Insight Extraction for Semiconductor Manufacturing Processes

In the semiconductor manufacturing environment it is very important to understand which factors have the most impact on process outcomes and to control them accordingly. This is usually achieved through design of experiments at process start-up and long term observation of production. As such it relies heavily on the expertise of the process engineer. In this work, we present an automatic approach to extracting useful insights about production processes and equipment based on state-of-the-art Machine Learning techniques. The main goal of this activity is to provide tools to process engineers to accelerate the learning-by-observation phase of process analysis. Using a Metal Deposition process as an example, we highlight various ways in which the extracted information can be employed.

Automatic extraction of Reduced order models from CFD simulations for building energy modeling

Thermal comfort is defined as “that condition of mind which expresses satisfaction with the thermal environment’ [1] [2]. Field studies have been completed in order to establish the governing conditions for thermal comfort [3]. These studies showed that the internal climate of a room was the strongest factor in establishing thermal comfort. Direct manipulation of the internal climate is necessary to retain an acceptable level of thermal comfort. In order for Building Energy Management Systems (BEMS) strategies to be efficiently utilised it is necessary to have the ability to predict the effect that activating a heating/cooling source (radiators, windows and doors) will have on the room. The numerical modelling of the domain can be challenging due to necessity to capture temperature stratification and/or different heat sources (radiators, computers and human beings). Computational Fluid Dynamic (CFD) models are usually utilised for this function because they provide the level of details required. Although they provide the necessary level of accuracy these models tend to be highly computationally expensive especially when transient behaviour needs to be analysed. Consequently they cannot be integrated in BEMS. This paper presents and describes validation of a CFD-ROM method for real-time simulations of building thermal performance. The CFD-ROM method involves the automatic extraction and solution of reduced order models (ROMs) from validated CFD simulations. The test case used in this work is a room of the Environmental Research Institute (ERI) Building at the University College Cork (UCC). ROMs have shown that they are sufficiently accurate with a total error of less than 1% and successfully retain a satisfactory representation of the phenomena modelled. The number of zones in a ROM defines the size and complexity of that ROM. It has been observed that ROMs with a higher number of zones produce more accurate results. As each ROM has a time to solution of less than 20 seconds they can be integrated into the BEMS of a building which opens the potential to real time physics based building energy modelling.

Automatic extraction of reduced-order models from CFD simulations for building energy modelling

Accurate modelling of the internal climate of buildings is essential if Building Energy Management Systems (BEMS) are to efficiently maintain adequate thermal comfort. Computational fluid dynamics (CFD) models are usually utilised to predict internal climate. Nevertheless CFD models, although providing the necessary level of accuracy, are highly computationally expensive, and cannot practically be integrated in BEMS. This paper presents and describes validation of a CFD-ROM method for real-time simulations of building thermal performance. The CFD-ROM method involves the automatic extraction and solution of reduced order models (ROMs) from validated CFD simulations. ROMs are shown to be adequately accurate with a total error below 5% and to retain satisfactory representation of the phenomena modelled. Each ROM has a time to solution under 20seconds, which opens the potential of their integration with BEMS, giving real-time physics-based building energy modelling. A parameter study was conducted to investigate the applicability of the extracted ROM to initial boundary conditions different from those from which it was extracted. The results show that the ROMs retained satisfactory total errors when the initial conditions in the room were varied by ±5°C. This allows the production of a finite number of ROMs with the ability to rapidly model many possible scenarios.

Robust Extraction of Text from Camera Images using Colour and Spatial Information Simultaneously

The importance and use of text extraction from camera based coloured scene images is rapidly increasing with time. Text within a camera grabbed image can contain a huge amount of meta data about that scene. Such meta data can be useful for identification, indexing and retrieval purposes. While the segmentation and recognition of text from document images is quite successful, detection of coloured scene text is a new challenge for all camera based images. Common problems for text extraction from camera based images are the lack of prior knowledge of any kind of text features such as colour, font, size and orientation as well as the location of the probable text regions. In this paper, we document the development of a fully automatic and extremely robust text segmentation technique that can be used for any type of camera grabbed frame be it single image or video. A new algorithm is proposed which can overcome the current problems of text segmentation. The algorithm exploits text appearance in terms of colour and spatial distribution. When the new text extraction technique was tested on a variety of camera based images it was found to out perform existing techniques (or something similar). The proposed technique also overcomes any problems that can arise due to an unconstraint complex background. The novelty in the works arises from the fact that this is the first time that colour and spatial information are used simultaneously for the purpose of text extraction.

Fast automatic two-stage nonlinear model identification based on the extreme learning machine

It is convenient and effective to solve nonlinear problems with a model that has a linear-in-the-parameters (LITP) structure. However, the nonlinear parameters (e.g. the width of Gaussian function) of each model term needs to be pre-determined either from expert experience or through exhaustive search. An alternative approach is to optimize them by a gradient-based technique (e.g. Newton’s method). Unfortunately, all of these methods still need a lot of computations. Recently, the extreme learning machine (ELM) has shown its advantages in terms of fast learning from data, but the sparsity of the constructed model cannot be guaranteed. This paper proposes a novel algorithm for automatic construction of a nonlinear system model based on the extreme learning machine. This is achieved by effectively integrating the ELM and leave-one-out (LOO) cross validation with our two-stage stepwise construction procedure [1]. The main objective is to improve the compactness and generalization capability of the model constructed by the ELM method. Numerical analysis shows that the proposed algorithm only involves about half of the computation of orthogonal least squares (OLS) based method. Simulation examples are included to confirm the efficacy and superiority of the proposed technique.

Effect of short-term macrophage depletion in the development of posterior capsule opacification in rodents

Aim: To evaluate the role of macrophages in the development of posterior capsule opacification (PCO). Methods: For this purpose, an extracapsular lens extraction was performed in 18 consecutive Sprague-Dawley rats. Animals were treated with liposomal clodronate (Cl MDP-lip-treated group, n = 10) or phosphate-buffered saline (PBS) (control group, n = 8) 1 day preoperatively and on the first day postoperatively, and sacrificed 3 days postoperatively. Masked clinical, light microscopy and immunohistochemistry studies were conducted. The Fisher exact test and randomisation test were used to assess statistically differences between groups. Results: A statistically significant reduction in the number of macrophages (ED1+, ED7+, ED8+) was found in the Cl MDP-lip-treated group compared with the PBS-lip-treated group (p = 0.048, p = 0.004, p = 0.027, respectively). There were no statistically significant differences with regards to the presence/absence of central opacification (p = 0.29) and capsular wrinkling (p = 0.21) as detected clinically between groups. Similarly, a qualitative evaluation of the degree of PCO with regards to lens epithelial cell (LEC) proliferation, capsular wrinkling and Soemmerring ring formation showed no statistically significance between groups (p = 0.27, p = 0.061, p = 1.0, respectively). However, a statistically significant reduction in the number of lens epithelial cells (LEC) counted in the centre of the posterior capsule was found in the Cl MDP-lip- treated group (p = 0.009). Conclusion: Depletion of macrophages was accompanied by a reduction in LEC in the centre of the posterior capsule in rodents.

Machine learning classification of surgical pathology reports and chunk recognition for information extraction noise reduction

Background and aims: Machine learning techniques for the text mining of cancer-related clinical documents have not been sufficiently explored. Here some techniques are presented for the pre-processing of free-text breast cancer pathology reports, with the aim of facilitating the extraction of information relevant to cancer staging.

Materials and methods: The first technique was implemented using the freely available software RapidMiner to classify the reports according to their general layout: ‘semi-structured’ and ‘unstructured’. The second technique was developed using the open source language engineering framework GATE and aimed at the prediction of chunks of the report text containing information pertaining to the cancer morphology, the tumour size, its hormone receptor status and the number of positive nodes. The classifiers were trained and tested respectively on sets of 635 and 163 manually classified or annotated reports, from the Northern Ireland Cancer Registry.

Results: The best result of 99.4% accuracy – which included only one semi-structured report predicted as unstructured – was produced by the layout classifier with the k nearest algorithm, using the binary term occurrence word vector type with stopword filter and pruning. For chunk recognition, the best results were found using the PAUM algorithm with the same parameters for all cases, except for the prediction of chunks containing cancer morphology. For semi-structured reports the performance ranged from 0.97 to 0.94 and from 0.92 to 0.83 in precision and recall, while for unstructured reports performance ranged from 0.91 to 0.64 and from 0.68 to 0.41 in precision and recall. Poor results were found when the classifier was trained on semi-structured reports but tested on unstructured.

Conclusions: These results show that it is possible and beneficial to predict the layout of reports and that the accuracy of prediction of which segments of a report may contain certain information is sensitive to the report layout and the type of information sought.