A CRITICAL Sample Clauses

A CRITICAL. EVALUATION OF THE XXXX ET AL.’S (2003) INTER-ANNOTATOR AGREEMENT APPROACH There are lots of properties that make Xxxx et al’s (2003) approach re-usable. Firstly, their approach is supported with well-developed annotation guidelines which are based on a well-defined discourse theory. The annotation guidelines are brought to maturity by the training sessions. The training sessions not only advanced the annotation guidelines but also the annotators. Therefore, including training sessions in the annotation protocol is a double win. As another note worthy aspect, Xxxx et al. (2003) conducted all their annotations with at least two individual annotators in order to apply chance-corrected inter-annotator agreement measurement methodologies. They presented adjusted chance-corrected statistic (the kappa statistic) according to the needs of their theoretical framework. The kappa statistic is implemented in four different ways to meet the needs of the RST-based hierarchical discourse structure. Moreover, they explicitly presented the results and the corresponding methodologies with the negligible exception that they did not mention the exact kappa statistic used to measure the unit level agreement. Finally, they presented two sets of results, one of which is useful to monitor the effect of trainings, and the other is useful to monitor the repeatability of the annotations among different texts and annotators. Besides such important suggestions and implementations, Xxxx et al’s approach is not flawless. The problems can be classified into two groups: which are inherited from Marcu et al.’s (1999) study, and those which emerged in the study of Xxxx et al. (2003). The problems inherited from Xxxxx et al. (1999) are also discussed in the study of Xxxxx et al. (1999). All the problems raised by Xxxxx et al. (1999) will not be presented here, only the problems that will contribute to the ultimate goal of this thesis will be presented. The most noticeable problem is that the agreement levels (units, spans, nuclearity, relation, and group of relations) are independent from each other. With this approach, it is impossible to assign the source level of disagreement. For instance, wrong unit segmentation may lead to false nuclearity annotations, but may not affect the other levels’ annotation. On the other hand, in another wrong unit segmentation case, all levels’ annotation may be affected. Secondly, because of the hierarchical nature of the annotations, there exist lots of non...
Sample 1
AutoNDA by SimpleDocs

Related to A CRITICAL

  • Critical Milestones (a) Subject to the provisions of Section 3.1(c), commencing on the Effective Date, Seller shall develop the Facility in order to achieve the following milestones (“Critical Milestones”) on or before the date set forth in this Section 3.1(a):

  • Critical Illness Three (3) days per year, with pay, shall be granted in the case of a critical illness or accident to a member of the employee's immediate family as defined in Section 9.4.2. A statement by the physician verifying the need for the employee to be present with the immediate family member shall be attached to the absence form.

  • Origin Criteria For the goods that meet the origin criteria, the exporter should indicate in Box 8 of this Form, the origin criteria met, in the manner shown in the following table: Circumstances of production or manufacture in the country named in Box 11 of this form: Insert in Box 8

  • Service Level Expectations Without limiting any other requirements of the Agreement, the Service Provider shall meet or exceed the following standards, policies, and guidelines:

  • Benchmarks for Measuring Accessibility For the purposes of this Agreement, the accessibility of online content and functionality will be measured according to the W3C’s Web Content Accessibility Guidelines (WCAG) 2.0 Level AA and the Web Accessibility Initiative Accessible Rich Internet Applications Suite (WAI-ARIA) 1.0 for web content, which are incorporated by reference. Adherence to these accessible technology standards is one way to ensure compliance with the College’s underlying legal obligations to ensure that people with disabilities are able to acquire the same information, engage in the same interactions, and enjoy the same benefits and services within the same timeframe as their nondisabled peers, with substantially equivalent ease of use; that they are not excluded from participation in, denied the benefits of, or otherwise subjected to discrimination in any College programs, services, and activities delivered online, as required by Section 504 and the ADA and their implementing regulations; and that they receive effective communication of the College’s programs, services, and activities delivered online.

  • CORRECTIVE MEASURE The contractor shall repair any deficiencies in excess of the performance guideline.

  • Criminal History Category With regard to determining defendant’s criminal history points and criminal history category, based on the facts now known to the government, defendant’s criminal history points equal zero and defendant’s criminal history category is I.

  • Corrective Measures If the Participating Generator fails to meet or maintain the requirements set forth in this Agreement and/or the CAISO Tariff, the CAISO shall be permitted to take any of the measures, contained or referenced in the CAISO Tariff, which the CAISO deems to be necessary to correct the situation.

  • General Criteria a) Approved varieties shall be those varieties recommended by the Department, the Corporation, or those listed in the Atlantic Provinces Field Crop Guide, Publication 100A.

  • Evaluation Criteria 5.2.1. The responses will be evaluated based on the following: (edit evaluation criteria below as appropriate for your project)

Time is Money Join Law Insider Premium to draft better contracts faster.