We investigate the performance of a machine-based lip-reading system using both shape-only parameters and full shape and appearance parameters. Furthermore, we contrast the performance of a machine-based lip-reading system with human lip-reading ability. We find that the automated system outperforms human lip-readers. Curiously however, for relatively simple tasks there is little improvement in...

In this paper, we address the problem of machine translation (MT) of domain-specific texts for which large amounts of parallel data for training are not available. We focus on the IT domain and on English to Portuguese machine translation, and compare different strategies for improving system performance over two baselines, the first using only large dataset of out-of-domain data, and the secon...

This proposal is a first pass at adding what is required in 08-015r3 for the phy and link layer state diagrams to include the low power transceiver options of partial and slumber modes for SAS. 1) The following are the proposed changes to the SP state machine. 6.8.1 SP state machine overview The SP state machine controls the phy reset sequence. This state machine consists of three four sets of ...

Poplog is a portable interactive AI development environment available on a "ur}g" of op""ating systems and machines. It includes incremental compilers for bommon Lisp, Pop-ll, Prolog and Standard ML, along with tools for adding new incremental compilers. All the languages share a common development environment and data structures can be shared between programs written in the different languages...

Schema mapping is essential to manage schema heterogeneity among different sources. Schema mapping can be conducted by using machine learning algorithms or by knowledge engineering approaches. These two approaches have advantages and disadvantages. The machine learning approaches can learn their model using the data, but they are static, so they cannot be modified to reflect the domain data cha...

Intelligent user interfaces, such as recommender systems and email classifiers, use machine learning algorithms to customize their behavior to the preferences of an end user. Although these learning systems are somewhat reliable, they are not perfectly accurate. Traditionally, end users who need to correct these learning systems can only provide more labeled training data. In this paper, we foc...

In this paper, we study the impact of using a domain-specific bilingual lexicon on the performance of an Example-Based Machine Translation system. We conducted experiments for the EnglishFrench language pair on in-domain texts from Europarl (European Parliament Proceedings) and out-of-domain texts from Emea (European Medicines Agency Documents), and we compared the results of the Example-Based ...

OBJECTIVES The purpose of this nationwide intervention was to improve machine safety in small metal fabrication businesses (3 to 150 employees). The failure to implement machine safety programs related to guarding and lockout/tagout (LOTO) are frequent causes of Occupational Safety and Health Administration (OSHA) citations and may result in serious traumatic injury. METHODS Insurance safety ...

in this research, a braiding-pultrusion production line is developed by modifying a conventional textile maypole braiding machine and adding it to a pultrusion line for production of braid-pultruded (bp) composite rods. in order to evaluate the mechanical properties of bp rods and study the effect of some braiding parameters, they were produced with different braid roving linear densities and b...