1,722,571 research outputs found
Using SVM to combine global heuristics for the Standard Quadratic Problem
No full textThe Standard Quadratic Problem (StQP) is an NP-hard problem with many local minimizers (stationary
points). In the literature, heuristics based on unconstrained continuous non-convex formulations have been
proposed (Bomze & Palagi, 2005; Bomze, Grippo, & Palagi, 2012) but none dominates the other in terms of
best value found. Following (Cassioli, DiLorenzo, Locatelli, Schoen, & Sciandrone, 2012) we propose to use
Support Vector Machines (SVMs) to define a multistart global strategy which selects the “best” heuristic.
We test our method on StQP arising from the Maximum Clique Problem on a graph which is a challenging
combinatorial problem. We use as benchmark the clique problems in the DIMACS challenge
Séminaire du 10/10/2016 - E. Palagi
No full textPremier séminaire des lundis de l'ergonomie. Programme Deux présentations vous seront proposées pour l'inauguration de ce cycle de séminaires par les deux organisatrices. Recherche exploratoire et élaboration de méthodes centrées utilisateurs Emilie Palagi, doctorante en Ergonomie des Interactions Homme-Machine (IHM) au sein de l'équipe Wimmics (Inria & I3S) et EURECOM. Résumé : Les moteurs de recherche exploratoire sont des applications qui ont pour objectif premier d'aider les utilisateurs..
Gli anni bolognesi Palagi - Regesto cronologico
No full textIl contributo è rivolto all’indagine degli anni giovanili e della prima attività di Palagi, in relazione alla pittura del Basoli ed alla committenza cittadina a cavallo tra XVIII e XIX secolo. Il regesto cronologico tratta la vita del pittore bolognese, dalla nascita alla sua partenza per Roma avvenuta nel 1805
[Bianca Palagi (1913), funerary sculpture]
No full textFrom Berresford: Bianca Palagi (1913), Antonio Bozzano, Cimitero Comunale, Viareggio.Depiction of girl hit by car.Title from Berresford
Getting rid of blinkers: the case of mirror self-recognition in horses (Equus caballus)
No full textThe commentary by Gallup and Anderson (Anim Cogn tittps://doi.org/10.1007/s10071-021-01538-9, 2021) on the original article by Baragli, Scopa, Maglieri, and Palagi (Anim Cogn https://doi.org/10.1007/s10071-021-01502-7, 2021) raised some concerns about the methodological approach used by the authors to demonstrate Mirror Self-Recognition (MSR) in horses. The commentary does not take into account horse physiology and psychology, leading Gallup and Anderson to inappropriately discredit the findings obtained by Baragli et al. Anim Cogn 2021. In this reply, we underlined the importance of a blinker-free approach to understand the evolutionary processes at the basis of animal cognition
Social play in bonobos (Pan paniscus) and chimpanzees (Pan troglodytes): Implications for natural social systems and inter-individual relationships.
No full textThis study compares adult play behavior
in the two Pan species in order to test the effects of phylogenetic
closeness and the nature of social systems on
play distribution. The social play (both with fertile and
immature subjects) performed by adults did not differ
between the two species. In contrast, in bonobos, play
levels among fertile subjects were higher than in chimpanzees.
Findings regarding levels of undecided conflicts
(more frequent in bonobos) and formal submission displays
(lacking in bonobos) confirm, in the two colonies
under study, that bonobos exhibit ‘‘egalitarianism’’ more
than chimpanzees. Some authors emphasized the importance
of play-fighting for social assessment when relationships
among individuals are not codified and structured
according to rank-rules. Indeed, adult bonobos
played more roughly than chimpanzees. Moreover, adult
bonobos displayed the full play-face at a high frequency
especially during rough play sessions, whereas in chimpanzees,
the frequency of play signals was not affected
by roughness of play. The frequency of social play among
bonobo females was higher than in any other sex combinations,
whereas no difference was found for chimpanzees.
As a matter of fact, social play can be viewed as a
balance between cooperation and competition. Among
bonobo females, characterized by social competence and
affiliation, social play might enhance their behavioral
flexibility and increase their socially symmetrical relationships
which, after all, are the basis for their egalitarian
society
Adult play and the evolution of tolerant and cooperative societies
Full text linkPlay is generally considered an immature affair. However, adult play is present in several mammal species living in complex social systems. Here, I hypothesize that adult social play is favored by natural selection in those species characterized by high level of social tolerance and/or by the need of others’ cooperation to reach a goal (i.e., leverage). The integration and comparison of bio-behavioral data on non-human primates and wild social carnivores allows drawing a comprehensive picture on the importance of adult play in facing unpredictable, novel social situations and in overcoming stressful experiences. The ability to cope with potentially competitive interactions through play can favor the emergence of egalitarian societies. A further interesting and beneficial aspect of adult play is its role in synchronizing group activities and favoring collective decision making by renovating the motivation to cooperate in groupmates. As a last step, some considerations about the presence of adult play in the most egalitarian and cooperative human groups (e.g., hunter-gatherer societies) allows discussing the apparent dichotomy between cultural and biological evolution of certain behavioral traits, including social play in adulthood
Global optimization issues in deep network regression: an overview
No full textThe paper presents an overview of global issues in optimizationmethods for training feedforward
neural networks (FNN) in a regression setting.We first recall the learning optimization
paradigm for FNN and we briefly discuss global scheme for the joint choice of the network
topologies and of the network parameters. The main part of the paper focuses on the
core subproblem which is the continuous unconstrained (regularized) weights optimization
problem with the aim of reviewing global methods specifically arising both in multi layer
perceptron/deep networks and in radial basis networks.We review some recent results on the
existence of non-global stationary points of the unconstrained nonlinear problem and the role
of determining a global solution in a supervised learning paradigm. Local algorithms that are
widespread used to solve the continuous unconstrained problems are addressed with focus on
possible improvements to exploit the global properties. Hybrid global methods specifically
devised for FNN training optimization problems which embed local algorithms are discussed
too
- …
