GEOTECHNICAL AND ENVIRONMENTAL IMPACTS OF STEEL SLAG IN EMBANKMENTS

Steel slag is a byproduct of iron and steel production by the metallurgical industries. Annually, 21 million tons of steel slag is produced in the United States. Most of the slag is landfilled, which represents a significant economic loss and a waste of valuable land space. Steel slag has great potential for the construction of highway embankments; however, its use has been limited due to its high swelling potential and alkalinity. The swelling potential of steel slags may lead to deterioration of the structural stability of highways, and high alkalinity poses an environmental challenge as it affects the leaching behavior of trace metals. This study seeks a methodology that promotes the use of steel slag in highway embankments by minimizing these two main disadvantages. Accelerated swelling tests were conducted to evaluate the swelling behavior of pure steel slag and water treatment residual (WTR) treated steel slag, where WTR is an alum-rich by-product of drinking water treatment plants. Sequential batch tests and column leach tests, as well as two different numerical analyses, UMDSurf and WiscLEACH, were carried out to check the environmental suitability of the methods. Tests were conducted to study the effect of a common borrow fill material that encapsulated the slag in the embankment and the effects of two subgrade soils on the chemical properties of slag leachate. The results indicated that an increase in WTR content in the steel slag-WTR mixtures yields a decrease in pH and most of the leached metal concentrations, except aluminum. The change in the levels of pH, after passing through encapsulation and subgrade, depends on the natural pHs of materials.

SEQUENTIAL DECISIONS AND PREDICTIONS IN NATURAL LANGUAGE PROCESSING

Natural language processing has achieved great success in a wide range of ap- plications, producing both commercial language services and open-source language tools. However, most methods take a static or batch approach, assuming that the model has all information it needs and makes a one-time prediction. In this disser- tation, we study dynamic problems where the input comes in a sequence instead of all at once, and the output must be produced while the input is arriving. In these problems, predictions are often made based only on partial information. We see this dynamic setting in many real-time, interactive applications. These problems usually involve a trade-off between the amount of input received (cost) and the quality of the output prediction (accuracy). Therefore, the evaluation considers both objectives (e.g., plotting a Pareto curve). Our goal is to develop a formal understanding of sequential prediction and decision-making problems in natural language processing and to propose efficient solutions. Toward this end, we present meta-algorithms that take an existent batch model and produce a dynamic model to handle sequential inputs and outputs. Webuild our framework upon theories of Markov Decision Process (MDP), which allows learning to trade off competing objectives in a principled way. The main machine learning techniques we use are from imitation learning and reinforcement learning, and we advance current techniques to tackle problems arising in our settings. We evaluate our algorithm on a variety of applications, including dependency parsing, machine translation, and question answering. We show that our approach achieves a better cost-accuracy trade-off than the batch approach and heuristic-based decision- making approaches. We first propose a general framework for cost-sensitive prediction, where dif- ferent parts of the input come at different costs. We formulate a decision-making process that selects pieces of the input sequentially, and the selection is adaptive to each instance. Our approach is evaluated on both standard classification tasks and a structured prediction task (dependency parsing). We show that it achieves similar prediction quality to methods that use all input, while inducing a much smaller cost. Next, we extend the framework to problems where the input is revealed incremen- tally in a fixed order. We study two applications: simultaneous machine translation and quiz bowl (incremental text classification). We discuss challenges in this set- ting and show that adding domain knowledge eases the decision-making problem. A central theme throughout the chapters is an MDP formulation of a challenging problem with sequential input/output and trade-off decisions, accompanied by a learning algorithm that solves the MDP.