scholarly journals Explaining Reinforcement Learning to Mere Mortals: An Empirical Study

Author(s):  
Andrew Anderson ◽  
Jonathan Dodge ◽  
Amrita Sadarangani ◽  
Zoe Juozapaitis ◽  
Evan Newman ◽  
...  

We present a user study to investigate the impact of explanations on non-experts? understanding of reinforcement learning (RL) agents. We investigate both a common RL visualization, saliency maps (the focus of attention), and a more recent explanation type, reward-decomposition bars (predictions of future types of rewards). We designed a 124 participant, four-treatment experiment to compare participants? mental models of an RL agent in a simple Real-Time Strategy (RTS) game. Our results show that the combination of both saliency and reward bars were needed to achieve a statistically significant improvement in mental model score over the control. In addition, our qualitative analysis of the data reveals a number of effects for further study.

2018 ◽  
Vol 25 (2) ◽  
pp. 153-170 ◽  
Author(s):  
Jennifer A. Griffith ◽  
Carter Gibson ◽  
Kelsey Medeiros ◽  
Alexandra MacDougall ◽  
Jay Hardy ◽  
...  

Leaders have been classified as having charismatic, ideological, or pragmatic (CIP) leadership styles, each characterized by distinct patterns in cognition and interaction. Although each CIP style has been shown to facilitate certain aspects of the creative process for followers, questions remain regarding the impact of leadership style on overall follower creative performance. One factor likely to influence this relationship is leader distance, composed of the physical distance, perceived social distance, and perceived task interaction among leaders and followers. Past research has also emphasized the role of leaders’ mental models as they relate to follower performance. Less understood, however, is how the mental models of followers may affect this process. Using the CIP model of leadership, this study explores leader distance and leader–follower mental model congruence on follower creative performance. Results indicated that while leadership style does not directly influence follower creativity, it interacts with leader distance to shape creative outcomes. Results further indicated that while general mental model congruence is not predictive, alignment on specific mental model dimensions contributes to enhanced creative performance among followers. Implications and future research directions are discussed.


Geosciences ◽  
2019 ◽  
Vol 9 (5) ◽  
pp. 238 ◽  
Author(s):  
Diego Corrochano ◽  
Alejandro Gómez-Gonçalves

This paper analyzes the impact of fieldwork on the development of students’ mental models concerning glaciers and their effects on the landscape. Data were collected by means of an open-ended questionnaire that was administered to 279 pre-service teachers before and after an educational field trip, which analyzed its impact on short-term and long-term outcomes. In general, students’ mental models about how glaciers function and how they create landforms are relatively simplistic and incomplete. Students are unaware of the major erosional properties associated with glaciers and many of them do not specify that glaciers are bodies of ice that have a tendency to move down slope. The analysis of the data yielded four mental model categories. Fieldwork influenced the short-term effects on mental model development even though its positive impact decreases over time. Mental models including scientific views were only found in the post-instruction group. On the other hand, the pre-instruction group was strongly influenced by a catastrophic event that occurred in the region in 1959 (the Ribadelago flooding), which interferes with students’ mental reasoning on the formation of landscape features. This way of thinking is reinforced and/or mixed with a religious myth (Villaverde de Lucerna legend), which also invokes a catastrophic origin of the lake. In this case, this includes mystic flooding.


2019 ◽  
Vol 118 (11) ◽  
pp. 303-312
Author(s):  
Jamal Asad Mezel ◽  
Adnan Fadhil Khaleel ◽  
Kiran Das Naik Eslavath

This empirical study show that the impact of all styles was well moderate. The means of effect of all styles were less than 3 out of 5. It means the expected impact of transformational affect upon the all dimensions of the activities, are not expected due to the traditional styles of leadership and the lack of information about the transformational leadership styles which can guide leaders to use such styles in the organization which may be this results due to lack of trained leaders and necessary knowledge with the leaders in all universities about transformational styles the traditional form of the leadership styles which used by the university leaders affect the communication between all levels of the administration and the faculty members which has consequence because decrease in motivation and a self-consideration from the administration.


2019 ◽  
Author(s):  
Jennifer R Sadler ◽  
Grace Elisabeth Shearrer ◽  
Nichollette Acosta ◽  
Kyle Stanley Burger

BACKGROUND: Dietary restraint represents an individual’s intent to limit their food intake and has been associated with impaired passive food reinforcement learning. However, the impact of dietary restraint on an active, response dependent learning is poorly understood. In this study, we tested the relationship between dietary restraint and food reinforcement learning using an active, instrumental conditioning task. METHODS: A sample of ninety adults completed a response-dependent instrumental conditioning task with reward and punishment using sweet and bitter tastes. Brain response via functional MRI was measured during the task. Participants also completed anthropometric measures, reward/motivation related questionnaires, and a working memory task. Dietary restraint was assessed via the Dutch Restrained Eating Scale. RESULTS: Two groups were selected from the sample: high restraint (n=29, score >2.5) and low restraint (n=30; score <1.85). High restraint was associated with significantly higher BMI (p=0.003) and lower N-back accuracy (p=0.045). The high restraint group also was marginally better at the instrumental conditioning task (p=0.066, r=0.37). High restraint was also associated with significantly greater brain response in the intracalcarine cortex (MNI: 15, -69, 12; k=35, pfwe< 0.05) to bitter taste, compared to neutral taste.CONCLUSIONS: High restraint was associated with improved performance on an instrumental task testing how individuals learn from reward and punishment. This may be mediated by greater brain response in the primary visual cortex, which has been associated with mental representation. Results suggest that dietary restraint does not impair response-dependent reinforcement learning.


Sign in / Sign up

Export Citation Format

Share Document