Repository of works by Caltech published authors.
Type = Book Section
Mostrando recursos 1 - 20 de 5,351
Hidden Markov Models in Molecular Biology: New Algorithms and Applications - Baldi, Pierre; Chauvin, Yves; Hunkapiller, Tim; McClure, Marcella A.
Hidden Markov Models (HMMs) can be applied to several important problems in molecular biology. We introduce a new convergent learning algorithm for HMMs that, unlike the classical Baum-Welch algorithm is smooth and can be applied on-line or in batch mode, with or without the usual Viterbi most likely path approximation. Left-right HMMs with insertion and deletion states are then trained to represent several protein families including immunoglobulins and
kinases. In all cases, the models derived capture all the important statistical properties of the families and can be used efficiently in a number of important tasks such as multiple alignment, motif detection,...
Visual Motion Computation in Analog VLSI Using Pulses - Sarpeshkar, Rahul; Bair, Wyeth; Koch, Christof
The real time computation of motion from real images
using a single chip with integrated sensors is a hard problem. We present two analog VLSI schemes that use pulse
domain neuromorphic circuits to compute motion. Pulses
of variable width, rather than graded potentials, represent
a natural medium for evaluating temporal relationships.
Both algorithms measure speed by timing a moving edge
in the image. Our first model is inspired by Reichardt's
algorithm in the fiy and yields a non-monotonic response
vs. velocity curve. We present data from a chip that
implements this model. Our second algorithm yields a
monotonic response vs. velocity curve and is currently
being translated into silicon.
A Method for Learning from Hints - Abu-Mostafa, Yaser S.
We address the problem of learning an unknown function by
putting together several pieces of information (hints) that we know about the function. We introduce a method that generalizes learning from examples to learning from hints. A canonical representation of hints is defined and illustrated for new types of hints. All the hints are represented to the learning process by examples, and
examples of the function are treated on equal footing with the rest of the hints. During learning, examples from different hints are selected for processing according to a given schedule. We present two types of schedules; fixed schedules that specify...
Probability Estimation from a Database Using a Gibbs Energy Model - Miller, John W.; Goodman, Rodney M.
We present an algorithm for creating a neural network which produces accurate probability estimates as outputs. The network implements a Gibbs probability distribution model of the training database. This model is created by a new transformation relating the joint probabilities of attributes in the database to the weights (Gibbs potentials) of the distributed network model. The theory
of this transformation is presented together with experimental results. One advantage of this approach is the network weights are prescribed without iterative gradient descent. Used as a classifier the network tied or outperformed published results on a variety of
Combined Neural Network and Rule-Based Framework for Probabilistic Pattern Recognition and Discovery - Greenspan, Hayit K.; Goodman, Rodney; Chellappa, Rama
A combined neural network and rule-based approach is suggested as a general framework for pattern recognition. This approach enables unsupervised and supervised learning, respectively, while providing probability estimates for the output classes. The probability maps are utilized for
higher level analysis such as a feedback for smoothing over the output label maps and the identification of unknown patterns (pattern "discovery"). The suggested approach is presented and demonstrated in the texture - analysis task. A correct classification rate in the 90 percentile is achieved for both unstructured and structured natural texture mosaics. The advantages of the probabilistic approach to pattern analysis are...
A Theory of Shape-Memory Thin Films with Applications - Bhattacharya, K.; James, R. D.
Shape-memory alloys have the largest energy output per unit volume per cycle of known actuator systems . Unfortunately, they are temperature activated and hence, their frequency is limited in bulk specimens. However, this is overcome in thin films; and hence shape-memory alloys are ideal actuator materials in micromachines. The heart of the shape-memory effect lies in a martensitic phase transformation and the resulting microstructure. It is well-known that microstructure can be significantly different in thin films as compared to bulk materials. In this paper, we report on a theory of single crystal martensitic this films. We show that single crystal...
Observation of Domain Motion in Single-Crystal Barium Titanate under Combined Electromechanical Loading Conditions - Burcsu, E.; Ravichandran, G.; Bhattacharya, K.
The nonlinear electromechanical behavior of ferroelectric materials is governed by the motion of domains. Since many common ferroelectric materials, such as barium titanate and PZT, are also ferroelastic, the domain motion is highly affected by stress as well as electric field. Experiments are performed on (001) and (100) oriented single crystals of barium titanate under combined electromechanical loading conditions. The crystal is subjected to a constant compressive stress (dead load) and an oscillating electric field along the  direction. Global deformation and polarization are measured as a function of electric field at different values of compressive stress. The use of...
Thin Films of Active Materials - Bhattacharya, K.
This paper summarizes some recent developments in the study of the mechanics of thin films motivated by the use of active materials in making microactuators.
Kinematics of Crossing Twins - Bhattacharya, K.
Twins are commonly observed in crystalline solids that undergo martensitic phase transformation. In many materials the twins are also seen to cross each other. This is surprising in view of the severe kinematic restrictions that such crossings impose. This paper presents a sufficient condition for satisfying these restrictions. It
turns out that this condition is automatically satisfied as a consequence of material symmetry in many common martensitic materials. This explains the common observation
of crossing twins. The result is also applied to the magnetostrictive material Terfenol.
Deformation Behavior of a Shape Memory Alloy, Nitinol - Daly, Samantha; Bhattacharya, Kaushik; Ravichandran, Guruswami
Nickel-Titanium, commonly referred to as Nitinol, is a shape-memory alloy with numerous
applications due to its superelastic nature and its ability to revert to a previously defined shape
when deformed and then heated past a set transformation temperature. While the
crystallography and the overall phenomenology are reasonably well understood, much remains
unknown about the deformation and failure mechanisms of these materials. These latter issues
are becoming critically important as Nitinol is being increasingly used in medical devices and
space applications. The talk will describe the investigation of the deformation and failure of
Nitinol using an in-situ optical technique called Digital Image Correlation (DIC). With this
technique, full-field quantitative...
Modeling large strain electrostriction of ferroelectrics under combined electromechanical loads - Zhang, Wei; Bhattacharya, Kaushik
A computational model for investigating domain switching and macroscopic electromechanical properties of ferroelectric materials is developed. Various aspects of domain nucleation and growth, and their effects on macroscopic hysteresis are examined. The model is validated against recent experimental observations. It is thus validated as a design tool to investigate various aspects of a novel thin film ferroelectric microactuator in future work.
Interaction of oxygen vacancies with domain walls and its impact on fatigue in ferroelectric thin films - Xiao, Yu; Bhattacharya, Kaushik
The role of oxygen vacancies in fatigue and dielectric breakdown has been a topic of intense research in ferroelectric perovskites like BaTiO_3. This paper presents a comprehensive model that treats the ferroelectrics as polarizable wide band-gap semiconductors where the oxygen vacancies act as donors. First, a fully coupled nonlinear model is developed with space charges, polarization, electric potential and elastic displacements as variables without making any a priori assumptions on the space charge distribution and the polarization. Second, a Pt/BaTiO_3/Pt structure is considered. Full-field coupled numerical simulations are used to investigate the structure of 180° and 90° domain walls in...
Modeling electromechanical properties of ionic polymers - Xiao, Yu; Bhattacharya, Kaushik
We present a multi-scale approach to modeling the electro-mechanical behavior of ionic polymers. We start with
a detailed elasto-electro-chemical model which allows for finite deformation. We reduce it to one space dimension
appropriate for the commonly used sheet configuration, and demonstrate that steady state solutions display an
important boundary layer effect. We conclude with a macroscopic model of a strip of ionic-polymer-metal-composite.
Electromechanical behavior of 90-degree domain motion in barium titanate single crystals - Burcsu, Eric; Ravichandran, G.; Bhattacharya, Kaushik
It is well known that many common ferroelectric materials are also ferroelastic, thus the nonlinear behavior of these materials, as governed by domain motion, is highly affected by stress, as well as electric field. The combined influence of stress and electric field on domain motion and the electrostrictive response of ferroelectric single crystals is investigated. Experiments are performed on (001) and (100) oriented single crystals of barium titanate under combined electro-mechanical loading. The crystal is exposed to a constant compressive stress and an oscillating electric field along the  direction. Global deformation and polarization are measured as a function of...
The Taylor Estimate of Recoverable Strains in Shape-Memory Polycrystals - Bhattacharya, K.; Kohn, R. V.; Shu, Y. C.
Shape-memory behavior ls the ability of ccrwin materials to recover, on heating, apparently plastic deformation sustained below a critical temperature. Some materials have good shape-memory behavior as single crystals but little or none as polycrystals, while others have good shape-memory behavior even as polycrystals. Bhattacharya and Kohn (1996. 1997) have proposed a framework to understand this difference. They use energy minimization and the Taylor estimate to argue that the recoverable strains in a polycrystal depend not only on the texture of the polycrystal and the transformation, but critically on the change in symmetry during the underlying martensitic phase transformation. Their...
Energy minimization and nonlinear problems in polycrystalline solids - Bhattacharya, Kaushik
Common engineering structural materials -- metallic alloys and ceramics -- are polycrystalline. They are made up of a very large number of grins which have identical crystal structure, but which are oriented differently. The properties of the material depend critically on the texture, by which one means the size, the shape and the orientation distributions of the different grains. If we can systematically understand this dependence, we can identify textures which provide the best possible properties and then try to design a processing technique which gives rise to the texture.
Linear properties -- elastic moduli, conductivity etc. -- have received...
Characteristics of Lead Isotope Evolution on a Continental Scale in the Earth - Patterson, Clair
Recent investigations of lead isotopes in oceanic sediments, beach and river sands, and
oceanic volcanics are reviewed from one comprehensive point of view: relatively large
amounts of uranium and small amounts of lead have been continuously transported from
the interior of the earth to an outer protocontinental layer throughout geologic time;
continental segments were formed at different times from this layer, but such activity
was largely confined to the interval 3500 m.yrs ago to 2500 m.yrs ago; uranium is
enriched in the surface rocks of the continents by magmatic activity, but such rocks
are well-mixed with uranium depleted material throughout a depth of a few tens of
Comparison of Oxygen Isotope Analyses of Tektites, Soils, and Impactite Glasses - Taylor, Hugh P., Jr.; Epstein, Samuel
New oxygen isotope analyses of tektites from various localities were made, including
two philippinites from Isabela, Philippine Islands, one of which contains small Ni-Fe
spherules (CHAO and others ); a billitonite; a Muong Nong-type tektite, as well as
a normal type, from Phang Daeng, Thailand; a Georgia tektite; and seven bediasites.
The total range in δ-values for 24 tektites from widely separated occurrences is 8.9 to
10.8 per mil, relative to mean ocean water. Eight isotopic analyses of bediasites exhibit
a good correlation with index of refraction, density, and SiO_2 content. The higher the
O(^18)/O^(16) ratio the lower is the SiO_2 content. The other analyzed tektites do...
Oxygen Isotope Studies on the Origin of Dolomites - Epstein, Samuel; Graf, Donald L.; Degens, Egon T.
Oxygen isotope measurements were made on coexisting dolomite and calcite pairs both
from recent sedimentary deposits and from high temperature laboratory syntheses.
In the limited number of synthetic carbonate samples thus far available, it appears that
there is an oxygen isotope fractionation between dolomite and calcite, similar to that
found in naturally occurring hydrothermal calcite-dolomite pairs. There is a lack of
such fractionation between dolomite and calcite in young sedimentary deposits. It is
therefore concluded that dolomite is a secondary product resulting from addition of
magnesium to existing crystalline CaCO_3, which takes place without an isotopic change
in the oxygen of the precursor CaCO_3. It is also shown...