Refinement methods for software architecture design using the software architecture model

Modern software systems are often large and complicated. To better understand, develop, and manage large software systems, researchers have studied software architectures that provide the top level overall structural design of software systems for the last decade. One major research focus on software architectures is formal architecture description languages, but most existing research focuses primarily on the descriptive capability and puts less emphasis on software architecture design methods and formal analysis techniques, which are necessary to develop correct software architecture design. ^ Refinement is a general approach of adding details to a software design. A formal refinement method can further ensure certain design properties. This dissertation proposes refinement methods, including a set of formal refinement patterns and complementary verification techniques, for software architecture design using Software Architecture Model (SAM), which was developed at Florida International University. First, a general guideline for software architecture design in SAM is proposed. Second, specification construction through property-preserving refinement patterns is discussed. The refinement patterns are categorized into connector refinement, component refinement and high-level Petri nets refinement. These three levels of refinement patterns are applicable to overall system interaction, architectural components, and underlying formal language, respectively. Third, verification after modeling as a complementary technique to specification refinement is discussed. Two formal verification tools, the Stanford Temporal Prover (STeP) and the Simple Promela Interpreter (SPIN), are adopted into SAM to develop the initial models. Fourth, formalization and refinement of security issues are studied. A method for security enforcement in SAM is proposed. The Role-Based Access Control model is formalized using predicate transition nets and Z notation. The patterns of enforcing access control and auditing are proposed. Finally, modeling and refining a life insurance system is used to demonstrate how to apply the refinement patterns for software architecture design using SAM and how to integrate the access control model. ^ The results of this dissertation demonstrate that a refinement method is an effective way to develop a high assurance system. The method developed in this dissertation extends existing work on modeling software architectures using SAM and makes SAM a more usable and valuable formal tool for software architecture design. ^

Signal processing of physiological signals for automated assessment of stress in computer users

The need to provide computers with the ability to distinguish the affective state of their users is a major requirement for the practical implementation of affective computing concepts. This dissertation proposes the application of signal processing methods on physiological signals to extract from them features that can be processed by learning pattern recognition systems to provide cues about a person's affective state. In particular, combining physiological information sensed from a user's left hand in a non-invasive way with the pupil diameter information from an eye-tracking system may provide a computer with an awareness of its user's affective responses in the course of human-computer interactions. In this study an integrated hardware-software setup was developed to achieve automatic assessment of the affective status of a computer user. A computer-based "Paced Stroop Test" was designed as a stimulus to elicit emotional stress in the subject during the experiment. Four signals: the Galvanic Skin Response (GSR), the Blood Volume Pulse (BVP), the Skin Temperature (ST) and the Pupil Diameter (PD), were monitored and analyzed to differentiate affective states in the user. Several signal processing techniques were applied on the collected signals to extract their most relevant features. These features were analyzed with learning classification systems, to accomplish the affective state identification. Three learning algorithms: Naïve Bayes, Decision Tree and Support Vector Machine were applied to this identification process and their levels of classification accuracy were compared. The results achieved indicate that the physiological signals monitored do, in fact, have a strong correlation with the changes in the emotional states of the experimental subjects. These results also revealed that the inclusion of pupil diameter information significantly improved the performance of the emotion recognition system. ^

A method for enhancing digital information displayed to computer users with visual refractive errors via spatial and spectral processing

This research pursued the conceptualization, implementation, and verification of a system that enhances digital information displayed on an LCD panel to users with visual refractive errors. The target user groups for this system are individuals who have moderate to severe visual aberrations for which conventional means of compensation, such as glasses or contact lenses, does not improve their vision. This research is based on a priori knowledge of the user's visual aberration, as measured by a wavefront analyzer. With this information it is possible to generate images that, when displayed to this user, will counteract his/her visual aberration. The method described in this dissertation advances the development of techniques for providing such compensation by integrating spatial information in the image as a means to eliminate some of the shortcomings inherent in using display devices such as monitors or LCD panels. Additionally, physiological considerations are discussed and integrated into the method for providing said compensation. In order to provide a realistic sense of the performance of the methods described, they were tested by mathematical simulation in software, as well as by using a single-lens high resolution CCD camera that models an aberrated eye, and finally with human subjects having various forms of visual aberrations. Experiments were conducted on these systems and the data collected from these experiments was evaluated using statistical analysis. The experimental results revealed that the pre-compensation method resulted in a statistically significant improvement in vision for all of the systems. Although significant, the improvement was not as large as expected for the human subject tests. Further analysis suggest that even under the controlled conditions employed for testing with human subjects, the characterization of the eye may be changing. This would require real-time monitoring of relevant variables (e.g. pupil diameter) and continuous adjustment in the pre-compensation process to yield maximum viewing enhancement.

Formal verification and testing of software architectural models

Ensuring the correctness of software has been the major motivation in software research, constituting a Grand Challenge. Due to its impact in the final implementation, one critical aspect of software is its architectural design. By guaranteeing a correct architectural design, major and costly flaws can be caught early on in the development cycle. Software architecture design has received a lot of attention in the past years, with several methods, techniques and tools developed. However, there is still more to be done, such as providing adequate formal analysis of software architectures. On these regards, a framework to ensure system dependability from design to implementation has been developed at FIU (Florida International University). This framework is based on SAM (Software Architecture Model), an ADL (Architecture Description Language), that allows hierarchical compositions of components and connectors, defines an architectural modeling language for the behavior of components and connectors, and provides a specification language for the behavioral properties. The behavioral model of a SAM model is expressed in the form of Petri nets and the properties in first order linear temporal logic.^ This dissertation presents a formal verification and testing approach to guarantee the correctness of Software Architectures. The Software Architectures studied are expressed in SAM. For the formal verification approach, the technique applied was model checking and the model checker of choice was Spin. As part of the approach, a SAM model is formally translated to a model in the input language of Spin and verified for its correctness with respect to temporal properties. In terms of testing, a testing approach for SAM architectures was defined which includes the evaluation of test cases based on Petri net testing theory to be used in the testing process at the design level. Additionally, the information at the design level is used to derive test cases for the implementation level. Finally, a modeling and analysis tool (SAM tool) was implemented to help support the design and analysis of SAM models. The results show the applicability of the approach to testing and verification of SAM models with the aid of the SAM tool.^

System-on-a-Chip (SoC) based hardware acceleration in Register Transfer Level (RTL) design

Today, modern System-on-a-Chip (SoC) systems have grown rapidly due to the increased processing power, while maintaining the size of the hardware circuit. The number of transistors on a chip continues to increase, but current SoC designs may not be able to exploit the potential performance, especially with energy consumption and chip area becoming two major concerns. Traditional SoC designs usually separate software and hardware. Thus, the process of improving the system performance is a complicated task for both software and hardware designers. The aim of this research is to develop hardware acceleration workflow for software applications. Thus, system performance can be improved with constraints of energy consumption and on-chip resource costs. The characteristics of software applications can be identified by using profiling tools. Hardware acceleration can have significant performance improvement for highly mathematical calculations or repeated functions. The performance of SoC systems can then be improved, if the hardware acceleration method is used to accelerate the element that incurs performance overheads. The concepts mentioned in this study can be easily applied to a variety of sophisticated software applications. The contributions of SoC-based hardware acceleration in the hardware-software co-design platform include the following: (1) Software profiling methods are applied to H.264 Coder-Decoder (CODEC) core. The hotspot function of aimed application is identified by using critical attributes such as cycles per loop, loop rounds, etc. (2) Hardware acceleration method based on Field-Programmable Gate Array (FPGA) is used to resolve system bottlenecks and improve system performance. The identified hotspot function is then converted to a hardware accelerator and mapped onto the hardware platform. Two types of hardware acceleration methods – central bus design and co-processor design, are implemented for comparison in the proposed architecture. (3) System specifications, such as performance, energy consumption, and resource costs, are measured and analyzed. The trade-off of these three factors is compared and balanced. Different hardware accelerators are implemented and evaluated based on system requirements. 4) The system verification platform is designed based on Integrated Circuit (IC) workflow. Hardware optimization techniques are used for higher performance and less resource costs. Experimental results show that the proposed hardware acceleration workflow for software applications is an efficient technique. The system can reach 2.8X performance improvements and save 31.84% energy consumption by applying the Bus-IP design. The Co-processor design can have 7.9X performance and save 75.85% energy consumption.

Fabrication of Dense Non-Circular Nanomagnetic Device Arrays Using Self-Limiting Low-Energy Glow-Discharge Processing

We describe a low-energy glow-discharge process using reactive ion etching system that enables non-circular device patterns, such as squares or hexagons, to be formed from a precursor array of uniform circular openings in polymethyl methacrylate, PMMA, defined by electron beam lithography. This technique is of a particular interest for bit-patterned magnetic recording medium fabrication, where close packed square magnetic bits may improve its recording performance. The process and results of generating close packed square patterns by self-limiting low-energy glow-discharge are investigated. Dense magnetic arrays formed by electrochemical deposition of nickel over self-limiting formed molds are demonstrated.

A mathematical resolution to log transformations and the binning effect in applied processing of data in flow cytometry

This dissertation develops a new mathematical approach that overcomes the effect of a data processing phenomenon known as "histogram binning" inherent to flow cytometry data. A real-time procedure is introduced to prove the effectiveness and fast implementation of such an approach on real-world data. The histogram binning effect is a dilemma posed by two seemingly antagonistic developments: (1) flow cytometry data in its histogram form is extended in its dynamic range to improve its analysis and interpretation, and (2) the inevitable dynamic range extension introduces an unwelcome side effect, the binning effect, which skews the statistics of the data, undermining as a consequence the accuracy of the analysis and the eventual interpretation of the data. Researchers in the field contended with such a dilemma for many years, resorting either to hardware approaches that are rather costly with inherent calibration and noise effects; or have developed software techniques based on filtering the binning effect but without successfully preserving the statistical content of the original data. The mathematical approach introduced in this dissertation is so appealing that a patent application has been filed. The contribution of this dissertation is an incremental scientific innovation based on a mathematical framework that will allow researchers in the field of flow cytometry to improve the interpretation of data knowing that its statistical meaning has been faithfully preserved for its optimized analysis. Furthermore, with the same mathematical foundation, proof of the origin of such an inherent artifact is provided. These results are unique in that new mathematical derivations are established to define and solve the critical problem of the binning effect faced at the experimental assessment level, providing a data platform that preserves its statistical content. In addition, a novel method for accumulating the log-transformed data was developed. This new method uses the properties of the transformation of statistical distributions to accumulate the output histogram in a non-integer and multi-channel fashion. Although the mathematics of this new mapping technique seem intricate, the concise nature of the derivations allow for an implementation procedure that lends itself to a real-time implementation using lookup tables, a task that is also introduced in this dissertation.

A method for enhancing digital information displayed to computer users with visual refractive errors via spatial and spectral processing

This research pursued the conceptualization, implementation, and verification of a system that enhances digital information displayed on an LCD panel to users with visual refractive errors. The target user groups for this system are individuals who have moderate to severe visual aberrations for which conventional means of compensation, such as glasses or contact lenses, does not improve their vision. This research is based on a priori knowledge of the user's visual aberration, as measured by a wavefront analyzer. With this information it is possible to generate images that, when displayed to this user, will counteract his/her visual aberration. The method described in this dissertation advances the development of techniques for providing such compensation by integrating spatial information in the image as a means to eliminate some of the shortcomings inherent in using display devices such as monitors or LCD panels. Additionally, physiological considerations are discussed and integrated into the method for providing said compensation. In order to provide a realistic sense of the performance of the methods described, they were tested by mathematical simulation in software, as well as by using a single-lens high resolution CCD camera that models an aberrated eye, and finally with human subjects having various forms of visual aberrations. Experiments were conducted on these systems and the data collected from these experiments was evaluated using statistical analysis. The experimental results revealed that the pre-compensation method resulted in a statistically significant improvement in vision for all of the systems. Although significant, the improvement was not as large as expected for the human subject tests. Further analysis suggest that even under the controlled conditions employed for testing with human subjects, the characterization of the eye may be changing. This would require real-time monitoring of relevant variables (e.g. pupil diameter) and continuous adjustment in the pre-compensation process to yield maximum viewing enhancement.

Formal verification and testing of software architectural models

Ensuring the correctness of software has been the major motivation in software research, constituting a Grand Challenge. Due to its impact in the final implementation, one critical aspect of software is its architectural design. By guaranteeing a correct architectural design, major and costly flaws can be caught early on in the development cycle. Software architecture design has received a lot of attention in the past years, with several methods, techniques and tools developed. However, there is still more to be done, such as providing adequate formal analysis of software architectures. On these regards, a framework to ensure system dependability from design to implementation has been developed at FIU (Florida International University). This framework is based on SAM (Software Architecture Model), an ADL (Architecture Description Language), that allows hierarchical compositions of components and connectors, defines an architectural modeling language for the behavior of components and connectors, and provides a specification language for the behavioral properties. The behavioral model of a SAM model is expressed in the form of Petri nets and the properties in first order linear temporal logic. This dissertation presents a formal verification and testing approach to guarantee the correctness of Software Architectures. The Software Architectures studied are expressed in SAM. For the formal verification approach, the technique applied was model checking and the model checker of choice was Spin. As part of the approach, a SAM model is formally translated to a model in the input language of Spin and verified for its correctness with respect to temporal properties. In terms of testing, a testing approach for SAM architectures was defined which includes the evaluation of test cases based on Petri net testing theory to be used in the testing process at the design level. Additionally, the information at the design level is used to derive test cases for the implementation level. Finally, a modeling and analysis tool (SAM tool) was implemented to help support the design and analysis of SAM models. The results show the applicability of the approach to testing and verification of SAM models with the aid of the SAM tool.

User Interface Design for Electronic Learning Software: Promoting Usability and Facilitating Learning for Children

Computing devices have become ubiquitous in our technologically-advanced world, serving as vehicles for software applications that provide users with a wide array of functions. Among these applications are electronic learning software, which are increasingly being used to educate and evaluate individuals ranging from grade school students to career professionals. This study will evaluate the design and implementation of user interfaces in these pieces of software. Specifically, it will explore how these interfaces can be developed to facilitate the use of electronic learning software by children. In order to do this, research will be performed in the area of human-computer interaction, focusing on cognitive psychology, user interface design, and software development. This information will be analyzed in order to design a user interface that provides an optimal user experience for children. This group will test said interface, as well as existing applications, in order to measure its usability. The objective of this study is to design a user interface that makes electronic learning software more usable for children, facilitating their learning process and increasing their academic performance. This study will be conducted by using the Adobe Creative Suite to design the user interface and an Integrated Development Environment to implement functionality. These are digital tools that are available on computing devices such as desktop computers, laptops, and smartphones, which will be used for the development of software. By using these tools, I hope to create a user interface for electronic learning software that promotes usability while maintaining functionality. This study will address the increasing complexity of computing software seen today – an issue that has risen due to the progressive implementation of new functionality. This issue is having a detrimental effect on the usability of electronic learning software, increasing the learning curve for targeted users such as children. As we make electronic learning software an integral part of educational programs in our schools, it is important to address this in order to guarantee them a successful learning experience.

User Interface Design for Electronic Learning Software: Promoting Usability and Facilitating Learning for Children

Computing devices have become ubiquitous in our technologically-advanced world, serving as vehicles for software applications that provide users with a wide array of functions. Among these applications are electronic learning software, which are increasingly being used to educate and evaluate individuals ranging from grade school students to career professionals. This study will evaluate the design and implementation of user interfaces in these pieces of software. Specifically, it will explore how these interfaces can be developed to facilitate the use of electronic learning software by children. In order to do this, research will be performed in the area of human-computer interaction, focusing on cognitive psychology, user interface design, and software development. This information will be analyzed in order to design a user interface that provides an optimal user experience for children. This group will test said interface, as well as existing applications, in order to measure its usability. The objective of this study is to design a user interface that makes electronic learning software more usable for children, facilitating their learning process and increasing their academic performance. This study will be conducted by using the Adobe Creative Suite to design the user interface and an Integrated Development Environment to implement functionality. These are digital tools that are available on computing devices such as desktop computers, laptops, and smartphones, which will be used for the development of software. By using these tools, I hope to create a user interface for electronic learning software that promotes usability while maintaining functionality. This study will address the increasing complexity of computing software seen today – an issue that has risen due to the progressive implementation of new functionality. This issue is having a detrimental effect on the usability of electronic learning software, increasing the learning curve for targeted users such as children. As we make electronic learning software an integral part of educational programs in our schools, it is important to address this in order to guarantee them a successful learning experience.

Mecanismo de tolerância a falhas através de escalonamento para uma arquitetura reconfigurável de grão grosso

The continuous evolution of integrated circuit technology has allowed integrating thousands of transistors on a single chip. This is due to the miniaturization process, which reduces the diameter of wires and transistors. One drawback of this process is that the circuit becomes more fragile and susceptible to break, making the circuit more susceptible to permanent faults during the manufacturing process as well as during their lifetime. Coarse Grained Reconfigurable Architectures (CGRAs) have been used as an alternative to traditional architectures in an attempt to tolerate such faults due to its intrinsic hardware redundancy and high performance. This work proposes a fault tolerance mechanism in a CGRA in order to increase the architecture fault tolerance even considering a high fault rate. The proposed mechanism was added to the scheduler, which is the mechanism responsible for mapping instructions onto the architecture. The instruction mapping occurs at runtime, translating binary code without the need for recompilation. Furthermore, to allow faster implementation, instruction mapping is performed using a greedy module scheduling algorithm, which consists of a software pipeline technique for loop acceleration. The results show that, even with the proposed mechanism, the time for mapping instructions is still in order of microseconds. This result allows that instruction mapping process remains at runtime. In addition, a study was also carried out mapping scheduler rate. The results demonstrate that even at fault rates over 50% in functional units and interconnection components, the scheduler was able to map instructions onto the architecture in most of the tested applications.

Otimização do controle do diagrama de radiação de radares de varredura para rastreio de foguetes usando o método GAMMC para o Caso Planar (GAMMC-P)

Launching centers are designed for scientific and commercial activities with aerospace vehicles. Rockets Tracking Systems (RTS) are part of the infrastructure of these centers and they are responsible for collecting and processing the data trajectory of vehicles. Generally, Parabolic Reflector Radars (PRRs) are used in RTS. However, it is possible to use radars with antenna arrays, or Phased Arrays (PAs), so called Phased Arrays Radars (PARs). Thus, the excitation signal of each radiating element of the array can be adjusted to perform electronic control of the radiation pattern in order to improve functionality and maintenance of the system. Therefore, in the implementation and reuse projects of PARs, modeling is subject to various combinations of excitation signals, producing a complex optimization problem due to the large number of available solutions. In this case, it is possible to use offline optimization methods, such as Genetic Algorithms (GAs), to calculate the problem solutions, which are stored for online applications. Hence, the Genetic Algorithm with Maximum-Minimum Crossover (GAMMC) optimization method was used to develop the GAMMC-P algorithm that optimizes the modeling step of radiation pattern control from planar PAs. Compared with a conventional crossover GA, the GAMMC has a different approach from the conventional one, because it performs the crossover of the fittest individuals with the least fit individuals in order to enhance the genetic diversity. Thus, the GAMMC prevents premature convergence, increases population fitness and reduces the processing time. Therefore, the GAMMC-P uses a reconfigurable algorithm with multiple objectives, different coding and genetic operator MMC. The test results show that GAMMC-P reached the proposed requirements for different operating conditions of a planar RAV.

An automated approach for performance deviation analysis of evolving software systems

A manutenção e evolução de sistemas de software tornou-se uma tarefa bastante crítica ao longo dos últimos anos devido à diversidade e alta demanda de funcionalidades, dispositivos e usuários. Entender e analisar como novas mudanças impactam os atributos de qualidade da arquitetura de tais sistemas é um pré-requisito essencial para evitar a deterioração de sua qualidade durante sua evolução. Esta tese propõe uma abordagem automatizada para a análise de variação do atributo de qualidade de desempenho em termos de tempo de execução (tempo de resposta). Ela é implementada por um framework que adota técnicas de análise dinâmica e mineração de repositório de software para fornecer uma forma automatizada de revelar fontes potenciais – commits e issues – de variação de desempenho em cenários durante a evolução de sistemas de software. A abordagem define quatro fases: (i) preparação – escolher os cenários e preparar os releases alvos; (ii) análise dinâmica – determinar o desempenho de cenários e métodos calculando seus tempos de execução; (iii) análise de variação – processar e comparar os resultados da análise dinâmica para releases diferentes; e (iv) mineração de repositório – identificar issues e commits associados com a variação de desempenho detectada. Estudos empíricos foram realizados para avaliar a abordagem de diferentes perspectivas. Um estudo exploratório analisou a viabilidade de se aplicar a abordagem em sistemas de diferentes domínios para identificar automaticamente elementos de código fonte com variação de desempenho e as mudanças que afetaram tais elementos durante uma evolução. Esse estudo analisou três sistemas: (i) SIGAA – um sistema web para gerência acadêmica; (ii) ArgoUML – uma ferramenta de modelagem UML; e (iii) Netty – um framework para aplicações de rede. Outro estudo realizou uma análise evolucionária ao aplicar a abordagem em múltiplos releases do Netty, e dos frameworks web Wicket e Jetty. Nesse estudo foram analisados 21 releases (sete de cada sistema), totalizando 57 cenários. Em resumo, foram encontrados 14 cenários com variação significante de desempenho para Netty, 13 para Wicket e 9 para Jetty. Adicionalmente, foi obtido feedback de oito desenvolvedores desses sistemas através de um formulário online. Finalmente, no último estudo, um modelo de regressão para desempenho foi desenvolvido visando indicar propriedades de commits que são mais prováveis a causar degradação de desempenho. No geral, 997 commits foram minerados, sendo 103 recuperados de elementos de código fonte degradados e 19 de otimizados, enquanto 875 não tiveram impacto no tempo de execução. O número de dias antes de disponibilizar o release e o dia da semana se mostraram como as variáveis mais relevantes dos commits que degradam desempenho no nosso modelo. A área de característica de operação do receptor (ROC – Receiver Operating Characteristic) do modelo de regressão é 60%, o que significa que usar o modelo para decidir se um commit causará degradação ou não é 10% melhor do que uma decisão aleatória.

An automated approach for performance deviation analysis of evolving software systems

A manutenção e evolução de sistemas de software tornou-se uma tarefa bastante crítica ao longo dos últimos anos devido à diversidade e alta demanda de funcionalidades, dispositivos e usuários. Entender e analisar como novas mudanças impactam os atributos de qualidade da arquitetura de tais sistemas é um pré-requisito essencial para evitar a deterioração de sua qualidade durante sua evolução. Esta tese propõe uma abordagem automatizada para a análise de variação do atributo de qualidade de desempenho em termos de tempo de execução (tempo de resposta). Ela é implementada por um framework que adota técnicas de análise dinâmica e mineração de repositório de software para fornecer uma forma automatizada de revelar fontes potenciais – commits e issues – de variação de desempenho em cenários durante a evolução de sistemas de software. A abordagem define quatro fases: (i) preparação – escolher os cenários e preparar os releases alvos; (ii) análise dinâmica – determinar o desempenho de cenários e métodos calculando seus tempos de execução; (iii) análise de variação – processar e comparar os resultados da análise dinâmica para releases diferentes; e (iv) mineração de repositório – identificar issues e commits associados com a variação de desempenho detectada. Estudos empíricos foram realizados para avaliar a abordagem de diferentes perspectivas. Um estudo exploratório analisou a viabilidade de se aplicar a abordagem em sistemas de diferentes domínios para identificar automaticamente elementos de código fonte com variação de desempenho e as mudanças que afetaram tais elementos durante uma evolução. Esse estudo analisou três sistemas: (i) SIGAA – um sistema web para gerência acadêmica; (ii) ArgoUML – uma ferramenta de modelagem UML; e (iii) Netty – um framework para aplicações de rede. Outro estudo realizou uma análise evolucionária ao aplicar a abordagem em múltiplos releases do Netty, e dos frameworks web Wicket e Jetty. Nesse estudo foram analisados 21 releases (sete de cada sistema), totalizando 57 cenários. Em resumo, foram encontrados 14 cenários com variação significante de desempenho para Netty, 13 para Wicket e 9 para Jetty. Adicionalmente, foi obtido feedback de oito desenvolvedores desses sistemas através de um formulário online. Finalmente, no último estudo, um modelo de regressão para desempenho foi desenvolvido visando indicar propriedades de commits que são mais prováveis a causar degradação de desempenho. No geral, 997 commits foram minerados, sendo 103 recuperados de elementos de código fonte degradados e 19 de otimizados, enquanto 875 não tiveram impacto no tempo de execução. O número de dias antes de disponibilizar o release e o dia da semana se mostraram como as variáveis mais relevantes dos commits que degradam desempenho no nosso modelo. A área de característica de operação do receptor (ROC – Receiver Operating Characteristic) do modelo de regressão é 60%, o que significa que usar o modelo para decidir se um commit causará degradação ou não é 10% melhor do que uma decisão aleatória.