restrict

About: restrict is a research topic. Over the lifetime, 284 publications have been published within this topic receiving 2699 citations.

Success and Failure: Static Concepts in a Dynamic Society.

Abstract: Since program participants are inclined to view evaluators as a threat to themselves and their program, and since such a perception can restrict the flow of information about a program to the evalu

EU regulations on algorithmic decision-making and a "right to explanation".

Abstract: We summarize the potential impact that the European Union's new General Data Protection Regulation will have on the routine use of machine learning algorithms. Slated to take effect as law across the EU in 2018, it will restrict automated individual decision-making (that is, algorithms that make decisions based on user-level predictors) which "significantly affect" users. The law will also effectively create a "right to explanation," whereby a user can ask for an explanation of an algorithmic decision that was made about them. We argue that while this law will pose large challenges for industry, it highlights opportunities for computer scientists to take the lead in designing algorithms and evaluation frameworks which avoid discrimination and enable explanation.

Code Completion with Neural Attention and Pointer Networks.

Abstract: Intelligent code completion has become an essential research task to accelerate modern software development. To facilitate effective code completion for dynamically-typed programming languages, we apply neural language models by learning from large codebases, and develop a tailored attention mechanism for code completion. However, standard neural language models even with attention mechanism cannot correctly predict the out-of-vocabulary (OoV) words that restrict the code completion performance. In this paper, inspired by the prevalence of locally repeated terms in program source code, and the recently proposed pointer copy mechanism, we propose a pointer mixture network for better predicting OoV words in code completion. Based on the context, the pointer mixture network learns to either generate a within-vocabulary word through an RNN component, or regenerate an OoV word from local context through a pointer component. Experiments on two benchmarked datasets demonstrate the effectiveness of our attention mechanism and pointer mixture network on the code completion task.

Innovation and Its Discontents

Abstract: In the last two decades, the role of patents in the U.S. innovation system has become increasingly troublesome, driven by two apparently mundane changes in patent law and policy. Economic analysis does not support abolishing patents, and even weakening the fundamental presumption of validity for appropriately issued patents would carry serious policy risks. The alternative is to make sure that, before they can be used to restrict the commercial activities of competitors, patents have gotten the appropriate scrutiny to ensure their validity. At the same time, we need to accept that the PTO will still make mistakes, and create a judicial system that deals with those mistakes in a balanced way. Doing this without an infeasible increase in resources for the Patent Office will require that the system be significantly modified. The modifications should be carefully tuned to create incentives so that private parties have the incentive and opportunity to bring information to bear, but have limited incentive and o...

Improving the Acquisition of Small Targets

Abstract: This paper describes the design and comparative evaluation of three methods that aid the acquisition of small targets. The first method, called ‘bubble targets’, increases the effective width of the target as the pointer approaches. The second method uses a form of’ stickiness’ to restrict movement as the pointer passes over an object. In the third method, called ‘goal-crossing’, the user simultaneously presses two mouse buttons before passing the pointer over the item. Goal-crossing overcomes the need for the user to decelerate the mouse when acquiring the target. Two evaluations were conducted, with the first (n = 37) based on the acquisition of abstract targets for Fitts’ Law modelling, and the second based on an ecologically oriented window resizing task (n = 11). Both showed that goal-crossing allowed the fastest target acquisition, but that it produced high error rates and was unpopular with participants. The ‘bubble’ and’ sticky’ techniques also allowed faster target acquisition than the traditional approach, and users were enthusiastic about them. Fitts’ Law accurately modelled all techniques. Implications of the results for general user interface design are briefly discussed.