This document analyzes screening strategies for COVID-19. It discusses analyzing different screening strategies using virological and serological tests given limited testing capacity over 10 months. Five strategies are proposed with different priority levels. Strategy 1 adopted in France performed worst, while strategies 4 and 5 performed best across criteria like percentage of untested patients and stock breaks. The analysis found importance for logistics in health crises and formalized screening strategy optimization.
On the Modeling of the Three Types of Non-Spiking Neurons of the Caenorhabdit...Juan Luis Jiménez Laredo
The nematode Caenorhabditis elegans (C. elegans) is a well-known model organism in neuroscience. The relative simplicity of its nervous system, made up of few hundred neurons, shares some essential features with more sophisticated nervous systems, including the human one. If we are able to fully characterize the nervous system of this organism, we will be one step closer to understanding the mechanisms underlying the behavior of living things. Following a recently conducted electrophysiological survey on different C. elegans neurons, this paper aims at modeling the three non-spiking RIM, AIY and AFD neurons (arbitrarily named with three upper case letters by convention). To date, they represent the three possible forms of non-spiking neuronal responses of the C. elegans. To achieve this objective, we propose a conductance-based neuron model adapted to the electrophysiological features of each neuron. These features are based on current biological research and a series of in-silico experiments which use differential evolution to fit the model to experimental data. From the obtained results, we formulate a series of biological hypotheses regarding currents involved in the neuron dynamics. These models reproduce experimental data with a high degree of accuracy while being biologically consistent with state-of-the-art research.
Au cours des derniers mois, la Blockchain a connu un engouement très important, surtout à cause du buzz qui a été fait autour des crypto-monnaies telles que Bitcoin. Si la Blockchain est la technologie qui se cache derrière ces dernières, elle n'est pas uniquement cela et promet de permettre de construire des applications décentralisées (et sans tiers de confiance) dans divers domaines d'application où un tiers de confiance était jusque là nécessaire. Dans cette présentations, nous présentons les principes de fonctionnement de la technologie Blockchain avec tous les mécanismes sous-jacents qui pour un grand nombre sont antérieurs à la Blockchain (réseaux pair-à-pair, cryptographie, fonctions de condensat, etc.).
This document analyzes screening strategies for COVID-19. It discusses analyzing different screening strategies using virological and serological tests given limited testing capacity over 10 months. Five strategies are proposed with different priority levels. Strategy 1 adopted in France performed worst, while strategies 4 and 5 performed best across criteria like percentage of untested patients and stock breaks. The analysis found importance for logistics in health crises and formalized screening strategy optimization.
On the Modeling of the Three Types of Non-Spiking Neurons of the Caenorhabdit...Juan Luis Jiménez Laredo
The nematode Caenorhabditis elegans (C. elegans) is a well-known model organism in neuroscience. The relative simplicity of its nervous system, made up of few hundred neurons, shares some essential features with more sophisticated nervous systems, including the human one. If we are able to fully characterize the nervous system of this organism, we will be one step closer to understanding the mechanisms underlying the behavior of living things. Following a recently conducted electrophysiological survey on different C. elegans neurons, this paper aims at modeling the three non-spiking RIM, AIY and AFD neurons (arbitrarily named with three upper case letters by convention). To date, they represent the three possible forms of non-spiking neuronal responses of the C. elegans. To achieve this objective, we propose a conductance-based neuron model adapted to the electrophysiological features of each neuron. These features are based on current biological research and a series of in-silico experiments which use differential evolution to fit the model to experimental data. From the obtained results, we formulate a series of biological hypotheses regarding currents involved in the neuron dynamics. These models reproduce experimental data with a high degree of accuracy while being biologically consistent with state-of-the-art research.
Au cours des derniers mois, la Blockchain a connu un engouement très important, surtout à cause du buzz qui a été fait autour des crypto-monnaies telles que Bitcoin. Si la Blockchain est la technologie qui se cache derrière ces dernières, elle n'est pas uniquement cela et promet de permettre de construire des applications décentralisées (et sans tiers de confiance) dans divers domaines d'application où un tiers de confiance était jusque là nécessaire. Dans cette présentations, nous présentons les principes de fonctionnement de la technologie Blockchain avec tous les mécanismes sous-jacents qui pour un grand nombre sont antérieurs à la Blockchain (réseaux pair-à-pair, cryptographie, fonctions de condensat, etc.).
Spatially structured Metaheuristics: Principles and Practical ApplicationsJuan Luis Jiménez Laredo
A relevant number of metaheuristics are based on population. Although conventions may establish different names, individuals in evolutionary algorithms, ants in ant colony optimization or particles in particle swarm optimization belong to the same side of a coin: they are all atomic elements of the population (a.k.a. building-blocks). In this context, spatially structured metaheuristics investigate how to improve the performance of metaheuristics by confining these elements in neighborhoods. This talk aims at presenting the working principles of spatially structured metaheuristics and practical applications to enhance diversity, scalability and robustness.
1) Tournament selection has issues of not sampling all individuals in a population and not selecting some individuals.
2) Cooperative selection is proposed as an improvement that is more altruistic. It works by having the better of two randomly sampled individuals be selected for breeding, while the other individual is helped by taking the average fitness of the two.
3) This process is illustrated through an example population where cooperative selection results in all individuals being both sampled and selected for breeding, avoiding the problems with tournament selection.
1) The document compares pool-based and island-based evolutionary algorithms (EAs) using CouchDB as the shared memory.
2) A proof of concept experiment evaluates pool-based EAs, island-based EAs, and a baseline EA on the OneMax problem using CouchDB for concurrent access and persistence.
3) The results show that an elite pool-based EA accessed CouchDB 25 times more than an island-based EA but achieved a higher speedup of 2.5 compared to 1.7 for the island model, demonstrating that the pool model can improve performance over the baseline and island models.
This document describes research validating a peer-to-peer evolutionary algorithm model in a real-world environment. The researchers conducted large-scale experiments on over 3008 computing agents distributed across 188 computers to test the scalability of their peer-to-peer evolutionary computation model. Their results showed seamless scalability up to the large number of agents used. Compared to prior simulation-based studies, the experiments validated that the parallel evolutionary algorithm could solve problems with similar or better performance than sequential algorithms, while achieving near-linear speedup through massive parallelization.
Analysing the Performance of Different Population Structures for an Agent-bas...Juan Luis Jiménez Laredo
The document compares the performance of different population structures (ring, Watts-Strogatz, newscast) for a peer-to-peer evolutionary algorithm. It models a fine-grained parallel evolutionary algorithm using a peer-to-peer protocol as the underlying population structure. The experimental analysis tests these population structures on a 2-trap problem and finds that regular lattices require smaller population sizes but more evaluations to find a solution, while different small-world methods produce equivalent performance. Future work is proposed to validate the model on a real peer-to-peer infrastructure and explore other population structures and metaheuristics.
The document discusses peer-to-peer evolutionary computation (P2P EC) as a parallel solution for hard optimization problems. It aims to find empirical evidence of the viability of the P2P EC paradigm. The model uses a fine-grained parallel evolutionary algorithm with a peer-to-peer protocol as the underlying population structure. Three test cases are analyzed to evaluate the scalability and fault tolerance of the approach.
This document presents research on improving genetic algorithms via deterministic population shrinkage. The researchers hypothesize that genetic algorithms require different population sizes at different convergence stages, with larger populations needed early on. They propose a Simple Variable Population Sizing (SVPS) scheme where the population shrinks deterministically over generations. Experiments on generalized l-trap functions of varying difficulty show that SVPS can improve performance over a fixed population size by minimizing the number of solutions evaluated while maintaining solution quality.
This document discusses addressing churn in a peer-to-peer evolutionary algorithm. It presents the key issues of decentralization, scalability, and fault tolerance in P2P EAs. The experimental setup models churn using a Weibull distribution and evaluates the P2P EA on massively multimodal deceptive problems of varying size under different churn rates. The results show the P2P EA maintains near optimal solutions even with high churn, with less than 1% of the initial population lost for the largest problem size tested.
This document discusses using peer-to-peer (P2P) networks to tackle large problem instances in difficult optimization problems. It proposes applying P2P technology to evolutionary algorithms (EAs) to achieve massive scalability. The authors outline introducing P2P concepts to EAs, their model for P2P evolutionary algorithms, the experimental setup, results, and conclusions.
Spatially structured Metaheuristics: Principles and Practical ApplicationsJuan Luis Jiménez Laredo
A relevant number of metaheuristics are based on population. Although conventions may establish different names, individuals in evolutionary algorithms, ants in ant colony optimization or particles in particle swarm optimization belong to the same side of a coin: they are all atomic elements of the population (a.k.a. building-blocks). In this context, spatially structured metaheuristics investigate how to improve the performance of metaheuristics by confining these elements in neighborhoods. This talk aims at presenting the working principles of spatially structured metaheuristics and practical applications to enhance diversity, scalability and robustness.
1) Tournament selection has issues of not sampling all individuals in a population and not selecting some individuals.
2) Cooperative selection is proposed as an improvement that is more altruistic. It works by having the better of two randomly sampled individuals be selected for breeding, while the other individual is helped by taking the average fitness of the two.
3) This process is illustrated through an example population where cooperative selection results in all individuals being both sampled and selected for breeding, avoiding the problems with tournament selection.
1) The document compares pool-based and island-based evolutionary algorithms (EAs) using CouchDB as the shared memory.
2) A proof of concept experiment evaluates pool-based EAs, island-based EAs, and a baseline EA on the OneMax problem using CouchDB for concurrent access and persistence.
3) The results show that an elite pool-based EA accessed CouchDB 25 times more than an island-based EA but achieved a higher speedup of 2.5 compared to 1.7 for the island model, demonstrating that the pool model can improve performance over the baseline and island models.
This document describes research validating a peer-to-peer evolutionary algorithm model in a real-world environment. The researchers conducted large-scale experiments on over 3008 computing agents distributed across 188 computers to test the scalability of their peer-to-peer evolutionary computation model. Their results showed seamless scalability up to the large number of agents used. Compared to prior simulation-based studies, the experiments validated that the parallel evolutionary algorithm could solve problems with similar or better performance than sequential algorithms, while achieving near-linear speedup through massive parallelization.
Analysing the Performance of Different Population Structures for an Agent-bas...Juan Luis Jiménez Laredo
The document compares the performance of different population structures (ring, Watts-Strogatz, newscast) for a peer-to-peer evolutionary algorithm. It models a fine-grained parallel evolutionary algorithm using a peer-to-peer protocol as the underlying population structure. The experimental analysis tests these population structures on a 2-trap problem and finds that regular lattices require smaller population sizes but more evaluations to find a solution, while different small-world methods produce equivalent performance. Future work is proposed to validate the model on a real peer-to-peer infrastructure and explore other population structures and metaheuristics.
The document discusses peer-to-peer evolutionary computation (P2P EC) as a parallel solution for hard optimization problems. It aims to find empirical evidence of the viability of the P2P EC paradigm. The model uses a fine-grained parallel evolutionary algorithm with a peer-to-peer protocol as the underlying population structure. Three test cases are analyzed to evaluate the scalability and fault tolerance of the approach.
This document presents research on improving genetic algorithms via deterministic population shrinkage. The researchers hypothesize that genetic algorithms require different population sizes at different convergence stages, with larger populations needed early on. They propose a Simple Variable Population Sizing (SVPS) scheme where the population shrinks deterministically over generations. Experiments on generalized l-trap functions of varying difficulty show that SVPS can improve performance over a fixed population size by minimizing the number of solutions evaluated while maintaining solution quality.
This document discusses addressing churn in a peer-to-peer evolutionary algorithm. It presents the key issues of decentralization, scalability, and fault tolerance in P2P EAs. The experimental setup models churn using a Weibull distribution and evaluates the P2P EA on massively multimodal deceptive problems of varying size under different churn rates. The results show the P2P EA maintains near optimal solutions even with high churn, with less than 1% of the initial population lost for the largest problem size tested.
This document discusses using peer-to-peer (P2P) networks to tackle large problem instances in difficult optimization problems. It proposes applying P2P technology to evolutionary algorithms (EAs) to achieve massive scalability. The authors outline introducing P2P concepts to EAs, their model for P2P evolutionary algorithms, the experimental setup, results, and conclusions.
1. Tas de sable dans
un tamis
Juan Luis Jiménez Laredo
Réunion 17.04.2018 Ri2C
2. « Que savons-nous si des créations
de mondes ne sont point déterminées
par des chutes de grains de sable ?
Qui donc connaît les flux et les
reflux réciproques de l'infiniment
grand et de l'infiniment petit, le
retentissement des causes dans les
précipices de l'être et les avalanches
de la création ? »
Les Misérables Victor Hugo
3. Avez-vous entendu parler du tas de sable?
Automate cellulaire par Bak, Tang & Wiesenfeld 1987
Règles
1. Les grains de sable sont déposés au hasard
2. Si un grain est déposé sur le site (x, y) sa hauteur change:
3. Si la hauteur d'un site est supérieure à un seuil h(x,y) ≥ 4
1),(),( yxhyxh
1)1,()1,(
1),1(),1(
4),(),(
yxhyxh
yxhyxh
yxhyxh
7. Exemple
Un grain est déposé au hasard
3. Si la hauteur d'un site est
supérieure à un seuil h(x,y) ≥ 4
1)1,()1,(
1),1(),1(
4),(),(
yxhyxh
yxhyxh
yxhyxh
14. Les résultats dans une courbe Fréquence / Taille
Dynamique d'exécution
L'une des empreintes digitales de SOC est la
relation de loi de puissance entre la taille et la
fréquence des événements
15. Peut-on l'utiliser pour équilibrer la charge?
Au lieu de sable, pensez aux tâches
et pensez aux sites en tant qu'éléments de calcul
Le tas de sable est dans un tamis !
16. RésultatsConfiguration expérimentale
A set of P = {p1 , . . . , p100 } processors
arranged in a 10 × 10 toroidal grid .
Each processor has a processing speed of 1 instr./cycle.
A set of workloads N = {N1, . . . , N120},
each Ni composed of 10000 tasks ‹n, a› of length (n) inst.
and arrival time (a) such that:
Ni = {‹n, 1›, . . . , ‹n, 10000›}
where .: instinNn i
The sandpile load-balancer is naturally driven to critical
states when the incoming workload is equal to the
maximum throughput of the system
17. Merci de votre attention !
En savoir plus :
Juan Luis Jiménez Laredo, Frédéric Guinand, Damien Olivier, Pascal Bouvry:
Load Balancing at the Edge of Chaos: How Self-Organized Criticality Can Lead to Energy-Efficient
Computing. IEEE Trans. Parallel Distrib. Syst. 28(2): 517-529 (2017)
Juan Luis Jiménez Laredo, Frédéric Guinand, Damien Olivier, Pascal Bouvry:
Trading Off Resource Utilization and Task Migrations in Dynamic Load-balancing. GECCO
(Companion) 2015: 1409-1410
Juan Luis Jiménez Laredo, Pascal Bouvry, Frédéric Guinand, Bernabé Dorronsoro, Carlos M.
Fernandes:
The sandpile scheduler - How self-organized criticality may lead to dynamic load-
balancing. Cluster Computing 17(2): 191-204 (2014)
Juan Luis Jiménez Laredo, Bernabé Dorronsoro, Johnatan E. Pecero, Pascal Bouvry, Juan José
Durillo, Carlos M. Fernandes:
Designing a Self-Organized Approach for Scheduling Bag-of-Tasks. 3PGCIC 2012: 315-320
Dune by nunavut @ flickr
https://creativecommons.org/licenses/by-nc/2.0/