Consultar ensayos de calidad
The changing face of methodological - individualism
THE CHANGING FACE OF METHODOLOGICAL
INDIVIDUALISM
CÓD: 53073935
RESEÑA DEL AUTOR: Catedratico de Ciencias Sociales de la
Universidad M¨alardalen. Su investigación centra
en temas de Economía y ciencias políticas. Tiene
publicaciones tales Como, The Limits of Public
Choice: A Sociological Critique of the Economic Theory of Politics
, Methodological Individualism
ANTECEDENTES: La primera definición del
concepto en el ambito de la economía la estableció
Schumpeter en 1908: el individualismo metodológico es una forma de
analisis que se inicia siempre a partir del comportamiento individual. No obstante,
ha sido el filósofo de la ciencia Karl Popper quien con mas
fuerza (e impacto) ha sostenido que este principio debe ser la piedra angular
de todas las ciencias sociales: todas las instituciones sociales y los
fenómenos colectivos en realidad son abstracciones hipotéticas
derivadas necesariamente de las decisiones de los individuos: “La tarea
de de las ciencias sociales consiste en construir y analizar nuestros modelos
sociológicos con todo cuidado en términos descriptivos o
nominalistas, es decir, en términos de los individuos, de sus actitudes,
expectativas, relaciones etc.” (Popper, 1957
Según este principio, los fenómenos
sociales son producidos por la interacción (voluntaria o involuntaria)
de los individuos. Esto no quiere decir que el resultado social haya sido
buscado conscientementey que sea el mejor posible, sino que la
conjunción de decisiones sociales conduce a un
determinado tipo de fenómeno social. Intentar entender el
fenómeno social como
si las decisiones individuales no fueran relevantes equivaldría a
renunciar a una verdadera explicación, ya que el fenómeno social
quedaría descrito como
una especie de caja negra.
PROBLEMATICA: En este articulo, el autor se centra en describir la
batalla constante sobre la verdadera naturaleza de la sociedad que ha tenido
lugar a lo largo de la historia del pensamiento social, y la mejor manera de
entender y explicarlo; El autor desarrolla el individualismo
metodológico, en todas sus características y con todos sus
componentes, tomando al individuo como un conjunto de componentes que lo hacen
decidir y elegir de acuerdo a ese conjunto y no solo desde su parte racional
HIPOTESIS DEL AUTOR:El autor plantea que no es fructífero concebir el
individualismo metodológico como una corriente que busca la
reducción de las leyes sociales. La alternativa mas
fructífera, es ver el individualismo metodológico como
principio de la explicación de los fenómenos sociales que tienen
lugar en el entorno de los individuos. Separando el individualismo
metodológico y el holistico se desdibujan las dos doctrinas ya que no
aparecen como
opuestas en su totalidad.
PRINCIPALES ARGUMENTOS QUE CORROBORAN O CONTRADICEN LA HIPÓTESIS:
PRINCIPALES RESULTADOS
POSTURA CRÍTICA:
II. LEARNING MOTION WITH
AN ARTICULATED ARM
Machine learning is one area where we expect algorithms inspired by
attentional selection strategy to outperform conventional ones. There are
several ways in which attention might facilitate learning. One is during
learning; if shown a single image of a car embedded in a dense background ï¬lled
with other objects, the learning algorithm does not know which features belong
to the object of relevance (here the car) and which ones are incidental. If
attention would segment the car from the rest of the scene, however, superior
performance can be obtained. This is particu-
larly relevant to one-shot learning algorithms. The same is true during the
recognition phase. Detecting the same car, say, under a different viewpoint, in
a novel scene is much facilitated if an attentional selection strategy can
segment the car from the background and just forward its associated features to
the recognition module (see Rutishauser et al., 2004, for an illustration of
this strategy). Of course, segmentation also helps in reducing the amount of
data that must be memorized
,thus improving learning
speed. Picking the right information to he learned and ignoring the rest is
probably one of the key functions of attentional selection. Indeed, the
resultant bottleneck appears to be necessary for the utilization of some kinds
of memory (Naveh-Benjamin and Guez, 2000). The test bed we use for exploring
attentional learning is the control of a segmented arm moving around in a
boxlike environment. It can pick up, move, and drop disks. At the most abstract
level, the arm is used to solve various kinds of puzzles. The problem we
explored was one of ordering various objects into target locations. This is
equivalent to the
Tower
of Hanoi problem (Claus,
1884). In our version of this problem (see Fig. 4.2), we begin with an
allotment of disks of various diameters. We assume that they have holes in
their middle, that these disks are stacked in order of decreasing size (i.e., a
larger disk must always be below a smaller one), and that the segmented arm can
transfer the disks from one target stack to another one. The arm moves around
the board and physically takes the top disk from each target and moves it to
another stack, with the end goal of placing them in increasing size on a
speciï¬c goal target. Various obstacles are placed on the board through which
the arm cannot pass. The arm’s segments can overlap as it moves. We assume that
the end effector, when placed over a target, takes or releases a single disk
automatically. Our problem, then, is to manipulate the joints of the arm to
move its endeffector between the appropriate targets in the correct order so as
to solve the puzzle. The details of the articulated arm, the playing board, and
targets are shown in Fig. 4.2. For our purposes, we give the arm segments
minimal dynamics involving a maximum torque and a momentum/friction decay
characteristic. These force relationships are solved by the logic subsystem
using a set of torquechange equations similar to those described by (Uno et
al., 1989) for modeling human limb control. Initially, the system has not yet
learned to drive its joints, and so must use its logic/planning functions to
solve the control problem via explicit equations. The threesegment arm has a
complicated inverse kinematics
SECTION
I. FOUNDATIONS
20
2
CHAPTER 4. SELECTIVE ATTENTION AS AN OPTIMAL COMPUTATIONAL STRATEGY
1.5
2
1
F2
3 4
F3
l 4
q3
l
q2
0.5
F1
l
1
0
3
-0.5
-1
1
-1.5
-2 -2
-1
0
1
2
Baskets
Obstacles
Reachable Area
FIGURE 4.2 The Tower of Hanoi problem arranged for solution
by an articulated arm. The arm must move between the marked targets without
colliding with the solid obstacles. Outlined squares are the positions of the
targets. Solid circles indicate obstacles around which the arm must navigate.
which requires an expensive optimization process to ï¬nd
the best trajectory to move from a present position to a target position. The
minimum torque-change model selects a conï¬guration out of the possible
solutions that requires theminimum angular change of the arm segments to
achieve. This is a very costly step in terms of required computational power,
and so at ï¬rst the attention of the logic unit is taken up by this low-level
fun
Política de privacidad
