-
Data Equity: Foundational Concepts for Generative AI
Authors:
JoAnn Stonier,
Lauren Woodman,
Majed Alshammari,
Renée Cummings,
Nighat Dad,
Arti Garg,
Alberto Giovanni Busetto,
Katherine Hsiao,
Maui Hudson,
Parminder Jeet Singh,
David Kanamugire,
Astha Kapoor,
Zheng Lei,
Jacqueline Lu,
Emna Mizouni,
Angela Oduor Lungati,
María Paz Canales Loebel,
Arathi Sethumadhavan,
Sarah Telford,
Supheakmungkol Sarin,
Kimmy Bettinger,
Stephanie Teeuwen
Abstract:
This briefing paper focuses on data equity within foundation models, both in terms of the impact of Generative AI (genAI) on society and on the further development of genAI tools. GenAI promises immense potential to drive digital and social innovation, such as improving efficiency, enhancing creativity and augmenting existing data. GenAI has the potential to democratize access and usage of technol…
▽ More
This briefing paper focuses on data equity within foundation models, both in terms of the impact of Generative AI (genAI) on society and on the further development of genAI tools. GenAI promises immense potential to drive digital and social innovation, such as improving efficiency, enhancing creativity and augmenting existing data. GenAI has the potential to democratize access and usage of technologies. However, left unchecked, it could deepen inequities. With the advent of genAI significantly increasing the rate at which AI is deployed and developed, exploring frameworks for data equity is more urgent than ever. The goals of the briefing paper are threefold: to establish a shared vocabulary to facilitate collaboration and dialogue; to scope initial concerns to establish a framework for inquiry on which stakeholders can focus; and to shape future development of promising technologies. The paper represents a first step in exploring and promoting data equity in the context of genAI. The proposed definitions, framework and recommendations are intended to proactively shape the development of promising genAI technologies.
△ Less
Submitted 27 October, 2023;
originally announced November 2023.
-
Random projection tree similarity metric for SpectralNet
Authors:
Mashaan Alshammari,
John Stavrakakis,
Adel F. Ahmed,
Masahiro Takatsuka
Abstract:
SpectralNet is a graph clustering method that uses neural network to find an embedding that separates the data. So far it was only used with $k$-nn graphs, which are usually constructed using a distance metric (e.g., Euclidean distance). $k$-nn graphs restrict the points to have a fixed number of neighbors regardless of the local statistics around them. We proposed a new SpectralNet similarity met…
▽ More
SpectralNet is a graph clustering method that uses neural network to find an embedding that separates the data. So far it was only used with $k$-nn graphs, which are usually constructed using a distance metric (e.g., Euclidean distance). $k$-nn graphs restrict the points to have a fixed number of neighbors regardless of the local statistics around them. We proposed a new SpectralNet similarity metric based on random projection trees (rpTrees). Our experiments revealed that SpectralNet produces better clustering accuracy using rpTree similarity metric compared to $k$-nn graph with a distance metric. Also, we found out that rpTree parameters do not affect the clustering accuracy. These parameters include the leaf size and the selection of projection direction. It is computationally efficient to keep the leaf size in order of $\log(n)$, and project the points onto a random direction instead of trying to find the direction with the maximum dispersion.
△ Less
Submitted 25 February, 2023;
originally announced February 2023.
-
A parameter-free graph reduction for spectral clustering and SpectralNet
Authors:
Mashaan Alshammari,
John Stavrakakis,
Masahiro Takatsuka
Abstract:
Graph-based clustering methods like spectral clustering and SpectralNet are very efficient in detecting clusters of non-convex shapes. Unlike the popular $k$-means, graph-based clustering methods do not assume that each cluster has a single mean. However, these methods need a graph where vertices in the same cluster are connected by edges of large weights. To achieve this goal, many studies have p…
▽ More
Graph-based clustering methods like spectral clustering and SpectralNet are very efficient in detecting clusters of non-convex shapes. Unlike the popular $k$-means, graph-based clustering methods do not assume that each cluster has a single mean. However, these methods need a graph where vertices in the same cluster are connected by edges of large weights. To achieve this goal, many studies have proposed graph reduction methods with parameters. Unfortunately, these parameters have to be tuned for every dataset. We introduce a graph reduction method that does not require any parameters. First, the distances from every point $p$ to its neighbors are filtered using an adaptive threshold to only keep neighbors with similar surrounding density. Second, the similarities with close neighbors are computed and only high similarities are kept. The edges that survive these two filtering steps form the constructed graph that was passed to spectral clustering and SpectralNet. The experiments showed that our method provides a stable alternative, where other methods performance fluctuated according to the setting of their parameters.
△ Less
Submitted 25 February, 2023;
originally announced February 2023.
-
The Effect of Points Dispersion on the $k$-nn Search in Random Projection Forests
Authors:
Mashaan Alshammari,
John Stavrakakis,
Adel F. Ahmed,
Masahiro Takatsuka
Abstract:
Partitioning trees are efficient data structures for $k$-nearest neighbor search. Machine learning libraries commonly use a special type of partitioning trees called $k$d-trees to perform $k$-nn search. Unfortunately, $k$d-trees can be ineffective in high dimensions because they need more tree levels to decrease the vector quantization (VQ) error. Random projection trees rpTrees solve this scalabi…
▽ More
Partitioning trees are efficient data structures for $k$-nearest neighbor search. Machine learning libraries commonly use a special type of partitioning trees called $k$d-trees to perform $k$-nn search. Unfortunately, $k$d-trees can be ineffective in high dimensions because they need more tree levels to decrease the vector quantization (VQ) error. Random projection trees rpTrees solve this scalability problem by using random directions to split the data. A collection of rpTrees is called rpForest. $k$-nn search in an rpForest is influenced by two factors: 1) the dispersion of points along the random direction and 2) the number of rpTrees in the rpForest. In this study, we investigate how these two factors affect the $k$-nn search with varying $k$ values and different datasets. We found that with larger number of trees, the dispersion of points has a very limited effect on the $k$-nn search. One should use the original rpTree algorithm by picking a random direction regardless of the dispersion of points.
△ Less
Submitted 25 February, 2023;
originally announced February 2023.
-
Random Projection Forest Initialization for Graph Convolutional Networks
Authors:
Mashaan Alshammari,
John Stavrakakis,
Adel F. Ahmed,
Masahiro Takatsuka
Abstract:
Graph convolutional networks (GCNs) were a great step towards extending deep learning to unstructured data such as graphs. But GCNs still need a constructed graph to work with. To solve this problem, classical graphs such as $k$-nearest neighbor are usually used to initialize the GCN. Although it is computationally efficient to construct $k$-nn graphs, the constructed graph might not be very usefu…
▽ More
Graph convolutional networks (GCNs) were a great step towards extending deep learning to unstructured data such as graphs. But GCNs still need a constructed graph to work with. To solve this problem, classical graphs such as $k$-nearest neighbor are usually used to initialize the GCN. Although it is computationally efficient to construct $k$-nn graphs, the constructed graph might not be very useful for learning. In a $k$-nn graph, points are restricted to have a fixed number of edges, and all edges in the graph have equal weights. We present a new way to construct the graph and initialize the GCN. It is based on random projection forest (rpForest). rpForest enables us to assign varying weights on edges indicating varying importance, which enhanced the learning. The number of trees is a hyperparameter in rpForest. We performed spectral analysis to help us setting this parameter in the right range. In the experiments, initializing the GCN using rpForest provides better results compared to $k$-nn initialization.
△ Less
Submitted 22 October, 2023; v1 submitted 22 February, 2023;
originally announced February 2023.
-
Graph Construction using Principal Axis Trees for Simple Graph Convolution
Authors:
Mashaan Alshammari,
John Stavrakakis,
Adel F. Ahmed,
Masahiro Takatsuka
Abstract:
Graph Neural Networks (GNNs) are increasingly becoming the favorite method for graph learning. They exploit the semi-supervised nature of deep learning, and they bypass computational bottlenecks associated with traditional graph learning methods. In addition to the feature matrix $X$, GNNs need an adjacency matrix $A$ to perform feature propagation. In many cases, the adjacency matrix $A$ is missi…
▽ More
Graph Neural Networks (GNNs) are increasingly becoming the favorite method for graph learning. They exploit the semi-supervised nature of deep learning, and they bypass computational bottlenecks associated with traditional graph learning methods. In addition to the feature matrix $X$, GNNs need an adjacency matrix $A$ to perform feature propagation. In many cases, the adjacency matrix $A$ is missing. We introduce a graph construction scheme that constructs the adjacency matrix $A$ using unsupervised and supervised information. Unsupervised information characterizes the neighborhood around points. We used Principal Axis trees (PA-trees) as a source for unsupervised information, where we create edges between points falling onto the same leaf node. For supervised information, we used the concept of penalty and intrinsic graphs. A penalty graph connects points with different class labels, whereas an intrinsic graph connects points with the same class labels. We used the penalty and intrinsic graphs to remove or add edges to the graph constructed via PA-tree. We tested this graph construction scheme on two well-known GNNs: 1) Graph Convolutional Network (GCN) and 2) Simple Graph Convolution (SGC). The experiments show that it is better to use SGC because it is faster and delivers better or the same results as GCN. We also test the effect of oversmoothing on both GCN and SGC. We found out that the level of smoothing has to be carefully selected for SGC to avoid oversmoothing.
△ Less
Submitted 7 November, 2023; v1 submitted 22 February, 2023;
originally announced February 2023.
-
Approximate spectral clustering density-based similarity for noisy datasets
Authors:
Mashaan Alshammari,
Masahiro Takatsuka
Abstract:
Approximate spectral clustering (ASC) was developed to overcome heavy computational demands of spectral clustering (SC). It maintains SC ability in predicting non-convex clusters. Since it involves a preprocessing step, ASC defines new similarity measures to assign weights on graph edges. Connectivity matrix (CONN) is an efficient similarity measure to construct graphs for ASC. It defines the weig…
▽ More
Approximate spectral clustering (ASC) was developed to overcome heavy computational demands of spectral clustering (SC). It maintains SC ability in predicting non-convex clusters. Since it involves a preprocessing step, ASC defines new similarity measures to assign weights on graph edges. Connectivity matrix (CONN) is an efficient similarity measure to construct graphs for ASC. It defines the weight between two vertices as the number of points assigned to them during vector quantization training. However, this relationship is undirected, where it is not clear which of the vertices is contributing more to that edge. Also, CONN could be tricked by noisy density between clusters. We defined a directed version of CONN, named DCONN, to get insights on vertices contributions to edges. Also, we provided filtering schemes to ensure CONN edges are highlighting potential clusters. Experiments reveal that the proposed filtering was highly efficient when noise cannot be tolerated by CONN.
△ Less
Submitted 22 February, 2023;
originally announced February 2023.
-
Approximate spectral clustering with eigenvector selection and self-tuned $k$
Authors:
Mashaan Alshammari,
Masahiro Takatsuka
Abstract:
The recently emerged spectral clustering surpasses conventional clustering methods by detecting clusters of any shape without the convexity assumption. Unfortunately, with a computational complexity of $O(n^3)$, it was infeasible for multiple real applications, where $n$ could be large. This stimulates researchers to propose the approximate spectral clustering (ASC). However, most of ASC methods a…
▽ More
The recently emerged spectral clustering surpasses conventional clustering methods by detecting clusters of any shape without the convexity assumption. Unfortunately, with a computational complexity of $O(n^3)$, it was infeasible for multiple real applications, where $n$ could be large. This stimulates researchers to propose the approximate spectral clustering (ASC). However, most of ASC methods assumed that the number of clusters $k$ was known. In practice, manual setting of $k$ could be subjective or time consuming. The proposed algorithm has two relevance metrics for estimating $k$ in two vital steps of ASC. One for selecting the eigenvectors spanning the embedding space, and the other to discover the number of clusters in that space. The algorithm used a growing neural gas (GNG) approximation, GNG is superior in preserving input data topology. The experimental setup demonstrates the efficiency of the proposed algorithm and its ability to compete with similar methods where $k$ was set manually.
△ Less
Submitted 22 February, 2023;
originally announced February 2023.
-
Refining a $k$-nearest neighbor graph for a computationally efficient spectral clustering
Authors:
Mashaan Alshammari,
John Stavrakakis,
Masahiro Takatsuka
Abstract:
Spectral clustering became a popular choice for data clustering for its ability of uncovering clusters of different shapes. However, it is not always preferable over other clustering methods due to its computational demands. One of the effective ways to bypass these computational demands is to perform spectral clustering on a subset of points (data representatives) then generalize the clustering o…
▽ More
Spectral clustering became a popular choice for data clustering for its ability of uncovering clusters of different shapes. However, it is not always preferable over other clustering methods due to its computational demands. One of the effective ways to bypass these computational demands is to perform spectral clustering on a subset of points (data representatives) then generalize the clustering outcome, this is known as approximate spectral clustering (ASC). ASC uses sampling or quantization to select data representatives. This makes it vulnerable to 1) performance inconsistency (since these methods have a random step either in initialization or training), 2) local statistics loss (because the pairwise similarities are extracted from data representatives instead of data points). We proposed a refined version of $k$-nearest neighbor graph, in which we keep data points and aggressively reduce number of edges for computational efficiency. Local statistics were exploited to keep the edges that do not violate the intra-cluster distances and nullify all other edges in the $k$-nearest neighbor graph. We also introduced an optional step to automatically select the number of clusters $C$. The proposed method was tested on synthetic and real datasets. Compared to ASC methods, the proposed method delivered a consistent performance despite significant reduction of edges.
△ Less
Submitted 22 February, 2023;
originally announced February 2023.
-
Enhanced magnetic properties in ZnCoAlO caused by exchangecoupling to Co nanoparticles
Authors:
Qi Feng,
Wala Dizayee,
Xiaoli Li,
David S Score,
James R Neal,
Anthony J Behan,
Abbas Mokhtari,
Marzook S Alshammari,
Mohammed S Al-Qahtani,
Harry J Blythe,
Roy W Chantrell,
Steve M Heald,
Xiao-Hong Xu,
A Mark Fox,
Gillian A Gehring
Abstract:
Wereport the results of a sequence of magnetisation and magneto-optical studies on laser ablated thin films of ZnCoAlO and ZnCoO that contain a small amount of metallic cobalt. The results are compared to those expected when all the magnetization is due to isolated metallic clusters of cobalt and with an oxide sample that is almost free from metallic inclusions. Using a variety of direct magnetic…
▽ More
Wereport the results of a sequence of magnetisation and magneto-optical studies on laser ablated thin films of ZnCoAlO and ZnCoO that contain a small amount of metallic cobalt. The results are compared to those expected when all the magnetization is due to isolated metallic clusters of cobalt and with an oxide sample that is almost free from metallic inclusions. Using a variety of direct magnetic measurements and also magnetic circular dichroism we find that there is ferromagnetism within both the oxide and the metallic inclusions, and furthermore that these magnetic components are exchange-coupled when aluminium is included. This enhances both the coercive field and the remanence. Hence the presence of a controlled quantity of metallic nanoparticles in ZnAlO can improve the magnetic response of the oxide, thus giving great advantages for applications in spintronics.
△ Less
Submitted 7 August, 2018;
originally announced August 2018.
-
Negative activation energy and dielectric signatures of excitons and excitonic Mott transitions in quantum confined laser structures
Authors:
Amit Bhunia,
Kanika Bansal,
Mohamed Henini,
Marzook S. Alshammari,
Shouvik Datta
Abstract:
Mostly, optical spectroscopies are used to investigate the physics of excitons, whereas their electrical evidences are hardly explored. Here, we examined a forward bias activated differential capacitance response of GaInP-AlGaInP based multi-quantum well laser diodes to trace the presence of excitons using electrical measurements. Occurrence of negative activation energy after light emission is un…
▽ More
Mostly, optical spectroscopies are used to investigate the physics of excitons, whereas their electrical evidences are hardly explored. Here, we examined a forward bias activated differential capacitance response of GaInP-AlGaInP based multi-quantum well laser diodes to trace the presence of excitons using electrical measurements. Occurrence of negative activation energy after light emission is understood as thermodynamical signature of steady state excitonic population under intermediate range of carrier injections. Similar corroborative results are also observed in an InGaAs-GaAs quantum dot laser structure grown by molecular beam epitaxy. With increasing biases, the measured differential capacitance response slowly vanishes. This represents gradual Mott transition of an excitonic phase into an electron-hole plasma in a GaInP-AlGaInP laser diode. This is further substantiated by more and more exponentially looking shapes of high energy tails in electroluminescence spectra with increasing forward bias, which originates from a growing non-degenerate population of free electrons and holes. Such an experimental correlation between electrical and optical properties of excitons can be used to advance the next generation excitonic devices.
△ Less
Submitted 20 October, 2016;
originally announced October 2016.
-
Dynamical patterns of p53 states driven by PML
Authors:
Md. Jahoor Alam,
Eyad M. AlShammari,
R. K. Brojen Singh
Abstract:
Recent studies suggest that PML (Promyelocytic Leukemia) suppresses p53 protein degradation by inhibiting MDM2 protein in the nucleus [1], and regulates a number of biological functions [2]. We modeled a PML-MDM2-p53 regulatory network by integrating p53-MDM2 autoregulatory model [3] and related PML pathways. We found that p53 dynamics switched at various dynamical states induced by PML which corr…
▽ More
Recent studies suggest that PML (Promyelocytic Leukemia) suppresses p53 protein degradation by inhibiting MDM2 protein in the nucleus [1], and regulates a number of biological functions [2]. We modeled a PML-MDM2-p53 regulatory network by integrating p53-MDM2 autoregulatory model [3] and related PML pathways. We found that p53 dynamics switched at various dynamical states induced by PML which corresponds to different cellular states. Our results show clear transitions among the stabilized, damped and sustain states under different stress conditions induced by PML in the system. These states in p53 dynamics is the signature of existence of various cellular states and the phase transition like behaviour of these states involve various biological significance.
△ Less
Submitted 20 April, 2015;
originally announced April 2015.
-
Dynamic frequency dependence of bias activated negative capacitance in semiconductor diodes under high forward bias
Authors:
Kanika Bansal,
Mohamed Henini,
Marzook S. Alshammari,
Shouvik Datta
Abstract:
We observed qualitatively dissimilar frequency dependence of negative capacitive response under high charge injection in two sets of junction diodes which are functionally different from each other i.e. electroluminescent diodes and non-luminescent Si-based diodes. Using the technique of bias-activated differential capacitance response, we investigated the mutual dynamics of different rate process…
▽ More
We observed qualitatively dissimilar frequency dependence of negative capacitive response under high charge injection in two sets of junction diodes which are functionally different from each other i.e. electroluminescent diodes and non-luminescent Si-based diodes. Using the technique of bias-activated differential capacitance response, we investigated the mutual dynamics of different rate processes in different diodes. We explain these observations as the mutual competition of fast and slow electronic transition rates albeit differently. This study provides a better understanding of the physics of junction diodes operating under high charge carrier injection and may lead to superior device functionalities.
△ Less
Submitted 18 June, 2014;
originally announced June 2014.
-
Optimizing Key Distribution in Peer to Peer Network Using B-Trees
Authors:
Abdulrahman Aldhaheri,
Hammoud Alshammari,
Majid Alshammari
Abstract:
Peer to peer network architecture introduces many desired features including self-scalability that led to achieving higher efficiency rate than the traditional server-client architecture. This was contributed to the highly distributed architecture of peer to peer network. Meanwhile, the lack of a centralized control unit in peer to peer network introduces some challenge. One of these challenges is…
▽ More
Peer to peer network architecture introduces many desired features including self-scalability that led to achieving higher efficiency rate than the traditional server-client architecture. This was contributed to the highly distributed architecture of peer to peer network. Meanwhile, the lack of a centralized control unit in peer to peer network introduces some challenge. One of these challenges is key distribution and management in such an architecture. This research will explore the possibility of develo** a novel scheme for distributing and managing keys in peer to peer network architecture efficiently.
△ Less
Submitted 22 August, 2013;
originally announced August 2013.
-
Design of a novel electrostatic ion storage ring at KACST
Authors:
M. O. A. El Ghazaly,
S. M. Alshammari,
C. P. Welsch,
H. H. Alharbi
Abstract:
A new electrostatic storage ring for beams at energies up to 30keV.q is currently under development at the National Centre for Mathematics and Physics (NCMP), King Abdulaziz City for Science and Technology (KACST). The ring design is based on the existing electrostatic storage rings, but stretches significantly beyond them in that it shall form the core of a unique flexible experimental facility a…
▽ More
A new electrostatic storage ring for beams at energies up to 30keV.q is currently under development at the National Centre for Mathematics and Physics (NCMP), King Abdulaziz City for Science and Technology (KACST). The ring design is based on the existing electrostatic storage rings, but stretches significantly beyond them in that it shall form the core of a unique flexible experimental facility at KACST. The lattice of this ring has been designed in a way that enables the use of state-of-the-art experimental methods to study electron-ion, laser-ion, and ion-neutral beams interactions. The lattice design also allows for a future upgrade of the ring to a double storage ring structure that would enable ion-ion beam interactions to be performed. In this paper, we present the design of this ring with a focus on beam dynamics calculations for the 7o single-bend racetrack layout. The study is principally based on the SIMION8 program. We complemented this study further by using purpose-written routine and MAD-X simulation code. An in-depth investigation into beam stability under consideration of non-linear field components in the electrostatic optical elements, is presented. Finally, different working points and stability regions are discussed.
△ Less
Submitted 22 February, 2013;
originally announced February 2013.
-
Magnetic and Optical properties of strained films of multiferroic GdMnO3
Authors:
Mohammed S. Al Qahtani,
Marzook Alshammari,
H. J Blythe,
A. M. Fox,
G. A. Gehring,
N. Andreev,
V. Chichkov,
Ya Mukovskii
Abstract:
The effects of strain on a film of mulitferroic GdMnO3 are investigated using both magnetometry and magneto-optic spectroscopy. Optical spectra, in the energy range 1.5eV - 3.5eV, were taken in Faraday geometry in an applied magnetic field and also at remanence. This yielded rich information on the effects of strain on the spin ordering in these films. Epitaxial films of GdMnO3 were grown on SrTiO…
▽ More
The effects of strain on a film of mulitferroic GdMnO3 are investigated using both magnetometry and magneto-optic spectroscopy. Optical spectra, in the energy range 1.5eV - 3.5eV, were taken in Faraday geometry in an applied magnetic field and also at remanence. This yielded rich information on the effects of strain on the spin ordering in these films. Epitaxial films of GdMnO3 were grown on SrTiO3 and LaAlO3 substrates. The LaAlO3 was twinned and so produced a highly strained film whereas the strain was less for the film grown on SrTiO3. The Neél temperatures and coercive fields were measured using zero field data and hysteresis loops obtained using a SQUID magnetometer. Optical absorption data agreed with earlier work on bulk materials. The two well known features in the optical spectrum, the charge transfer transition between Mn d states at ~2eV and the band edge transition from the oxygen p band to the d states at ~3eV are observed in the magnetic circular dichroism; however they behaved very differently both as a function of magnetic field and temperature. This is interpreted in terms of the magnetic ordering of the Mn spins.
△ Less
Submitted 18 November, 2010;
originally announced November 2010.
-
On Using Magnetic and optical methods to determine the size and characteristics of nanoparticles embedded in oxide semiconductors
Authors:
Gillian A. Gehring,
Harry J. Blythe,
Qi Feng,
David S. Score,
Abbas Mokhtari,
Marzook Alshammari,
Mohammed S. Al Qahtani,
A. Mark Fox
Abstract:
Films of oxides doped with transition metals are frequently believed to have magnetic inclusions. Magnetic methods to determine the amount of nanophases and their magnetic characteristics are described. The amount of the sample that is paramagnetic may also be measured. Optical methods are described and shown to be very powerful to determine which defects are also magnetic.
Films of oxides doped with transition metals are frequently believed to have magnetic inclusions. Magnetic methods to determine the amount of nanophases and their magnetic characteristics are described. The amount of the sample that is paramagnetic may also be measured. Optical methods are described and shown to be very powerful to determine which defects are also magnetic.
△ Less
Submitted 15 January, 2010;
originally announced January 2010.
-
Magneto-optical properties of Co/ZnO multilayer films
Authors:
David S Score,
Marzook Alshammari,
Qi Feng,
Harry J Blythe,
A Mark Fox,
Gillian A Gehring,
Zhi-Yong Quan,
Xiao-Li Li,
Xiao-Hong Xu
Abstract:
Multilayer films of ZnO with Co were deposited on glass substrates then annealed in a vacuum. The magnetisation of the films increased with annealing but not the magnitude of the magneto-optical signals. The dielectric functions for the films were calculated using the MCD spectra. A Maxwell Garnett theory of a metallic Co/ZnO mixture is presented. The extent to which this explains the MCD spectr…
▽ More
Multilayer films of ZnO with Co were deposited on glass substrates then annealed in a vacuum. The magnetisation of the films increased with annealing but not the magnitude of the magneto-optical signals. The dielectric functions for the films were calculated using the MCD spectra. A Maxwell Garnett theory of a metallic Co/ZnO mixture is presented. The extent to which this explains the MCD spectra taken on the films is discussed.
△ Less
Submitted 23 September, 2009;
originally announced September 2009.
