How Goals Erase Framing Effects in Risky Decision Making

Evidence has shown that goals systematically change risk preferences in repeated decisions under risk. For instance, decision makers could aim to reach goals in a limited time, such as “making at least $1000 with ten stock investments within a year.” We test whether goal-based risky decisions differ when facing gains as compared to losses. More specifically, we examine the impact of outcome framing (gains vs. losses) and state framing (positive vs. negative resource states) on goal-based risky decisions. Our results (N=100) reveal no framing effects; instead, we find a consistently strong effect of the goal on risk preferences independent of framing. Computational modeling showed that a dynamic version of prospect theory, with a goal-dependent reference point, described 87% of participants best. This model treats outcomes as gains and losses depending on the state-goal distance. Our results show how goals can erase standard framing effects observed in risky choices without goals.

Novel Is Not Always Better: On the Relation between Novelty and Dominance Pruning

Novelty pruning is a planning technique that focuses on exploring states that are novel, i.e., those containing facts that have not been seen before. This seemingly simple idea has had a huge impact on the state of the art in planning though its effectiveness is not entirely understood yet.We relate novelty to dominance pruning, which compares states to previously seen states to eliminate those that are provably worse in terms of goal distance. Novelty can be interpreted as an unsafe approximation of dominance, where states containing novel facts are relevant because they enable new paths to the goal and, therefore, they are less likely to be dominated by others. This provides a framework to understand the success of novelty, resulting in new variants that combine both techniques.

Aiming for the Goal with SInE

The Sumo INference Engine (SInE) is a well-established premise selection algorithm for first-order theorem provers, routinely used, especially on large theory problems. The main idea of SInE is to start from the goal formula and to iteratively add other formulas to those already added that are related by sharing signature symbols. This implicitly defines a certain heuristical distance of the individual formulas and symbols from the goal.In this paper, we show how this distance can be successfully used for other purposes than just premise selection. In particular, biasing clause selection to postpone introduction of input clauses further from the goal helps to solve more problems. Moreover, a precedence which respects such goal distance of symbols gives rise to a goal sensitive simplification ordering. We implemented both ideas in the automatic theorem prover Vampire and present their experimental evaluation on the TPTP benchmark.

Navigational Control Analysis of Two-Wheeled Self-Balancing Robot in an Unknown Terrain Using Back-Propagation Neural Network Integrated Modified DAYANI Approach

SummaryThe present paper discusses on development and implementation of back-propagation neural network integrated modified DAYANI method for path control of a two-wheeled self-balancing robot in an obstacle cluttered environment. A five-layered back-propagation neural network has been instigated to find out the intensity of various weight factors considering seven navigational parameters as obtained from the modified DAYANI method. The intensity of weight factors is found out using the neural technique with input parameters such as number of visible intersecting obstacles along the goal direction, minimum visible front obstacle distances as obtained from the sensors, minimum left side obstacle distance within the visible left side range of the robot, average of left side obstacle distances, minimum right side obstacle distance within the visible right side range of the robot, average of right side obstacle distances and goal distance from the robot’s probable next position. Comparison between simulation and experimental exercises is carried out for verifying the robustness of the proposed controller. Also, the authenticity of the proposed controller is verified through a comparative analysis between the results obtained by other existing techniques with the current technique in an exactly similar test scenario and an enhancement of the results is witnessed.

Completeness-Preserving Dominance Techniques for Satisficing Planning

Dominance pruning methods have recently been introduced for optimal planning. They compare states based on their goal distance to prune those that can be proven to be worse than others. In this paper, we introduce dominance techniques for satisficing planning. We extend the definition of dominance, showing that being closer to the goal is not a prerequisite for dominance in the satisficing setting. We develop a new method to automatically find dominance relations in which a state dominates another if it has achieved more serializable sub-goals. We take advantage of dominance relations in different ways; while in optimal planning their usage focused on dominance pruning and action selection, we also use it to guide enforced hill-climbing search, resulting in a complete algorithm.

Infants prospectively control reaching based on the difficulty of future actions – To what extent can infants’ multiple step actions be explained by Fitts’ law?

Prospective motor control, a key element of action planning, is the ability to adjust one’s actions with respect to task demands and action goals in an anticipatory manner. The current study investigates whether 14-month-olds are able to prospectively control their reaching actions based on the difficulty of the subsequent action. We used a reach-to-place task, with difficulty of the placing action varied by goal size and goal distance. To target prospective motor control, we determined the kinematics of the prior reaching movements using a motion-tracking system. Peak velocity of the first movement unit of the reach served as indicator for prospective motor control. Both difficulty aspects (goal size and goal distance) affected prior reaching, suggesting that both these aspects of the subsequent action have an impact on the prior action. The smaller the goal size and the longer the distance to the goal, the slower infants were in the beginning of their reach towards the object. Additionally we modeled movement times of both reaching and placing actions using a formulation of Fitts’ law. The model was significant for placement and reaching movement times. These findings suggest that 14-month-olds are able to plan their future actions and prospectively control their related movements with respect to future task difficulties.

Reminding customers to be loyal: does message framing matter?

Purpose Marketers may proactively give customers personalized notices regarding their progress toward certain rewards as a means to stimulate ongoing behaviors. This paper aims to investigate the effect on customer repatronage intention by framed messages concerning either goal-distance or consequences of an action and it also seeks to identify important variables moderating those responses. Design/methodology/approach Five between-subjects experiments examined how participants’ repatronage intentions, in response to the framing of goal-distance (Study 1a) and consequences of an action (Study 2a), varied as a function of their level of progress toward goal completion and also tested if the framing effects might be attenuated when relationship benefit was high rather than low (Studies 1b and 2b). They further adopted perceived reciprocity as an underlying mechanism for examining the interplay between these two kinds of framing in stimulating ongoing behavior (Study 3). Findings Although messages which emphasized what individuals need to spend more to attain a reward (versus how short they are from earning a reward) or loss following inaction (versus gain following action) were likely to erode intention, such effects were confined to individuals with a moderate level of progress. This intention-eroding effect was further attenuated by attractive reward. The persuasive advantages of short-from-the-end framing of goal-distance over more-to-the-end counterparts were found to be diminished when paired with a loss-framed message concerning consequences of an action. Furthermore, the observed effects on intention were mediated by perceived reciprocity. Originality/value The studies add to the current understanding of how the way in which information is presented might enhance loyalty or fail to do so.

From Qualitative to Quantitative Dominance Pruning for Optimal Planning

Dominance relations compare states to determine whether one is at least as good as another in terms of their goal distance. We generalize these qualitative yes/no relations to functions that measure by how much a state is better than another. This allows us to distinguish cases where the state is strictly closer to the goal. Moreover, we may obtain a bound on the difference in goal distance between two states even if there is no qualitative dominance.We analyze the multiple advantages that quantitative dominance has, like discovering coarser dominance relations, or trading dominance by g-value. Moreover, quantitative dominance can also be used to prove that an action starts an optimal plan from a given state. We introduce a novel action selection pruning that uses this to prune any other successor. Results show that quantitative dominance pruning greatly reduces the search space, significantly increasing the planners' performance.

Effective Heuristics for Suboptimal Best-First Search

Suboptimal heuristic search algorithms such as weighted A* and greedy best-first search are widely used to solve problems for which guaranteed optimal solutions are too expensive to obtain. These algorithms crucially rely on a heuristic function to guide their search. However, most research on building heuristics addresses optimal solving. In this paper, we illustrate how established wisdom for constructing heuristics for optimal search can fail when considering suboptimal search. We consider the behavior of greedy best-first search in detail and we test several hypotheses for predicting when a heuristic will be effective for it. Our results suggest that a predictive characteristic is a heuristic's goal distance rank correlation (GDRC), a robust measure of whether it orders nodes according to distance to a goal. We demonstrate that GDRC can be used to automatically construct abstraction-based heuristics for greedy best-first search that are more effective than those built by methods oriented toward optimal search. These results reinforce the point that suboptimal search deserves sustained attention and specialized methods of its own.