Allan Christensen

Time-based reward shaping in real-time strategy games

Real-Time Strategy (RTS) is a challenging domain for AI, since it involves not only a large state space, but also dynamic actions that agents execute concurrently. This problem cannot be optimally solved through general Q-learning techniques, so we propose a solution using a Semi Markov Decision Process (SMDP). We present a time-based reward shaping technique, TRS, to speed up the learning process in reinforcement learning. Especially, we show that our technique preserves the solution optimality for some SMDP problems. We evaluate the performance of our method in the Spring game Balanced Annihilation, and provide some benchmarks showing the performance of our approach.

Predictive Model for Group Selection Performance on Touch Devices

Users spend hours making selections with ineffective tools, we therefore examine selection methods for efficiency in various touch trials. In a preliminary study three alternative selection methods were identified, we compared these to a smart selection tool. The study showed that a single selection method was the fastest; however, when the amount of targets increased, a multiple selection tool became more efficient. A secondary study with more targets revealed similar results. We therefore examined the single selection method against traditional selection methods in a user study. The results reveal a model of the average action and time cost for all methods within the parameters of mental preparation and target addition. The study revealed that the most favored selection methods were a lasso and brush selection tool. The study provides evidence towards a predictive model of selection performance for multiple target selection trials.

Social integration i arbejdslivet: belyst gennem en undersøgelse inden for voksenhandicapsektoren

I organisationen og dens arbejdsmiljo er der altid udstodte personer eller grupper, som af den ene eller anden grund ikke er integreret i virksomheden. Dette forhold kan fremtraede i en social proces pa arbejdspladsen, hvor enkelte ansatte bliver mobbet, chikaneret eller bliver gjort til syndebuk af faellesskabet i organisationen. Eksempelvis udpeger medlemmer i sociale grupper syndebukke, nar noget gar galt, nar der opstar konflikter, eller hvis der er en person ansat, som er uonsket i gruppen. Personer, som er placeret yderst i gruppen, udvaelges ofte til at vaere syndebukke (Lenneer-Axelson & Thylefors 1993, 76). I arbejdslivets sociale processer dannes og udvikles grupper, som udgor en »slags kollektiv forsvarsmekanisme rettet mod et mere eller mindre uskyldigt gruppemedlem« (Einarsen & Mikkelsen 2000, 18). Efter vor opfattelse findes nogle bagvedliggende arsagssammenhaenge til sociale processer, som fremtraeder som udstodning eller mangelfuld social integration af ansatte i enhver organisation. Disse sammenhaenge finder vi i arbejdsmiljoog kulturforholdene i den enkelte organisation, og det er disse forhold vi vil beskaeftige os med i denne artikel. I en kombineret kulturog arbejdsmiljoundersogelse fik vi afdaekket et tilsyneladende velfungerende arbejdsmiljo (Nielsen & Christensen 2001). Men pa tvaers af personalegrupper identificerede vi ansatte, som ikke blev integreret socialt i institutionen. Det er vor antagelse, at hovedarsagen la bade i et ikke-intakt socialt netvaerk og i en organisationskultur. Dvs. et netvaerk som ikke er tilstraekkeligt udbygget til at fastholde og integrere ansatte, som har det svaert med kollegaer, ledere og beboere, og en

A Comparison of Gamified HCI Studies with Lab and Crowd Participants

We compared a game-based experiment carried out in the lab with crowdsourced set ups (informed and uninformed participants) on the device’s human resolution (DHR) the minimum size for dragging the finger onto a target on a touch screen. Lab participants produced fewer errors than the crowd. From lab participants, we found the smallest selectable target width for dragging onto non-occluded targets with visual target position feedback, was between 2mm and 4mm on mobile touch devices. Performance data on error not time allowed for drawing this conclusion as participants from all groups did not take enough care and time to acquire the targets. The bi-modal performance distributions of crowd participants required filtering data. Received on 18 November 2016; accepted on 04 February 2017; published on 05 April 2017

Transition Times for Manipulation Tasks in Hybrid Interfaces

Previous work has shown that uninstrumented in-air interaction is slower and not as precise for pointing tasks compared to the mouse. Yet, there are interaction scenarios where it is preferable or advantageous to use in-air input. Therefore, we examine a three-device hybrid setup involving the mouse, keyboard, and a Leap Motion. We performed a user study to quantify the costs associated with transitioning between these interaction devices and also for performing simple 2D manipulation tasks using the mouse and Leap Motion. We discovered that transitioning to and from the Leap Motion takes on average 0.87 seconds longer than transitions between the mouse and keyboard.