Qualities. The Logo has been custom typeset and spaced and must never be redrawn. The Logo must not be modified or distorted by any means, including by dropshadows, bevels, 3-D effects, embosses, glows, or outlines. The Logo must not be placed over a “busy” or patterned background that may interfere with its presentation clarity.
Qualities. During the Service Offering term, the Service Offering will substantially conform to the applicable Service Description unless otherwise stated herein. We do not commit that the Service Offering will be uninterrupted or that the Service Offering does fulfill any merchantability or fitness for a particular purpose, or will meet (or is designed to meet) your business requirements, unless stated otherwise in the applicable Service Description or individually agreed between the parties.
Qualities. 2.1 Unless otherwise indicated to Buyer in writing by Seller, any information provided to Buyer regarding the characteristics of goods and services (in particular of Marine Fuels, shall not be construed as specifications to be delivered hereunder, but only as indications of the general characteristics available from time to time.
Qualities. When taking into account the qualities of the AI subsystem (vs the functional requirements mandated by goal and tooling), the following will be addressed: - Accuracy: The accuracy required by the solution will be determined by a few numbers, for example reaching goal within 20nm. The accuracy metrics need to be a guiding force towards the development in the solution even in training phase. During the training phase the accuracy is a guiding factor and not yet a must. In the TEM use case the important requirement is to always reach the goal within 20 nm of the goal and then stop when in that position. In the training phase this means that the training and the testing data need to reach roughly the same results. Not being able to reproduce training results for testing data results in a failure. In UUV Use Case 1, the objective is to automatically create new and challenging scenarios to fine-tune the parameters of unmanned vehicles. The novelty and criticality of these generated scenarios are assessed using various Key Performance Indicators (KPIs). Unlike the TEM use case, there is no predefined optimal goal in this scenario, and any enhancement in KPIs represents a favourable solution for the use case. Additionally, one can involve experts’ opinion to review the performance of the RL- solution. In UUV Use Case 2, since the objective is to improve the performance of vehicle sensors based on their observations in scenarios generated by UUV Use Case 1, we have access to ground truth data in the form of an image stream from the environment. The accuracy requirement may be quantified, for instance, by achieving a high percentage of correctly identified objects and minimizing discrepancies between the sensor input and the actual image stream. The information resulting from failures and successes determines quality standards for the DT, in the case of the DT requirements, it is hard to have them all upfront well-established and it is an iterative process. Of course, failing to produce qualitatively good data will result in failure of the AI, but the accuracy of the AI can be determined independently upfront, even when this dependency is known.
