Efficient distribution-free population learning of simple concepts

We consider a variant of the ‘population learning model’ proposed by Kearns and Seung [8], in which the learner is required to be ‘distribution-free’ as well as computationally efficient. A population learner receives as input hypotheses from a large population of agents and produces as output its final hypothesis. Each agent is assumed to independently obtain labeled sample for the target concept and output a hypothesis. A polynomial time population learner is said to PAC-learn a concept class, if its hypothesis is probably approximately correct whenever the population size exceeds a certain bound which is polynomial, even if the sample size for each agent is fixed at some constant. We exhibit some general population learning strategies, and some simple concept classes that can be learned by them. These strategies include the ‘supremum hypothesis finder’, the ‘minimum superset finder’ (a special case of the ‘supremum hypothesis finder’), and various voting schemes. When coupled with appropriate agent algorithms, these strategies can learn a variety of simple concept classes, such as the ‘high–low game’, conjunctions, axis-parallel rectangles and others. We give upper bounds on the required population size for each of these cases, and show that these systems can be used to obtain a speed up from the ordinary PAC-learning model [11], with appropriate choices of sample and population sizes. With the population learner restricted to be a voting scheme, what we have is effectively a model of ‘population prediction’, in which the learner is to predict the value of the target concept at an arbitrarily drawn point, as a threshold function of the predictions made by its agents on the same point. We show that the population learning model is strictly more powerful than the population prediction model. Finally, we consider a variant of this model with classification noise, and exhibit a population learner for the class of conjunctions in this model. This revised version was published online in June 2006 with corrections to the Cover Date.     

A new hybrid GA solution to combinatorial optimization problems— an application to the multiprocessor scheduling problem

The multiprocessor scheduling problem is one of the classic examples of NP-hard combinatorial optimization problems. Several polynomial time optimization algorithms have been proposed for approximating the multiprocessor scheduling problem. In this paper, we suggest a geneticizedknowledge genetic algorithm (gkGA) as an efficient heuristic approach for solving the multiprocessor scheduling and other combinatorial optimization problems. The basic idea behind the gkGA approach is that knowledge of the heuristics to be used in the GA is also geneticized alongiside the genetic chromosomes. We start by providing four conversion schemes based on heuristics for converting chromosomes into priority lists. Through experimental evaluation, we observe that the performance of our GA based on each of these schemes is instance-dependent. However, if we simultaneously incorporate these schemes into our GA through the gkGA approach, simulation results show that the approach is not problem-dependent, and that the approach outperforms that of the previous GA. We also show the effectiveness of the gkGA approach compared with other conventional schemes through experimental evaluation. This work was presented, in part, at the Second International Symposium on Artifiical Life and Robotics, Oita, Japan, February 18–20, 1997     

Linear bipolar OTAs employing exponential‐law circuits

This paper presents design of linear bipolar OTAs, which are composed of two function blocks; one is an exponential‐law circuit and the other is a core cell. Multi‐tanh cells are employed as the core cell. This kind of OTA has lower power dissipation relatively to the conventional multi‐tanh cell. According as the order of the multi‐tanh core cell becomes higher, the number of circuit realization for the core cell increases. For example, we have two OTAs for the core circuit of an emitter‐coupled pair and four OTAs for the doublet core cell. Thus, we consider the generalized OTAs for an arbitrary order n of the core cell and obtain a formula to give the realization number of the linear OTAs for n. According to the formula, there must be eight OTAs in the case of n=3. All of the eight OTAs are examined. Analysis and simulation results show that the OTAs have advantage in their characteristics, such as linear input range, power dissipation, noise, and frequency response. Copyright 2004 John Wiley & Sons, Ltd.     

Phase Diagram of Hydrogen in Palladium

Hydrogen in palladium, Pd-H(D), is an interesting system because of the highly mobile hydrogen and the presence of a phase boundary below 100 K. Experimentally, however, the nature of this transition has not been established. Historically this transition around 55 to 100 K has been thought to be an order-disorder transition. Such a transition would produce a phase boundary with anomalies at specific hydrogen concentrations corresponding to the specific ordered structures. In order to check this phase boundary we have performed a detailed study of the hydrogen concentration dependence of the specific heat of PdH _x over the temperature range from below 0.5 K to above 100 K using PdH _x specimens with x up to 0.8753. The measured heat capacity has been analyzed as the sum of contributions due to the lattice specific heat of Pd, the electronic specific heat of PdH _x , and the excess contribution caused by hydrogenation of the specimen. The excess specific heat result shows a sharp peak which indicates a phase boundary with transition temperature T ₁=55 K to 85 K depending linearly on the hydrogen concentration from x=0.6572 to 0.8753. We do not observe anomalies at specific x values as would be expected for the specific ordered structures.     

Design and performance of compensator for decremental persistent current in HTS magnets using linear type magnetic flux pump

This paper has described the characteristics of a stationary linear type magnetic flux pump newly developed this time. The linear type flux pump aims to compensate a little decremental persistent current of the HTS magnet in NMR and MRI systems. The flux pump mainly consists of DC bias coil, 3-phase AC coil and Nb foil. Analytical results by the FEM are proved to nearly agree with experimental ones. In the experiment, it has been investigated that the flux pump can effectively charge the current in the load coil of 1.3 mH for various frequencies in 150 s under two following conditions: (1) DC bias of 10 A and AC of 4 A_rms, (2) DC bias of 10 A and AC of 5 A_rms. The maximum magnitudes of initial increasing rate of pumping current and load magnet voltage are: (1) 0.2 A/s and 25 mV and (2) 0.22 A/s and 34 mV, respectively.     

Magnetic Behavior of Some Rare-Earth Transition-Metal Perovskite Oxide Systems

Magnetic properties were investigated for the rare-earth 3d-transition metal oxides with the perovskite structure. Intriguing magnetic phenomena were reviewed for a few systems:magnetization peak effect in the titanates, magnetization reversal in the chromites and metallic ferromagnetism in the cobahites. The results suggest an important role of the rare-earth ions for the magnetic properties of such complex oxides.     

Bending Behavior of Steel Pipe Girders Filled with Ultralight Mortar

Bending behavior of steel pipes filled with ultralight mortar was studied by bending tests using a steel pipe, steel pipes filled with ultralight mortar, and steel pipes filled with light aggregate concrete and normal concrete. The steel pipe model filled with normal concrete had 1.8 times higher bending strength than the steel pipe model. The bending behavior of the steel pipe filled model with ultralight mortar was not improved when the compressive strength of the ultralight mortar was less than 1 MPa. However, ductility was much improved when the compressive strength was over 5 MPa, and the ultimate steel strain was more than double of the steel pipe model. The strains of steel and concrete in all the models were proportional to the distance from the neutral axis until the steel plate yielded. A simple analytical method was proposed to calculate the bending moments of the ultralight mortar filled steel pipes. The calculated values agreed very well with the test results.     

Glasslike Transitions in Thin Polymer-Melt Films Due to Thickness Constraint

The shear properties of thin films of star and linear polyisoprene (PIP) melts under high pressure were investigated as a function of sliding velocity (shear rate) using the surface forces apparatus. The results were contrasted with their bulk rheological properties; effects of thickness constraint on the shear behavior were discussed. The melts of PIP in bulk exhibit Newtonian-like constant viscosity at least at low shear rates (frequencies), which suggests that individual molecules flow with lateral sliding motion. However, thin films of PIP melts show tribological features involving apparent shear-thinning behavior, indicative of the correlated motions in confined geometries. The shear-property change from bulk rheological behavior to thin-film tribological behavior along with the thickness decrease reflects the physical states and their transitions in the systems; the thickness constraint induces glasslike transitions. Effects of molecular weights and molecular architecture (star-branched or linear) on the shear properties are also discussed.     

Permeation measurements for investigating atomic hydrogen flux and wall pumping/fuelling dynamics in QUEST

In order to investigate the overall atomic hydrogen background and the dynamic characteristics of wall pumping/fuelling phenomenon, a permeation probe system has been developed and applied in the spherical tokamak QUEST. Reliability of measurements, within ±3% accuracy and a positive correlation with the hydrogen line emission over three orders of magnitude have been demonstrated for more than 3000 various plasma discharges. By comparison of the experimental permeation (flux) curves with the numerically simulated curves, the net incident atomic hydrogen flux is evaluated in the range of 1 × 10¹⁹ H m^?2 s^?1 to 4 × 10²⁰ H m^?2 s^?1. The atomic flux has been investigated as a function of various plasma operation parameters like RF power, gas pressure and magnetic configuration. Using the static particle balance and permeation measurements, the progress in wall conditioning has been investigated. An inverse correlation between the atomic hydrogen flux and improvement in wall pumping has been observed over the two campaigns.     

PBPAF／PBPA共混物相容性的研究

研究了双酚ＡＦ基聚合物和双酚Ａ基聚合物的共混相容性。通过用ＤＳＣ，ＴＭＡ测定单一聚合物和共混物的玻璃化温度来判别它们的相容性．结果表明，两种聚合物在部分组成比时有一定的相容性。