No description available.

Many tasks can easily be posed as the problem of responding to the states of an external world with actions that maximise the reward received over time. Algorithms that reliably solve such problems exist. However, their worst-case complexities are typically more than proportional to the size of the state space in which a task is to be performed. Many simple tasks involve enormous numbers of states, which can make the application of such algorithms impractical. This thesis examines reinforcement learning algorithms which effectively learn to perform tasks by constructing mappings from states to suitable actions. In problems involving large numbers of states, these algorithms usually must construct approximate, rather than exact, solutions and the primary issue examined in the thesis is the way in which the complexity of constructing adequate approximations scales as the size of a state space increases. The vast majority of reinforcement learning algorithms operate by constructing estimates of the long-term value of states and using these estimates to select actions. The potential effects of errors in such estimates are examined and shown to be severe. Empirical results are presented which suggest that minor errors are likely to result in significant losses in many problems, and where such losses are most likely to occur. The complexity of constructing estimates accurate enough to prevent significant losses is also examined empirically and shown to be substantial.

The primary objective of this thesis is to evolve a method for assessing and developing features of a learning organization . To fulfill this, I approached the thesis by examining several research questions and using multiple research methodologies. The research questions were not all established at the outset. Rather, they evolved as features of a journey down a road less traveled. With this journey came the decision to write the thesis in the first person. The first research question was Q1: What will bridge the divide between organizational learning and the learning organization? By reviewing the extant literature on organizational learning and the learning organization, I developed a theoretical framework that linked these two streams. The framework suggests that the extent of divide between the two streams is determined by the extent of learning transfer. The learning transfer is affected by the learning barriers operating at the levels of learning (i.e., individuals, groups, and organizational). This led me to my second research question Q2: What are these barriers to learning transfer and how do they impact the levels of learning in the organization? I cumulated the dispersed literature on learning barriers, and synthesized the learning barriers into five key dimensions: Intrapersonal, relational, cultural, structural, and societal. I then used the Delphi technique on 17 individuals to investigate the impact of the learning barriers on the levels of learning. This generated two additional research questions. The third research question was Q3: How do individuals initiate a double-loop change? This deals with the little researched area of initiation of double-loop change whilst engaging with the interfaces at the levels of learning. I used multiple case studies to examine this question and found that individuals transit through four distinct stages when initiating double-loop change: 'embedded', 'embedded discomfited', 'scripted', and 'unscripted'. Once double-loop learning has been initiated at the individual level, it is important that it is transferred across the organization. Therefore, my fourth research question was Q4: How does a new shared understanding for a double-loop change develop across the organization? I did an in-depth, single case based investigation of an organization. Using Identity and Complexity theory perspectives, I tracked the evolving new shared understanding through four phases: de-identification phase, situated re-identification phase, transition phase, and identification with core ideology phase. The key insights from examining these research questions, particularly insights from examining Q3 and Q4, enabled me to suggest nine key organizational interventions necessary to overcome the learning barriers and develop a learning organization: Identifying, developing, and dispersing double-loop mastery; Enabling constructive contradictions; Creating a superordinate organizational identity; Building emotional intelligence (in individuals and groups); Ambidextrous leadership; Strategic support for experimentation; Promoting 'systems doing'; Accessibility of valid information; Institutionalizing scanning across industry boundaries. When these nine organizational interventions are implemented, they produce five new learning organization orientations: genetic diversity, organizational ideology, organizational dualism, organizational coupling, and strategic play. These five new learning organizational orientations provide the archetypes of the learning organization. I then developed an instrument to assess these five new orientations, and did a preliminary testing of the instrument. While aspects of my work overlaid with previous knowledge, new advances in knowledge were established by: Postulating a link between the streams of organizational learning and learning organization Synthesizing learning barriers into the five key dimensions, and investigating their impact on the levels of learning Understanding the stages of double-loop learning initiation by an individual, whilst engaging with the interfaces at the levels of learning Understanding the process of a new shared understanding evolving Postulating five new orientations of the learning organization

This thesis examined the processes underlying implicit learning. Six candidate components thought to be involved in implicit learning were identified: task demands, affordances of experimental materials, subjects????? intentions, content of knowledge, form of representation and awareness. A series of 15 experiments are reported that investigated the learning in the invariant digit, invariant clocks, and the artificial grammar task in an attempt to determine the role of these six components. The empirical findings were interpreted as supporting the view that learning under implicit task conditions: i) is interactive and not passive, ii) does not result in the acquisition of a default representation of knowledge, but that any structure remains latent within an experience until appropriate retrieval cues are provided at test, and iii) does not proceed ?????outside????? awareness but that the lack of insight that often accompanies learning is due to poor understanding about the basis of performance. A framework was proposed that demonstrated how this pattern of objective performance and subjective experience arises, through the interaction of the candidate components. This framework is more consistent with a process-oriented view (e.g. Whittlesea & Wright, 1997) than a separate systems account of implicit learning (e.g. Reber, 1989).

Zugl.: Münster (Westfalen), Universiẗat, Diss.

This study examined the teaching experiences of six elementary student interns from Oregon State University, Corvallis, OR during fall and winter quarters, 1991. The primary intent of the study was to examine critically the roles and effects of mind styles as applied in actual classroom instruction and learning, in terms of feasibility and student outcomes. The research focus was to examine the roles and effects of mind styles in an elementary educational setting by addressing the following questions: (1) Do elementary school students exhibit mind styles that are dominant or less dominant and which vary among individuals, as has been found in adults; (2) do mind styles exist among elementary students, (3) are mind styles determinable among elementary students; (4) can a researcher-developed self-assessment instrument for elementary students be used to accurately identify their dominant mind styles? The research methodology was qualitative. Specifically, the nature of the invisible driving forces that shaped students behavior within the classroom setting were observed. Data was gathered through journals and lesson plans from the student interns, video tapes of teaching experiences, task analyses, field notes, and maintenance of a diary by the researcher. The Self-Assessment Children's Instrument was administered to each elementary student who took part in the study, involving them in self-ranking themselves in relation to a variety of qualities. The research methodology and instrument were field-tested with satisfactory results. After testing 129 students and completing 43 observations on each student in six classrooms, the following was concluded from the results of the research: 1) Some learners are generalists who enjoy understanding the big picture before focusing on specifics. They want a context in which to put new ideas, and they are good at seeing relationships as they learn. 2) Some learners display an active, hands-on, exploratory pattern. These students learn physically and through concrete experiences and activities. 3) In every classroom there are students for whom personal relationships are important. These learners enjoy helping each other and working in groups. 4) Other students are structured and systematic in their approach. These students want rules for new materials to be presented clearly and logically with examples that build from the simple to the complex. / Graduation date: 1994

Networks that solve specific visual tasks, such as the evaluation of spatial relations with hyperacuity precision, can be eastily synthesized from a small set of examples. This may have significant implications for the interpretation of many psychophysical results in terms of neuronal models.

INCORPORATING LITERATURE CIRCLES IN A THIRD-GRADE CLASSROOM Margaret Marlene Morrow, Ed.D. University of Pittsburgh, 2005 This study describes how one third-grade classroom became literature discussants utilizing role-playing in a literature circle format. Over a span of nine months, I documented students initiation, training, and practice utilizing role-playing to conduct peer-led discussions about books. Types of responses, role-playing, and interactions resulting from role-playing activities while engaged in literature circle were examined. Audio and video taping, student self-evaluations and reactions, student role-playing schedules, and student artifacts based on culminating activities, provided a data set to analyze interactions and responses. Findings of the study indicated roles used to promote individual participation in discussions about books were easily adapted, particularly for the low and average-ability students who require more support in their learning. Over time and texts, students appeared to shift naturally in and out of roles while making contributions to discussions. Interactions known as Informative ranked highest in frequency for all students followed by Facilitative and Solicitation. Acknowledging interactions ranked third in frequency for the low-ability student. Responses known as Personal were most frequent and progressive across time and texts for all students. Interpretive responses showed the greatest percentage increase compared to other categories, but the lowest in cumulative frequency. Evaluative responses were most frequent and consistent across time and texts for the high-ability student compared to low and average ability students. Descriptive responses were lowest cumulative tally for high and low students only. The role of the teacher was viewed as shifting from voice of authority to facilitator assisting youngsters in finding their own voices.

It was the purpose of this study to explore if there is a more effective way to practice aural skills, one that could be individualized as well as computer based: a method of practice that would enhance what the instructor does in the classroom and help the student reinforce their recognition of the basic elements of aural skills and dictation. Concept mapping, which has been beneficial in other educational settings, especially math and science, might be such a way. The goal of this study was to examine if this procedure could be applied to aural stimuli in the same way it has been applied to verbal information. Sixty-four students in a first year college aural skills class were randomly assigned to two groups. Following a pretest of twenty-five chords randomly chosen from a pool of fifty, one group was presented with a lesson on concept mapping while listening to the chords and the other group received a distractor task. The results of these tests showed a higher gain score for the treatment group, M = 5.33, than for the control group, M = 1.25. An Analysis of Covariance (ANCOVA), showed a moderately significant gain of p = .030 (one-tailed) for the treatment group. Overall, the test results indicated a positive effect of the treatment on the ability of the participants to recognize chords in the posttest. Due to the limitations in the present study, more research is suggested to determine the effectiveness of the process.

This study investigated the effects of repeated reading of decodable and less decodable text on oral reading fluency, specifically accuracy and correct words read per minute. In addition the study sought to determine if a particular text type facilitated oral reading fluency for good, average, or poor readers. Participants read either decodable or less decodable texts or received regular classroom instruction. The two treatment groups reread two narrative and one informational text using four different techniques, (e.g. echo, choral and two variations of partner reading). The comparison group read a different leveled text each week as part of their regular classroom instruction. The leveled text was not reread. Treatment and comparison groups did not differ on post measures on nonsense word fluency or passage reading. Both treatment groups regardless of ability level increased significantly in words read correctly for each story. Thus, the repeated reading of text did have an effect on correct words read per minute. There was no significant difference between the groups reading decodable and less decodable text for two of the three stories read, regardless of reading ability. There was a significant difference between the two groups of low ability readers on one story, with the students reading decodable text making a significant gain in oral reading fluency.