape-theory-and-appli/

1 0

Reminder: IFISC Seminar, TODAY - Wednesday, Nov 02, 2022 at 14:30
by semfis＠ifisc.uib-csic.es 02 Nov '22

02 Nov '22

IFISC Seminar on Wednesday Nov 02, 2022 at 14:30 Place: IFISC Seminar Room Title: Learn one size to infer all: Exploiting symmetries in dynamical systems using scalable neural networks Speaker: Mirko Goldmann, IFISC Abstract: Tackling unavailable analytical models, machine learning methods emerged that learn from large amounts of data to forecast the dynamics of complex systems. Yet such predictions are mainly restricted to the dynamical regime observed during training. Here, we attempt to answer the question - Is it possible to infer untrained (size-dependent) dynamical regimes of a complex system while learning from one example related to a certain system size only? Therefore, we design scalable neural networks capable of inferring untrained high-dimensional dynamics for different system sizes. We train these networks to predict the dynamics of delay-dynamical and spatio-temporal systems for a single size. Then, we drive the network by its own predictions. Subsequently, by scaling the size of the trained network, we can predict complex dynamics for larger or smaller system sizes. Thus, the network learns from a single example and by exploiting symmetry properties, infers entire bifurcation diagrams. Presential seminar, with parallel Zoom stream:&nbsp;&nbsp; https://zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09 Url event:https://ifisc.uib-csic.es/en/events/seminars/learn-one-size-to-infer-all-exploiting-symmetries/

1 0

IFISC Seminar on Wednesday, Nov 02, 2022 at 14:30
by semfis＠ifisc.uib-csic.es 28 Oct '22

28 Oct '22

IFISC Seminar on Wednesday Nov 02, 2022 at 14:30 Place: IFISC Seminar Room Title: Learn one size to infer all: Exploiting symmetries in dynamical systems using scalable neural networks Speaker: Mirko Goldmann, IFISC Abstract: Tackling unavailable analytical models, machine learning methods emerged that learn from large amounts of data to forecast the dynamics of complex systems. Yet such predictions are mainly restricted to the dynamical regime observed during training. Here, we attempt to answer the question - Is it possible to infer untrained (size-dependent) dynamical regimes of a complex system while learning from one example related to a certain system size only? Therefore, we design scalable neural networks capable of inferring untrained high-dimensional dynamics for different system sizes. We train these networks to predict the dynamics of delay-dynamical and spatio-temporal systems for a single size. Then, we drive the network by its own predictions. Subsequently, by scaling the size of the trained network, we can predict complex dynamics for larger or smaller system sizes. Thus, the network learns from a single example and by exploiting symmetry properties, infers entire bifurcation diagrams. Presential seminar, with parallel Zoom stream:&nbsp;&nbsp; https://zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09 Url event:https://ifisc.uib-csic.es/en/events/seminars/learn-one-size-to-infer-all-exploiting-symmetries/

1 0

Reminder: IFISC Seminar, TODAY - Wednesday, Oct 26, 2022 at 14:30
by semfis＠ifisc.uib-csic.es 26 Oct '22

26 Oct '22

IFISC Seminar on Wednesday Oct 26, 2022 at 14:30 Place: IFISC Seminar Room Title: How not to construct functional brain networks Speaker: Onerva Korhonen, Aalto University Abstract: Human brain function relies on interactions between brain areas. Therefore, a complex network appears as a natural model for the brain, and network science indeed holds potential for understanding the brain. However, constructing a functional brain network is not trivial. I will address common problems in brain network construction, concentrating on a step that strongly affects obtained network properties: definition of network nodes. Using multiple fMRI datasets, I will show that the currently used node definition strategies produce nodes that suffer from low functional homogeneity, which may lead to spuriosity in observed network structure and data losses. Further, I will show that functional homogeneity changes in time, reflecting nodes&rsquo; roles in network topology. Finally, as a potential solution for node definition problem, I will introduce a flexible node definition strategy that accounts for the time-dependent nature of brain function. Presential seminar, with parallel Zoom stream:&nbsp;&nbsp; https://zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09 Url event:https://ifisc.uib-csic.es/en/events/seminars/how-not-to-construct-functional-brain-networks/

1 0

Reminder: Master Thesis, TODAY - Monday, Oct 24, 2022 at 10:20
by semfis＠ifisc.uib-csic.es 24 Oct '22

24 Oct '22

Master Thesis on Monday Oct 24, 2022 at 10:20 Place: IFISC Meeting Room Title: Analysis and the Action Curve of Agent Zero: Theory and Simulation Speaker: Benjamin Maxwell Fried, IFISC / UIB Abstract: &nbsp; In this thesis I explore a theoretical neurocognitively ground agent put forth by Joshua Epstein to advance social science research. In particular I begin to develop a probabilistic framework in which theoretical predictions can be made about the dynamics of Agent Zero. Furthermore, I study the collective behavior of an ensemble of Einstein&#39;s agents and compare the results with a modified version. Supervisor: Maxi San Miguel Jury: Tobias Galla, Maxi San Miguel, Raul Toral Presential in the IFISC meeting room (room 215), with parallel Zoom stream:&nbsp;https://zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09 Url event:https://ifisc.uib-csic.es/en/events/seminars/analysis-and-the-action-curve-of-agent-zero-theory/

1 0

Reminder: Talk, TODAY - Monday, Oct 24, 2022 at 10:00
by semfis＠ifisc.uib-csic.es 24 Oct '22

24 Oct '22

Talk on Monday Oct 24, 2022 at 10:00 Place: IFISC Seminar Room Title: Time delay reservoir computing with short-external-cavity laser dynamics and phase encoding Speaker: Lucas Talandier, IFISC (CSIC-UIB) Abstract: Semiconductor lasers with optical feedback or optical injection have been studied in recent years, especially for their capabilities for complex signal generation and information processing. The nonlinear behavior of this system can be used for solving nonlinear computing tasks, when operating in a transient input transformation regime and exploiting memory induced by the feedback. For tasks that require memory, increasing the feedback rate is often beneficial. However, this might be accompanied by the onset of complex dynamics resulting in inconsistent input signal transformations, which are detrimental for computing. In this talk, I will discuss a semiconductor laser system that operates in a short-cavity regime, for which chaotic dynamics appear only under strong feedback conditions. The extended feedback operating regime avoiding chaos, compared to the long-cavity regime, allows to realize better memory properties. I will discuss the&nbsp;trade-off between memory , speed of computation, dimensionality, and consistency. I will show examples of various parameter conditions, identifying short external cavity regimes with strong optical feedback and injection, which offer high memory capacity avoiding the laser&rsquo;s coherence collapse. I will also discuss how this system behaves when the input information is phase-encoded, and I will compare its performance to the commonly used amplitude encoding. Zoom:&nbsp;https://uibuniversitat.zoom.us/j/87008849910 Url event:https://ifisc.uib-csic.es/en/events/seminars/time-delay-reservoir-computing-with-short-external/

1 0

Talk on Monday, Oct 24, 2022 at 10:00
by semfis＠ifisc.uib-csic.es 21 Oct '22

21 Oct '22

Talk on Monday Oct 24, 2022 at 10:00 Place: IFISC Seminar Room Title: Time delay reservoir computing with short-external-cavity laser dynamics and phase encoding Speaker: Lucas Talandier, IFISC (CSIC-UIB) Abstract: Semiconductor lasers with optical feedback or optical injection have been studied in recent years, especially for their capabilities for complex signal generation and information processing. The nonlinear behavior of this system can be used for solving nonlinear computing tasks, when operating in a transient input transformation regime and exploiting memory induced by the feedback. For tasks that require memory, increasing the feedback rate is often beneficial. However, this might be accompanied by the onset of complex dynamics resulting in inconsistent input signal transformations, which are detrimental for computing. In this talk, I will discuss a semiconductor laser system that operates in a short-cavity regime, for which chaotic dynamics appear only under strong feedback conditions. The extended feedback operating regime avoiding chaos, compared to the long-cavity regime, allows to realize better memory properties. I will discuss the&nbsp;trade-off between memory , speed of computation, dimensionality, and consistency. I will show examples of various parameter conditions, identifying short external cavity regimes with strong optical feedback and injection, which offer high memory capacity avoiding the laser&rsquo;s coherence collapse. I will also discuss how this system behaves when the input information is phase-encoded, and I will compare its performance to the commonly used amplitude encoding. Zoom:&nbsp;https://uibuniversitat.zoom.us/j/87008849910 Url event:https://ifisc.uib-csic.es/en/events/seminars/time-delay-reservoir-computing-with-short-external/

1 0