Search

In interactive e-learning environments such as Intelligent Tutoring Systems, pedagogical decisions can be made at different levels of granularity. In this work, we focus on making decisions at "two levels": whole problems vs. single steps and explore three types of granularity: "problem-level only" ("Prob-Only"), "step-level only" ("Step-Only") and "both problem and step levels" ("Both"). More specifically, for Prob-Only, our pedagogical agency decides whether the next problem should be a worked example (WE) or a problem-solving (PS). In WEs, students observe how the tutor solves a problem while in PSs students solve the problem themselves. For Step-Only, the agent decides whether to elicit the student's next solution step or to tell the step directly. Here the student and the tutor "co-construct" the solution and we refer to this type of task as collaborative problem-solving (CPS). For Both, the agency first decides whether the next problem should be a WE, a PS, or a CPS and based on the problem-level decision, the agent then makes step-level decisions on whether to elicit or tell each step. In a series of classroom studies, we compare the three types of granularity under random yet reasonable pedagogical decisions. Results showed that while Prob-Only may be less effective for High students, Step-Only may be less effective for Low ones, Both can be effective for both High and Low students. Motivated by these findings, we propose and apply an offline, off-policy Gaussian Processes based "Hierarchical Reinforcement Learning" ("HRL") framework to induce a "hierarchical pedagogical policy" that makes adaptive, effective decisions at both the problem and step levels. In an empirical classroom study, our results showed that the HRL policy is significantly more effective than a Deep Q-Network (DQN) induced step-level policy and a random "yet reasonable" step-level baseline policy.

In this work, we investigate how two factors, metacognitiveskills and motivation, would impact student learning acrossdomains. More specifically, our primary goal is to identify thecritical, yet robust, interaction patterns of these two factors thatwould contribute to students’ performance in learning logicfirst and then their performance on a subsequent new domain,probability. We are concerned with two types of metacognitiveskills: strategy-awareness and time-awareness, that is, whichproblem-solving strategy to use and when to use it. Our datawere collected from 495 participants across three consecutivesemesters, and our results show that the only students who con-sistently outperform their peers across both domains are thosewho are not only highly motivated but also strategy-aware andtime-aware.

We explored the impact of three types of decision granularity,problem level (Prob), step level (Step), and both problem andstep levels (Both), on student learning. We first conducted anempirical study to directly compare the three conditions andthen three subsequent studies to evaluate one or two of thethree conditions. Overall our empirical results showed therewas no significant difference among the three conditions. Wefurther split students into different groups based on their per-formances on the single-principle and the multiple-principleproblems in the pre-test. Solving the single-principle problemsonly involves one step while solving the multiple-principleones involves generating multiple steps in a logic order. Wedefine High students as those who were correct on all single-principle problems and at least one multiple-principle ones inthe pre-test, Low students as those who were correct on someor all single-principle problems but no multiple-principle ones,and the rest are in the Medium group. Our empirical resultsshowed that for Low students, Both can be better than Step.For the Medium and High students, no clear conclusions couldbe drawn because of small sample sizes. As a result, in apost-hoc analysis all students were combined by their assignedconditions. Overall, while no significant difference was foundamong the three conditions, we found that the impact of threetypes of granularity, Prob, Step, and Both differs significantlyfor High vs. Low students: Both, Step > Prob for the Highstudents and Both, Prob > Step for the Low students. No clearconclusions could be drawn for the Medium group due to itssmall sample sizes. In short, while Prob could be effective forLow students but ineffective for High ones and Step could beeffective for High students but ineffective for Low ones, Bothseemed to be effective for both High and Low students.

In this study, we explored the impact of the decision agency(Student vs. Tutor) and granularity (Problem vs. Step) acrossstudents with different levels of incoming competence (Highvs. Low). Students were randomly assigned to four conditionsand split into High and Low groups based on their pre-testscores. All students used the same Intelligent Tutoring Sys-tem (ITS) called Pyrenees, followed the same general proce-dure, studied the same training materials, and worked throughthe same training problems. The only substantive differencesamong the four conditions were decision agency and granular-ity. That is: who decided to present an example or to solvea problem: the student or the ITS tutor; and was the deci-sion made problem-by-problem or step-by-step? Our overallresults showed that there were significant different impacts ofthe decision agency and granularity between High and Lowstudents on learning performance. More specifically, for Highstudents granularity was the more dominant factor in that steplevel decisions can be more effective than problem level deci-sions regardless of the decision agency; for Low students therewas a significant interaction effect in that: Low students ben-efit significantly more when they were making problem-leveldecisions than making step-level decisions, but no significantdifference was not found when the decisions were made bythe tutor. Much to our surprise, both High and Low groupsshowed strong decision-making preference for problem solv-ing over worked example at both problem and step levels.

In this study we explored the impact of student versus tutorpedagogical decision-making on learning. More specifically,we examined what would happen if we let students decidehow to handle the next task: to view it as a worked exam-ple or to solve it as a problem solving. We examined this im-pact at two levels of task granularity: problem vs. step. This2 × 2 study was conducted on an existing Intelligent TutoringSystem (ITS) called Pyrenees. 279 students were randomlyassigned to four conditions and the domain content and re-quired steps were strictly controlled to be equivalent acrossfour conditions: all students used the same system, followedthe same general procedure, studied the same training materi-als, and worked through the same training problems. The onlysubstantive differences among the four conditions were deci-sion agency {Student vs. Tutor} and granularity {Problem vs.Step}. That is: who decided to present an example or to solvea problem; and was the decision made problem-by-problem orstep-by-step? Our results showed that there was a significantinteraction effect between decision agency and granularity onstudent learning and a significant main effect of granularity ontime on training. That is, step level decisions can be more ef-fective than problem level decisions but the students were morelikely to make effective pedagogical decisions at problem levelthan step level. In general, on both problem and step levels, thestudents were significantly more likely to decide to do problemsolving rather than study it as a worked example.

[Objectives] The fault of centrifugal pump unit would influence the missions' accomplishment and the improvement of operational readiness. In order to solve the problem of detection and location of common faults of centrifugal pump units,a fault diagnosis method of antifriction bearing of centrifugal pump is presented.[Methods] the vibration signal of the antifriction bearing was decomposed adaptively with the method of local characteristic-scale decomposition(LCD). The random forest was used for fault diagnose based on the sample entropy generated by the intrinsic scale component(ISC)extracting. Finally, combined with the fault diagnosis test,the correctness of the above method was verified based on the monitoring data of bearing fault injected into the centrifugal pump unit.[Results] The results show that the diagnose method based on LCD and random forest can effectively diagnose the fault mode of the antifriction bearing of centrifugal pump.[Conclusions] The study in this paper can lay a foundation for the improvement of equipment fault diagnosis capability, and provide technical support for the establishment of centrifugal pump fault prediction and health management system.

In this paper, we explore the impact of two types of instructionalinterventions, worked examples and problem solving, attwo levels of granularity: problems and steps. This study drewon an existing Intelligent Tutoring System (ITS) for Probabilitycalled Pyrenees and involved 266 students who were randomlyassigned to five conditions. All students experiencedthe same procedure, studied the same training problems in thesame order, and used the same ITS. The conditions differedonly in how the training problems were presented. Our resultsshow that when the domain content and required steps arestrictly equivalent, different granularities of pedagogical decisionscan significantly impact students’ time on task. Morespecifically, the fine-grained step level decisions can have astronger pedagogical impact than the problem-level ones.

Background: In this study, we aimed to evaluate hemodynamic influence of the dissected aortic system via various ex vivo type B aortic dissection (AD) models. Methods: Twenty-four raw porcine aortas were harvested and randomly divided into 4 groups to create various aortic models. Model A was the control group, while models B to D indicated the AD group, where models B and C presented a proximal primary entry with the false lumen (FL) lengths of 15 and 20 cm, respectively, and model D presented a 20 cm FL with a proximal primary entry and a distal reentry. All the aortic models were connected to a mock circulation loop to attain the realistic flow and pressure status. The flow distribution rate (FDR) of the aortic branches was calculated. Doppler ultrasound was applied to visualize the AD structure and to attain the velocity of flow in both the true and false lumens. Several sections of the AD were stained with hematoxylin and eosin for histologic evaluation after the experiment. Results: This study demonstrated that higher pressures were found for the AD group compared with the control group. The mean systolic pressures at the inlet of models A to D were 113.34±0.81, 120.58±0.52, 117.76±0.82, and 115.87±0.42 mm Hg, respectively. The FDRs of the celiac artery in models A to D were 8.65%, 8.32%±0.15%, 7.87%±0.13%, and 8.03%±0.21%, respectively. By ultrasound visualization, the velocity of the flow at the entry to the FL in the AD group ranged in 10 to 92 cm/s. The dissection flap presented pulsatile movement, especially in the models B and C which contained 1 primary entry without distal reentries. Histological examinations indicated that AD was located between the intimal and medial layers. Conclusions: Our ex vivo models demonstrated that the configuration of the dissected aorta influenced the pressure distribution. Moreover, the dissection flap affected the FDR of the aortic branches that possibly inducing malperfusion syndrome. [ABSTRACT FROM AUTHOR]