Structured learning with constrained conditional models

Making complex decisions in real world problems often involves assigning values to sets of interdependent variables where an expressive dependency structure among these can influence, or even dictate, what assignments are possible. Commonly used models typically ignore expressive dependencies since the traditional way of incorporating non-local dependencies is inefficient and hence leads to expensive training and inference. The contribution of this paper is two-fold. First, this paper presents Constrained Conditional Models (CCMs), a?framework that augments linear models with declarative constraints as a way to support decisions in an expressive output space while maintaining modularity and tractability of training. The paper develops, analyzes and compares novel algorithms for CCMs based on Hidden Markov Models and Structured Perceptron. The proposed CCM framework is also compared to task-tailored models, such as semi-CRFs. Second, we propose CoDL, a?constraint-driven learning algorithm, which makes use of constraints to guide semi-supervised learning. We provide theoretical justification for CoDL along with empirical results which show the advantage of using declarative constraints in the context of semi-supervised training of probabilistic models.     

Simultaneous recrystallization, phosphorous diffusion and antireflection coating of silicon films using laser treatment

Laser technique application to polycrystalline silicon thin-film solar cell fabrication on glass substrates has received appreciable attention. In this paper, a laser-doping technique is developed for plasma-deposited amorphous silicon film. A process involving recrystallization, phosphorous diffusion and antireflection coating can be achieved simultaneously using the laser annealing process. The doping precursor, a phosphorous-doped titanium dioxide (TiO₂) solution, is synthesized using a sol-gel method and spin-coated onto the sample. After laser irradiation, the polycrystalline silicon grain size was about 0.5∼1.0 μm with a carrier concentration of 2 × 10¹⁹ cm^− 3 and electron mobility of 92.6 cm²/V s. The average polycrystalline silicon reflectance can be reduced to a value of 4.65% at wavelengths between 400 and 700 nm, indicating the upper TiO₂ film of antireflection coating.     

Transformation of microcrystalline silicon films by excimer-laser-induced crystallization

We describe the excimer-laser-induced crystallization of microcrystalline silicon films deposited by plasma-enhanced chemical vapor deposition (PECVD). Microcrystalline silicon films containing 2 at.% hydrogen can be used as precursor films for the laser recrystallization process without a dehydrogenation step, and provide a wider laser energy fluence process window than the previous explosive recrystallization for low temperature polysilicon (poly-Si) thin-film transistor (TFT) fabrication. Ellipsometry, transmission electron microscopy (TEM), and atomic force microscopy (AFM) are used to evaluate the laser irradiated films. Specially, we describe using atomic force microscopy to obtain plane-view grain microstructure images.     

Knowledge-based system for railway scheduling

In this paper, we describe a Knowledge-Based railway scheduling system (called RSS) for Taiwan Railway Administration's (TRA) railway scheduling operations. In our approach, the scheduling process is divided into two levels: global scheduling and local scheduling. In global scheduling, an initial train diagram is established by the Global Scheduler according to the master scheduling plan without considering conflicts. In local scheduling, scheduling conflicts are repaired by Local Scheduler. An embedded knowledge base provides advice for conflict resolution. Since cycles may exist among a sequence of repairs during the repair process, we also propose a cycle detection and resolution scheme in this paper. Through a graphic user interface, the user can interact with the Automatic Scheduler to solve the large-scale complex train scheduling problem in a cooperative and flexible fashion.     

Fault-tolerant deadlock avoidance algorithm for assembly processes

Unreliable resources pose challenges in design of deadlock avoidance algorithms as resources failures have negative impacts on scheduled production activities and may bring the system to dead states or deadlocks. This paper focuses on the development of a suboptimal polynomial complexity deadlock avoidance algorithm that can operate in the presence of unreliable resources for assembly processes. We formulate a fault-tolerant deadlock avoidance controller synthesis problem for assembly processes based on controlled assembly Petri net (CAPN), a class of Petri nets (PNs) that can model such characteristics as multiple resources and subassembly parts requirement in assembly production processes. The proposed fault-tolerant deadlock avoidance algorithm consists of a nominal algorithm to avoid deadlocks for nominal system state and an exception handling algorithm to deal with resources failures. We analyze the fault-tolerant property of the nominal deadlock avoidance algorithm based on resource unavailability models. Resource unavailability is modeled as loss of tokens in nominal Petri Net models to model unavailability of resources in the course of time-consuming recovery procedures. We define three types of token loss to model 1) resource failures in a single operation, 2) resource failures in multiple operations of a production process and 3) resource failures in multiple operations of multiple production processes. For each type of token loss, we establish sufficient conditions that guarantee the liveness of a CAPN after some tokens are removed. An algorithm is proposed to conduct feasibility analysis by searching for recovery control sequences and to keep as many types of production processes as possible continue production so that the impacts on existing production activities can be reduced.     

Design, optimization, and performance analysis of new photodiode structures for CMOS active-pixel-sensor (APS) imager applications

The dark current in the active-pixel-sensor (APS) cell of a CMOS imager is known to be mainly generated in the regions of bird's beak after the local oxidation of silicon process as well as the surface damage caused by the implantation of high doping concentration. Furthermore, shallow and deep pn-junctions can improve the photo-sensitivity for light of short and long wavelengths, respectively. In this paper, two new photodiode structures using p-substrate and lightly-doped sensor implant SN- as pn-junction photodiode with the regions of bird's beak embraced by SN- and p-field implants, respectively, are proposed and analyzed to reduce dark current and enhance the overall spectral response. 5 /spl mu/m/spl times/5 /spl mu/m APS cells fabricated in a 0.35-/spl mu/m single-poly-triple-metal (1P3M) 3.3-V CMOS process are designed by using the proposed photodiode structures. As shown from the experimental results, the two proposed photodiode structures of 5 /spl mu/m/spl times/5 /spl mu/m APS cells have lower dark currents of 30.6 mV/s and 35.2 mV/s at the reverse-biased voltage of 2 V and higher spectral response, as compared to the conventional structure and other photodiode structures. Thus, the two proposed new photodiode structures can be applied to CMOS imager systems with small pixel size, high resolution, and high quality.     

Bibliometric analysis of Patent Ductus Arteriosus treatments

A bibliometric analysis was performed to assess the quantitative trend of Patent Ductus Arteriosus (PDA) treatment research, including intravenous injection of indomethacin and surgery. The documents studied were retrieved from the Science Citation Index (SCI) for the period from 1991 to 2002. The publication pattern concerning authorship, collaboration, original countries, citation frequency, document type, language of publication, distribution of journals, page count and the most frequently cited papers were performed. The results indicated that either treatment was not the recent emphasis of PDA research. The publishing countries of both treatments have also denoted that these researches were mostly done in Europe and North America. Both surgery and drug treatments had few international collaboration papers. English was the dominant language, and collaboration of two to six authors was the most popular level of co-authorship. This revised version was published online in June 2006 with corrections to the Cover Date.     

Seismic Retrofit of Hollow Rectangular Bridge Columns

The seismic performance of rectangular hollow bridge columns is a significant issue of the high-speed rail project in Taiwan. The flexural ductility and shear capacity of such columns with the configuration of lateral reinforcement used in Taiwan have been studied recently. This paper reports that hollow rectangular bridge columns retrofitted with fiber-reinforced polymer (FRP) sheets were tested under a constant axial load and a cyclically reversed horizontal load to investigate their seismic behavior, including flexural ductility, dissipated energy, and shear capacity. An analytical model was also developed to predict the moment-curvature curve of sections and the load-displacement relationship of columns. Based on the test results, the seismic behavior of such columns will be presented. The test results were also compared to the proposed analytical model. It was found that the ductility factors of the tested piers are in the range from 3.4 to 6.3, and the proposed analytical model can predict the load-displacement relationship of such columns with acceptable accuracy. All in all, FRP sheets can effectively improve both the ductility factor and shear capacity of hollow rectangular bridge columns.     

Lithium plasma emitter for collisionless magnetized plasma experiment

This paper presents a newly developed lithium plasma emitter, which can provide quiescent and low-temperature collisionless conditions for magnetized plasma experiments. This plasma emitter generates thermal emissions of lithium ions and electrons to produce a lithium plasma. Lithium type beta-eucryptite and lanthanum-hexaboride (LaB(6)) powders were mixed and directly heated with a tungsten heater to synthesize ion and electron emissions. As a result, a plasma with a diameter of ~15 cm was obtained in a magnetic mirror configuration. The typical range of electron density was 10(12)-10(13) m(-3) and that of electron temperature was 0.1-0.8 eV with the emitter operation temperature of about 1500 K. The amplitude fluctuations for the plasma density were lower than 1%.     

Process quality improvement of a qualitative response with dynamic characteristic

To achieve quality improvement in manufacturing environments, the design of experiment (DOE) and Taguchi methods are two efficiency approaches to address such problems. Applying those methods to resolve quality improvement frequently focuses on the measurable quality response (or quantitative response). However, the complexity of products or processes is gradually increased and due to the limitation of measuring equipment, the quality response could not be directly measured. Visual inspection or measurement is then used to judge the quality for nonmeasurable response (or qualitative response). Basically, for assessing the quality of a qualitative response, it is initially divided into several classes or ordered categories. As for customers’ requirements gradually changing and applications’ flexibility gradually increasing, a dynamic structure for this process will be another important consideration for manufacturers. That is, it causes the work of quality improvement to be dynamically analyzed. Dr. Taguchi had proposed a dynamic method to analyze such issues. However, only several studies focused on quality improvement of a qualitative response during the static characteristic. Parameter optimization approaches were seldom proposed to address quality improvement of a qualitative response with the dynamic characteristic. Therefore, in this study, we proposed an integrated parameter-optimization approach to resolve quality improvement of a qualitative response with dynamic characteristics. An illustrative example, quality improvement of lead twist during the stamping process for lead frame manufacturing at Science-based park in Taiwan, is employed to demonstrate the effectiveness of the proposed approach.