Proxy (statistics)

About: Proxy (statistics) is a research topic. Over the lifetime, 5257 publications have been published within this topic receiving 94504 citations. The topic is also known as: proxy variable & proxy measurement.

Social comparison and task prediction: ability similarity and the use of a proxy.

Abstract: It is argued that an important aspect of the self-appraisal of ability involves a person's prediction of his or her own performance level on untried tasks through acquisition of knowledge about how other people have fared at the task. These others are treated as substitutes for a person's own participation in the task--as proxies. It is further argued that social comparison plays a significant role in both the selection of proxies and the impact of proxy information on task prediction. The first experiment, following earlier research on comparison preference in the context of performance prediction, showed that participants interested in performance prediction chose a proxy who had performed at about the same level as they had on a reference task, while those interested in improving their performances chose to see the performances of someone who had done far better than they on the reference task. The second study indicated that participants made more confident predictions for themselves from a similar proxy's task outcome than from that of a proxy whose ability was unknown to them. Participants predicted they would do worse than a more able proxy and better than a less able proxy, but were less confident in their predictions than were those with a similar proxy. Ratings of usefulness of the proxy followed the patterns of confidence ratings closely.

System and method for object persistence

Abstract: Embodiments of the invention are generally directed to a system and method for object persistence. In an embodiment, a proxy object is created based, at least in part, on object metadata stored in an internal data store of a persistence manager. The persistence manager may receive a request to access member data of a persistent object associated with the proxy object. In one embodiment, the persistence manager provides the member data from the proxy object.

Correlated color temperature: Use and limitations:

Abstract: Correlated color temperature (CCT) is a one-dimensional metric that aims to quantify the perceived visual quality of nominal white light sources. It is often used as a proxy for the color quality o...

Method of authorization by proxy within a computer network

Abstract: A method for enabling participants in an information technology (IT) system or a computer network to delegate user authority to other system participants is provided. The method of the present invention includes the generation of a proxy authorization. The proxy authorization, or proxy, is used by the IT system to insure that a given participant may have access to resources on the basis of a permission granted and intended by another user or agent of the IT system, and that the grantor of the permission is authorized to issue the access and/authorities as designated by or within the proxy authorization. A medical record repository, for example may allow unlimited access to particular individual patient records to an individual medical doctor. The doctor can then authorize a specific pharmacy to have limited access to designated portions of the medical records of certain of the patients to whom the doctor is authorized access. The pharmacy may then allow access to distinct and different subsets of the portions of the records, to which the pharmacy is authorized access to by a proxy issued by the doctor, to an insurance company, to a billing clerk, and to pharmacists. The use of proxies thereby allows for efficient B2B collaborative message processing using languages such as XML.

Meta-Gradient Reinforcement Learning

Abstract: The goal of reinforcement learning algorithms is to estimate and/or optimise the value function. However, unlike supervised learning, no teacher or oracle is available to provide the true value function. Instead, the majority of reinforcement learning algorithms estimate and/or optimise a proxy for the value function. This proxy is typically based on a sampled and bootstrapped approximation to the true value function, known as a return. The particular choice of return is one of the chief components determining the nature of the algorithm: the rate at which future rewards are discounted; when and how values should be bootstrapped; or even the nature of the rewards themselves. It is well-known that these decisions are crucial to the overall success of RL algorithms. We discuss a gradient-based meta-learning algorithm that is able to adapt the nature of the return, online, whilst interacting and learning from the environment. When applied to 57 games on the Atari 2600 environment over 200 million frames, our algorithm achieved a new state-of-the-art performance.