Showing papers by "University of Texas at Austin published in 2002"
TL;DR: Although the new index is mathematically defined and no human visual system model is explicitly employed, experiments on various image distortion types indicate that it performs significantly better than the widely used distortion metric mean squared error.
Abstract: We propose a new universal objective image quality index, which is easy to calculate and applicable to various image processing applications. Instead of using traditional error summation methods, the proposed index is designed by modeling any image distortion as a combination of three factors: loss of correlation, luminance distortion, and contrast distortion. Although the new index is mathematically defined and no human visual system model is explicitly employed, our experiments on various image distortion types indicate that it performs significantly better than the widely used distortion metric mean squared error. Demonstrative images and an efficient MATLAB implementation of the algorithm are available online at http://anchovy.ece.utexas.edu//spl sim/zwang/research/quality_index/demo.html.
5,285 citations
TL;DR: Neural Evolution of Augmenting Topologies (NEAT) as mentioned in this paper employs a principled method of crossover of different topologies, protecting structural innovation using speciation, and incrementally growing from minimal structure.
Abstract: An important question in neuroevolution is how to gain an advantage from evolving neural network topologies along with weights. We present a method, NeuroEvolution of Augmenting Topologies (NEAT), which outperforms the best fixed-topology method on a challenging benchmark reinforcement learning task. We claim that the increased efficiency is due to (1) employing a principled method of crossover of different topologies, (2) protecting structural innovation using speciation, and (3) incrementally growing from minimal structure. We test this claim through a series of ablation studies that demonstrate that each component is necessary to the system as a whole and to each other. What results is significantly faster learning. NEAT is also an important contribution to GAs because it shows how it is possible for evolution to both optimize and complexify solutions simultaneously, offering the possibility of evolving increasingly complex solutions over generations, and strengthening the analogy with biological evolution.
3,265 citations
TL;DR: In this paper, it was shown that the range of slopes arises mostly due of sys- tematic differences in the velocity dispersions used by different groups for the same galaxies, and that one significant component of the difference results from Ferrarese & Merritt's extrapolation of central velocity dispersion to re= 8( re is the effective radius) using an empirical formula.
Abstract: Observations of nearby galaxies reveal a strong correlation between the mass of the central dark object MBH and the velocity dispersionof the host galaxy, of the form logðMBH=M� Þ¼ � þ � logð�=� 0Þ; how- ever, published estimates of the slopespan a wide range (3.75-5.3). Merritt & Ferrarese have argued that low slopes (d4) arise because of neglect of random measurement errors in the dispersions and an incorrect choice for the dispersion of the Milky Way Galaxy. We show that these explanations and several others account for at most a small part of the slope range. Instead, the range of slopes arises mostly because of sys- tematic differences in the velocity dispersions used by different groups for the same galaxies. The origin of these differences remains unclear, but we suggest that one significant component of the difference results from Ferrarese & Merritt's extrapolation of central velocity dispersions to re= 8( re is the effective radius) using an empirical formula. Another component may arise from dispersion-dependent systematic errors in the mea- surements. A new determination of the slope using 31 galaxies yields � ¼ 4:02 � 0:32, � ¼ 8:13 � 0:06 for � 0 ¼ 200 km s � 1 . The MBH-� relation has an intrinsic dispersion in log MBH that is no larger than 0.25-0.3 dex and may be smaller if observational errors have been underestimated. In an appendix, we present a simple kinematic model for the velocity-dispersion profile of the Galactic bulge. Subject headings: black hole physics — galaxies: bulges — galaxies: fundamental parameters — galaxies: nuclei — Galaxy: bulge — Galaxy: kinematics and dynamics
2,742 citations
TL;DR: In this paper, the authors analyzed data from the National Violence Against Women Survey (NVAWS) of women and men aged 18 to 65 and found that both physical and psychological IPV are associated with significant physical and mental health consequences for both male and female victims.
1,874 citations
TL;DR: This survey covers rollback-recovery techniques that do not require special language constructs and distinguishes between checkpoint-based and log-based protocols, which rely solely on checkpointing for system state restoration.
Abstract: This survey covers rollback-recovery techniques that do not require special language constructs. In the first part of the survey we classify rollback-recovery protocols into checkpoint-based and log-based.Checkpoint-based protocols rely solely on checkpointing for system state restoration. Checkpointing can be coordinated, uncoordinated, or communication-induced. Log-based protocols combine checkpointing with logging of nondeterministic events, encoded in tuples called determinants. Depending on how determinants are logged, log-based protocols can be pessimistic, optimistic, or causal. Throughout the survey, we highlight the research issues that are at the core of rollback-recovery and present the solutions that currently address them. We also compare the performance of different rollback-recovery protocols with respect to a series of desirable properties and discuss the issues that arise in the practical implementations of these protocols.
1,772 citations
TL;DR: In this article, the authors find that institutional ownership concentration is positively related to the pay-for-performance sensitivity of executive compensation and negatively related with the level of compensation, even after controlling for firm size, industry, investment opportunities and performance.
Abstract: We find that institutional ownership concentration is positively related to the pay-for-performance sensitivity of executive compensation and negatively related to the level of compensation, even after controlling for firm size, industry, investment opportunities and performance. These results suggest that the institutions serve a monitoring role in mitigating the agency problem between shareholders and managers. Additionally, we find that clientele effects exist among institutions for firms with certain compensation structures, suggesting that institutions also influence compensation structures through their preferences.
1,755 citations
TL;DR: To provide more conclusive evidence, the available evidence lends itself to the argument that a combination of urban design, land use patterns, and transportation systems that promotes walking and bicycling will help create active, healthier, and more livable communities.
1,718 citations
TL;DR: This online RNA sequence and structure information, the result of extensive analysis, interpretation, data collection, and computer program and web development, is accessible at the Comparative RNA Web (CRW) Site.
Abstract: Background: Comparative analysis of RNA sequences is the basis for the detailed and accurate predictions of RNA structure and the determination of phylogenetic relationships for organisms that span the entire phylogenetic tree. Underlying these accomplishments are very large, wellorganized, and processed collections of RNA sequences. This data, starting with the sequences organized into a database management system and aligned to reveal their higher-order structure, and patterns of conservation and variation for organisms that span the phylogenetic tree, has been collected and analyzed. This type of information can be fundamental for and have an influence on the study of phylogenetic relationships, RNA structure, and the melding of these two fields. Results: We have prepared a large web site that disseminates our comparative sequence and structure models and data. The four major types of comparative information and systems available for the three ribosomal RNAs (5S, 16S, and 23S rRNA), transfer RNA (tRNA), and two of the catalytic intron RNAs (group I and group II) are: (1) Current Comparative Structure Models; (2) Nucleotide Frequency and Conservation Information; (3) Sequence and Structure Data; and (4) Data Access Systems. Conclusions: This online RNA sequence and structure information, the result of extensive analysis, interpretation, data collection, and computer program and web development, is accessible at our Comparative RNA Web (CRW) Site [http://www.rna.icmb.utexas.edu] . In the future, more data and information will be added to these existing categories, new categories will be developed, and additional RNAs will be studied and presented at the CRW Site.
1,676 citations
TL;DR: This article found that African American college students tend to obtain lower grades than their white counterparts, even when they enter college with equivalent test scores, and that negative stereotypes impugning Black students' intellectual abilities play a role in this underperformance.
1,649 citations
TL;DR: The reliability of recharge estimates using different tech- niques is variable as mentioned in this paper, and uncertainties in each approach to estimating recharge underscore the need for application of multiple techniques to increase the expected recharge rates at a site.
Abstract: DOI 10.1007/s10040-0010176-2 Abstract Various techniques are available to quantify recharge; however, choosing appropriate techniques is often difficult. Important considerations in choosing a technique include space/time scales, range, and reliabili- ty of recharge estimates based on different techniques; other factors may limit the application of particular tech- niques. The goal of the recharge study is important be- cause it may dictate the required space/time scales of the recharge estimates. Typical study goals include water-re- source evaluation, which requires information on re- charge over large spatial scales and on decadal time scales; and evaluation of aquifer vulnerability to contam- ination, which requires detailed information on spatial variability and preferential flow. The range of recharge rates that can be estimated using different approaches should be matched to expected recharge rates at a site. The reliability of recharge estimates using different tech- niques is variable. Techniques based on surface-water and unsaturated-zone data provide estimates of potential recharge, whereas those based on groundwater data gen- erally provide estimates of actual recharge. Uncertainties in each approach to estimating recharge underscore the need for application of multiple techniques to increase
1,562 citations
TL;DR: Support is provided for the claim that sociocultural processes foster body dissatisfaction, which in turn increase the risk for bulimic pathology, and it is suggested that prevention and treatment interventions might be enhanced by focusing greater attention on body image disturbances.
28 Jul 2002
TL;DR: This paper presents an elegant and effective framework for combining content and collaboration, which uses a content-based predictor to enhance existing user data, and then provides personalized suggestions through collaborative filtering.
Abstract: Most recommender systems use Collaborative Filtering or Content-based methods to predict new items of interest for a user. While both methods have their own advantages, individually they fail to provide good recommendations in many situations. Incorporating components from both methods, a hybrid recommender system can overcome these shortcomings. In this paper, we present an elegant and effective framework for combining content and collaboration. Our approach uses a content-based predictor tc enhance existing user data, and then provides personalized suggestions through collaborative filtering. We present experimental results that show how this approach, Content-Boosted Collaborative Filtering, performs better than a pure content-based predictor, pure collaborative filter, and a naive hybrid approach.
TL;DR: In this paper, the rheological behavior of compression molded mixtures of polycarbonate containing between 0.5 and 15% carbon nanotubes was investigated using oscillatory rheometry at 260°C.
TL;DR: Leptin-replacement therapy improved glycemic control and decreased triglyceride levels in patients with lipodystrophy and leptin deficiency and was well tolerated.
Abstract: Background The adipocyte hormone leptin is important in regulating energy homeostasis. Since severe lipodystrophy is associated with leptin deficiency, insulin resistance, hypertriglyceridemia, and hepatic steatosis, we assessed whether leptin replacement would ameliorate this condition. Methods Nine female patients (age range, 15 to 42 years; eight with diabetes mellitus) who had lipodystrophy and serum leptin levels of less than 4 ng per milliliter (0.32 nmol per milliliter) received recombinant methionyl human leptin (recombinant leptin). Recombinant leptin was administered subcutaneously twice a day for four months at escalating doses to achieve low, intermediate, and high physiologic replacement levels of leptin. Results During treatment with recombinant leptin, the serum leptin level increased from a mean (±SE) of 1.3±0.3 ng per milliliter to 11.1±2.5 ng per milliliter (0.1±0.02 to 0.9±0.2 nmol per milliliter). The absolute decrease in the glycosylated hemoglobin value was 1.9 percent (95 percent co...
TL;DR: This paper seeks to reconcile and integrate two independent research efforts into a significantly evolved functional basis, and provides a mechanism for evaluating whether future revisions are needed to the functional basis and, if so, how to proceed.
Abstract: In engineering design, all products and artifacts have some intended reason behind their existence: the product or artifact function. Functional modeling provides an abstract, yet direct, method for understanding and representing an overall product or artifact function. Functional modeling also strategically guides design activities such as problem decomposition, physical modeling, product architecting, concept generation, and team organization. A formal function representation is needed to support functional modeling, and a standardized set of function-related terminology leads to repeatable and meaningful results from such a representation. We refer to this representation as a functional basis; in this paper, we seek to reconcile and integrate two independent research efforts into a significantly evolved functional basis. These efforts include research from the National Institute of Standards and Technology and two US universities, and their industrial partners. The overall approach for integrating the functional representations and the final results are presented. This approach also provides a mechanism for evaluating whether future revisions are needed to the functional basis and, if so, how to proceed. The integration process is discussed relative to differences, similarities, insights into the representations, and product validation. Based on the results, a more versatile and comprehensive design vocabulary emerges. This vocabulary will greatly enhance and expand the frontiers of research in design repositories, product architecture, design synthesis, and general product modeling.
Book•
01 Jan 2002TL;DR: The most extensive chapter of the book examines possible mechanisms of peer influence and the evidence in favor of each as mentioned in this paper, concluding that deviant behavior is predominantly social behavior, and criminologists must eventually determine the significance of that fact.
Abstract: Criminologists often allude to 'peer influence' in explanations of crime and delinquency, but the meaning of that concept rarely receives careful attention. Companions in Crime organizes the extensive literature on peer influence and group delinquency into a coherent form for the first time. Chapters focus on the role of peers over the life course, the group nature of delinquent behavior, and the applicability of peer influence for explaining the major features of delinquent behavior. The most extensive chapter of the book examines possible mechanisms of peer influence and the evidence in favor of each. The principal thesis of Companions in Crime is that deviant behavior is predominantly social behavior, and criminologists must eventually determine the significance of that fact.
TL;DR: In this paper, the authors use social network theory to develop a systematic conceptual understanding of how firms located in different countries influence each other's rates of adoption as a result of cohesive and equivalent network relationships.
Abstract: We use panel data on ISO 9000 quality certification in 85 countries between 1993 and 1998 to better understand the cross-national diffusion of an organizational practice. Following neoinstitutional theory, we focus on the coercive, normative, and mimetic effects that result from the exposure of firms in a given country to a powerful source of critical resources, a common pool of relevant technical knowledge, and the experiences of firms located in other countries. We use social network theory to develop a systematic conceptual understanding of how firms located in different countries influence each other's rates of adoption as a result of cohesive and equivalent network relationships. Regression results provide support for our predictions that states and foreign multinationals are the key actors responsible for coercive isomorphism, cohesive trade relationships between countries generate coercive and normative effects, and role-equivalent trade relationships result in learning-based and competitive imitation.
TL;DR: It is discovered that physical dispersion of nonporous, nanoscale, fumed silica particles in glassy amorphous poly(4-methyl-2-pentyne) simultaneously and surprisingly enhances both membrane permeability and selectivity for large organic molecules over small permanent gases.
Abstract: Polymer nanocomposites continue to receive tremendous attention for application in areas such as microelectronics, organic batteries, optics, and catalysis. We have discovered that physical dispersion of nonporous, nanoscale, fumed silica particles in glassy amorphous poly(4-methyl-2-pentyne) simultaneously and surprisingly enhances both membrane permeability and selectivity for large organic molecules over small permanent gases. These highly unusual property enhancements, in contrast to results obtained in conventional filled polymer systems, reflect fumed silica-induced disruption of polymer chain packing and an accompanying subtle increase in the size of free volume elements through which molecular transport occurs, as discerned by positron annihilation lifetime spectroscopy. Such nanoscale hybridization represents an innovative means to tune the separation properties of glassy polymeric media through systematic manipulation of molecular packing.
TL;DR: A liquid crystal system was used for the fabrication of a highly ordered composite material from genetically engineered M13 bacteriophage and zinc sulfide nanocrystals, which spontaneously evolved a self-supporting hybrid film material.
Abstract: A liquid crystal system was used for the fabrication of a highly ordered composite material from genetically engineered M13 bacteriophage and zinc sulfide (ZnS) nanocrystals. The bacteriophage, which formed the basis of the self-ordering system, were selected to have a specific recognition moiety for ZnS crystal surfaces. The bacteriophage were coupled with ZnS solution precursors and spontaneously evolved a self-supporting hybrid film material that was ordered at the nanoscale and at the micrometer scale into approximately 72-micrometer domains, which were continuous over a centimeter length scale. In addition, suspensions were prepared in which the lyotropic liquid crystalline phase behavior of the hybrid material was controlled by solvent concentration and by the use of a magnetic field.
TL;DR: These results demonstrate that the chemical stability of silicon NCs could enable their use as redox-active macromolecular species with the combined optical and charging properties of semiconductor quantum dots.
Abstract: Reversible electrochemical injection of discrete numbers of electrons into sterically stabilized silicon nanocrystals (NCs) (∼2 to 4 nanometers in diameter) was observed by differential pulse voltammetry (DPV) in N , N ′-dimethylformamide and acetonitrile. The electrochemical gap between the onset of electron injection and hole injection—related to the highest occupied and lowest unoccupied molecular orbitals—grew with decreasing nanocrystal size, and the DPV peak potentials above the onset for electron injection roughly correspond to expected Coulomb blockade or quantized double-layer charging energies. Electron transfer reactions between positively and negatively charged nanocrystals (or between charged nanocrystals and molecular redox-active coreactants) occurred that led to electron and hole annihilation, producing visible light. The electrogenerated chemiluminescence spectra exhibited a peak maximum at 640 nanometers, a significant red shift from the photoluminescence maximum (420 nanometers) of the same silicon NC solution. These results demonstrate that the chemical stability of silicon NCs could enable their use as redox-active macromolecular species with the combined optical and charging properties of semiconductor quantum dots.
TL;DR: A clear negative relationship between precipitation and changes in soil organic carbon and nitrogen content when grasslands were invaded by woody vegetation is found, with drier sites gaining, and wetter sites losing, soilorganic carbon.
Abstract: The invasion of woody vegetation into deserts, grasslands and savannas is generally thought to lead to an increase in the amount of carbon stored in those ecosystems. For this reason, shrub and forest expansion (for example, into grasslands) is also suggested to be a substantial, if uncertain, component of the terrestrial carbon sink1,2,3,4,5,6,7,8,9,10,11,12,13,14. Here we investigate woody plant invasion along a precipitation gradient (200 to 1,100 mm yr-1) by comparing carbon and nitrogen budgets and soil δ13C profiles between six pairs of adjacent grasslands, in which one of each pair was invaded by woody species 30 to 100 years ago. We found a clear negative relationship between precipitation and changes in soil organic carbon and nitrogen content when grasslands were invaded by woody vegetation, with drier sites gaining, and wetter sites losing, soil organic carbon. Losses of soil organic carbon at the wetter sites were substantial enough to offset increases in plant biomass carbon, suggesting that current land-based assessments may overestimate carbon sinks. Assessments relying on carbon stored from woody plant invasions to balance emissions may therefore be incorrect.
TL;DR: Brain activation patterns in dyslexic and nonimpaired children during pseudoword and real-word reading tasks that required phonologic analysis provided neurobiological evidence of an underlying disruption in the neural systems for reading in children with dyslexia.
TL;DR: In this article, the authors estimate the amount of spillovers from R&D expenditures on a geographic basis, using a new data set which encompasses most of the world's innovative activity between 1970 and 1995, and find that technology is to a substantial degree local, not global, as the benefits from spillovers are declining with distance.
Abstract: Income convergence across countries turns on whether technological knowledge spillovers are global or local. I estimate the amount of spillovers from R&D expenditures on a geographic basis, using a new data set which encompasses most of the world's innovative activity between 1970 and 1995. I find that technology is to a substantial degree local, not global, as the benefits from spillovers are declining with distance. The distance at which the amount of spillovers is halved is about 1,200 kilometers. I also find that over time, technological knowledge has become considerably more global. Moreover, language skills are important for spillover diffusion. (JEL F0, O1, O3)
TL;DR: This article argued that the underlying principles of state-endorsed'multiculturalism' stand in tension with neoliberal political-economic policies, and that the most effective response to this menace is probably not to engage in frontal opposition to neoliberal regimes, but rather to refuse the dichotomy altogether.
Abstract: This article challenges the assumption that the underlying principles of state-endorsed 'multiculturalism' stand in tension with neoliberal political- economic policies. Based on ethnographic research in Guatemala, it is argued that neoliberalism's cultural project entails pro-active recognition of a minimal package of cultural rights, and an equally vigorous rejection of the rest. The result is a dichotomy between recognised and recalcitrant indigenous subjects, which confronts the indigenous rights movement as a 'menace' even greater than the assimilationist policies of the previous era. It is suggested that the most effective response to this menace is probably not to engage in frontal opposition to neoliberal regimes, but rather to refuse the dichotomy altogether.
TL;DR: In this paper, the authors used live plants for the fabrication of nanoparticles and found that the Au nanoparticles are in a crystalline state and showed defects such as twins in the crystal structure and icosahedral nanoparticles were found.
Abstract: In modern nanotechnology one of the most exciting areas is the interaction between inorganic quantum dots and biological structures. For instance gold clusters surrounded by a shell of organic ligands covalently attach to proteins or other biological substances and can be used for labeling in structural biology. In the present report we show the possibility of using live plants for the fabrication of nanoparticles. Alfalfa plants were grown in an AuCl4 rich environment. The absorption of Au metal by the plants was confirmed by X-ray absorption studies (XAS), and transmission electron microscopy (TEM). Atomic resolution analysis confirmed the nucleation and growth of Au nanoparticles inside the plant and that the Au nanoparticles are in a crystalline state. Images also showed defects such as twins in the crystal structure, and in some cases icosahedral nanoparticles were found. X-ray EDS studies corroborated that the nanoparticles are pure gold. This is the first report on the formation of gold nanoparticl...
13 May 2002
TL;DR: In this paper, insights on why image quality assessment is so difficult are provided by pointing out the weaknesses of the error sensitivity based framework and a new philosophy in designing image quality metrics is proposed.
Abstract: Image quality assessment plays an important role in various image processing applications. A great deal of effort has been made in recent years to develop objective image quality metrics that correlate with perceived quality measurement. Unfortunately, only limited success has been achieved. In this paper, we provide some insights on why image quality assessment is so difficult by pointing out the weaknesses of the error sensitivity based framework, which has been used by most image quality assessment approaches in the literature. Furthermore, we propose a new philosophy in designing image quality metrics: The main function of the human eyes is to extract structural information from the viewing field, and the human visual system is highly adapted for this purpose. Therefore, a measurement of structural distortion should be a good approximation of perceived image distortion. Based on the new philosophy, we implemented a simple but effective image quality indexing algorithm, which is very promising as shown by our current results.
TL;DR: The Ice, Cloud and Land Elevation Satellite (ICESat) mission as discussed by the authors measured changes in elevation of the Greenland and Antarctic ice sheets as part of NASA's Earth Observing System (EOS) of satellites.
TL;DR: Investigation of putative risk factors for binge eating in a longitudinal study of adolescent girls verified that binge eating predicted obesity onset and supported the assertion that these psychosocial and biological factors increase risk for binge Eating.
Abstract: Because little is known about the predictors of binge eating (a risk factor for obesity), a set of putative risk factors for binge eating was investigated in a longitudinal study of adolescent girls. Results verified that binge eating predicted obesity onset. Elevated dieting, pressure to be thin, modeling of eating disturbances, appearance overvaluation, body dissatisfaction, depressive symptoms, emotional eating, body mass, and low self-esteem and social support predicted binge eating onset with 92% accuracy. Classification tree analysis revealed an interaction between appearance overvaluation, body mass, dieting, and depressive symptoms, suggesting qualitatively different pathways to binge eating and identifying subgroups at extreme risk for this outcome. Results support the assertion that these psychosocial and biological factors increase risk for binge eating.
TL;DR: The history, taxonomy, geographic distribution, clinical disease, and therapy of the pathogenic nonpigmented or late-pigmenting rapidly growing mycobacteria (RGM) are reviewed and laboratory diagnosis of RGM using phenotypic methods such as biochemical testing and high-performance liquid chromatography and molecular methods of diagnosis are discussed.
Abstract: The history, taxonomy, geographic distribution, clinical disease, and therapy of the pathogenic nonpigmented or late-pigmenting rapidly growing mycobacteria (RGM) are reviewed. Community-acquired disease and health care-associated disease are highlighted for each species. The latter grouping includes health care-associated outbreaks and pseudo-outbreaks as well as sporadic disease cases. Treatment recommendations for each species and type of disease are also described. Special emphasis is on the Mycobacterium fortuitum group, including M. fortuitum, M. peregrinum, and the unnamed third biovariant complex with its recent taxonomic changes and newly recognized species (including M. septicum, M. mageritense, and proposed species M. houstonense and M. bonickei). The clinical and taxonomic status of M. chelonae, M. abscessus, and M. mucogenicum is also detailed, along with that of the closely related new species, M. immunogenum. Additionally, newly recognized species, M. wolinskyi and M. goodii, as well as M. smegmatis sensu stricto, are included in a discussion of the M. smegmatis group. Laboratory diagnosis of RGM using phenotypic methods such as biochemical testing and high-performance liquid chromatography and molecular methods of diagnosis are also discussed. The latter includes PCR-restriction fragment length polymorphism analysis, hybridization, ribotyping, and sequence analysis. Susceptibility testing and antibiotic susceptibility patterns of the RGM are also annotated, along with the current recommendations from the National Committee for Clinical Laboratory Standards (NCCLS) for mycobacterial susceptibility testing.
01 Oct 2002
TL;DR: This paper proposes physical designs for these Non-Uniform Cache Architectures (NUCAs) and extends these physical designs with logical policies that allow important data to migrate toward the processor within the same level of the cache.
Abstract: Growing wire delays will force substantive changes in the designs of large caches. Traditional cache architectures assume that each level in the cache hierarchy has a single, uniform access time. Increases in on-chip communication delays will make the hit time of large on-chip caches a function of a line's physical location within the cache. Consequently, cache access times will become a continuum of latencies rather than a single discrete latency. This non-uniformity can be exploited to provide faster access to cache lines in the portions of the cache that reside closer to the processor. In this paper, we evaluate a series of cache designs that provides fast hits to multi-megabyte cache memories. We first propose physical designs for these Non-Uniform Cache Architectures (NUCAs). We extend these physical designs with logical policies that allow important data to migrate toward the processor within the same level of the cache. We show that, for multi-megabyte level-two caches, an adaptive, dynamic NUCA design achieves 1.5 times the IPC of a Uniform Cache Architecture of any size, outperforms the best static NUCA scheme by 11%, outperforms the best three-level hierarchy--while using less silicon area--by 13%, and comes within 13% of an ideal minimal hit latency solution.