No description available.

Thesis (M.S. in Operations Research)--Naval Postgraduate School, March 2010. / Thesis Advisor: Alderson, David L. Second Reader: Church, Richard L. "March 2010." Author(s) subject terms: Evacuation, Network, Flow Optimization, Space-Time network. Includes bibliographical references (p. 41-42). Also available in print.

No description available.

No description available.

The current study investigates the effects of preparing learners for an online debate through a worked example in terms of student perception, participation, level of cognitive skill, and electronic interaction patterns. There has been a change in the focus of distance learning research from comparative media studies to the means to improve the quality of distance education. One of the key elements in this changing impetus are strategies to promote interaction such as the introduction of structure or scaffolding argumentation (McIsaac & Blocher, 1998). One such strategy is the online debate in which students are organized into teams to take a position on an issue and argue on its behalf (Jeong, 2004). The debate is constrained through the addition of rules and specific message headers. Well-designed quality interaction holds the potential to create more satisfied learners and higher quality learning outcomes (Muirhead, 2002), but in the absence of quality, interaction has been found to actually lead to a decrease in satisfaction, participation, and performance (Joung & Keller, 2004; Kreijns, Kirschner, & Jochems, 2002). One way to ensure quality within interactive exercises such as online debate is to prepare students through a worked example. A worked example models an expert’s work and demonstrates desired behaviors for the learner to study (Atkinson, Derry, Renkl, & Wortham, 2000). Students were randomly assigned to teams to participate in an online debate with half being given access to a worked example before participating. In order to examine the effects of the worked example on students’ perceived satisfaction and level of preparedness, a survey was administered at various points throughout the semester. Additionally, debate transcripts were analyzed for participation, cognitive skill, and interaction patterns. The results demonstrate that students prepared through a worked example participated more frequently, wrote more words or phrases that encouraged the participation of others, and used higher-order thinking skills. The conclusion was that worked examples can be used to model behaviors for students to emulate. The implication being that instructors should consider providing worked examples before engaging students in online debate and future research should examine the efficacy of a worked example in preparing learners for other types of interactive activities.

付記する学位プログラム名: デザイン学大学院連携プログラム / 京都大学 / 新制・課程博士 / 博士(情報学) / 甲第24033号 / 情博第789号 / 新制||情||134(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 黒橋 禎夫, 教授 河原 達也, 教授 楠見 孝 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

We discuss a formulation for active example selection for function learning problems. This formulation is obtained by adapting Fedorov's optimal experiment design to the learning problem. We specifically show how to analytically derive example selection algorithms for certain well defined function classes. We then explore the behavior and sample complexity of such active learning algorithms. Finally, we view object detection as a special case of function learning and show how our formulation reduces to a useful heuristic to choose examples to reduce the generalization error.

Syftet med studien var att undersöka priming med exempels påverkan på skapelsers kreativitet. Effekterna av priming skulle testas i formen av egenskapspriming såväl som konceptuell priming och mätas utifrån både en användbarhetsaspekt och en originalitetsaspekt. Ett delsyfte var också att undersöka sambandet mellan intresset för att ta fram en skapelse och en skapelses grad av originalitet samt användbarhet. Primingeffekterna prövades genom ett experiment där 36 studenter delades in i 3 grupper som fick i uppgift att ta fram IT-lösningar för kollektivtrafiken. 2 av grupperna primades med var sitt exempel på hög originalitet respektive hög användbarhet. Effekterna av egenskapspriming undersöktes sedan genom en jämförelse mellan grupperna över hur stor andel av de framtagna lösningarna som innehöll egenskaper från exemplen. För att undersöka effekterna av konceptuell primingen gjordes en jämförelse av originaliteten och användbarhet hos dessa gruppers lösningar jämfört med de framtagna av en kontrollgrupp utan tillgång till exempel. Lösningarnas originalitet och användbarhet bedömdes av en expertgrupp bestående av forskare inom kreativitet och branschaktiva inom området. Denna expertgrupp bedömde även sitt intresse för att ta fram de olika lösningarna. Denna bedömning tillsammans med de två tidigare användes för att undersöka sambandet mellan intresset för en lösning och lösningens originalitet respektive användbarhet. Resultatet av studien visade på ett signifikant positivt samband hos båda grupperna av bedömare mellan en lösnings användbarhet och intresset för att ta fram lösningen. Ett signifikant positivt samband gällande en lösnings originalitet och intresset för denne kunde dock endast uppvisas hos forskare medan de branschaktiva visade på ett osignifikant negativt samband mellan de två. Resultatet av egenskapspriming och konceptuell priming visade inte överlag på några signifikanta skillnader mellan grupperna. Studien anses vara av vikt då den visar på hur en ökning av kreativiteten på skapelser inom en organisation inte behöver betyda en ökning av organisationens kreativitet p.g.a. ointresset i att omfamna och satsa på dessa lösningar. Vidare visar studien på hur primingeffekter är svåra att skapa i situationer som i högre grad ska efterlikna de i verkliga livet och hur vidare forskning behövs på området för att kartlägga variabler som påverkar priming och primings begränsningar. / The aim of the study was to examine priming with example’s impact on ideas creativity, regarding usability and novelty both as a result of targeting priming and conceptual priming. A subsidiary aim was to examine the relationship between an idea’s degree of originality and usefulness and the interest in pursuing the idea. The effects of priming were tried in an experiment involving 36 students who were asked to develop IT solutions for public transport. The students were randomly divided into two experimental groups and one control group. The members of the experimental group were primed with an example with high originality respectively usability. The solutions were given two indexes on how many of the properties of respective examples were found in the solutions. Their degree of originality and usefulness were also assessed by a group of experts. This group also assessed their interest in the respective solution. The study revealed a significant positive correlation between a solution's usability and interest of the solution. A positive significant correlation could also be seen between a solution's originality and the interest of the solution, but only in the group of analysts consisting of scientists. Among the industry-experts the results instead pointed toward a negative relationship between originality and interest, although a non-significant relationship. There were no significant differences between the group’s solutions regarding their degree of originality and usefulness nor the frequency in which features from the example was found in the solutions. The study is of importance because it shows how increases in the creativity of solutions produced by individuals and employees not necessarily lead to an increase in the creativity of an organization, due to lack of interest in embracing and invest in these creative solutions. The study also reflects the difficulties in transferring priming effects discovered in a lab environment to situations that more closely resembles those in real life. Further research is necessary to broaden the knowledge regarding variables that affects priming and the limits between which priming exist.

Humans often learn to manipulate objects by observing other people. In much the same way, robots can use imitation learning to pick up useful skills. A system is detailed here for using imitation learning to teach a robot to grasp objects using both hand and whole-body grasps, which use the arms and torso as well as hands. Demonstration grasp trajectories are created by teleoperating a simulated robot to pick up simulated objects. When presented with a new object, the system compares it against the objects in a stored database to pick a demonstrated grasp used on a similar object. Both objects are modeled as a combination of primitives—boxes, cylinders, and spheres—and by considering the new object to be a transformed version of the demonstration object, contact points are mapped from one object to the other. The best kinematically feasible grasp candidate is chosen with the aid of a grasp quality metric. To test the success of the chosen grasp, a full, collision-free grasp trajectory is found and an attempt is made to execute in the simulation. The implemented system successfully picks up 92 out of 100 randomly generated test objects in simulation. / Singapore-MIT Alliance (SMA)

The use of Machine Learning for malware detection is essential to counter the massive growth in malware types compared with the traditional signature-based detection system. However, machine learning models could also be extremely vulnerable and sensible to transferable adversarial example (AE) attacks. The transfer AE attack does not require extra information from the victim model such as gradient information. Researchers explore mainly 2 lines of transfer-based adversarial example attacks: ensemble models and ensemble samples. \\ Although comprehensive innovations and progress have been achieved in transfer AE attacks, few works have investigated how these techniques perform in malware data. Besides, generating adversarial examples on an android APK file is not as easy and convenient as it is on image data since the generated AE of malware should also remain its functionality and executability after perturbation. Therefore, it is urgent to validate whether previous methodologies could still have their effect on malware considering the differences compared to image data. \\ In this thesis, we first have a thorough literature review for the AE attacks on malware data and general transfer AE attacks. Then we design our algorithm for the transfer AE attack. We formulate the optimization problem based on the intuition that the contribution evenness of features towards the final prediction result is highly correlated to the AE transferability. We then solve the optimization problem by gradient descent and evaluate it through extensive experiments. Implementing and experimenting with the state-of-the-art AE algorithms and transferability enhancement techniques, we analyze and summarize the weaknesses and strengths of each method. / Master of Science / Machine learning models have been widely applied to malware detection systems in recent years due to the massive growth in malware types. However, these models are vulnerable to adversarial attacks. Malicious attackers can add some small imperceptible perturbations to the original testing samples and mislead the classification results at a very low cost. Research on adversarial attacks would help us gain a better understanding of the attacker's side and inspire defenses against them. Among all adversarial attacks, the transfer-based adversarial example attack is one of the most devastating attacks since it does not require extra information from the targeted victim model such as gradient information or query from the model. Although plenty of researchers has explored the transfer AE attack lately, few works focus on malware (e.g., Android) data. Compared with image data, perturbing malware is more complicated and challenging since the generated adversarial examples of malware need to remain functionality and executability. To validate how transfer AE attack methods perform on malware, we implement the state-of-the-art (SOTA) works in this thesis and experiment with them on real Android data. Besides, we develop a new transfer-based AE attack method based on the contribution of each feature for generating AE. We then do comprehensive evaluations and draw comparisons between SOTA works and our proposed method.