Matching Items (996)
Description
Reinforcement learning (RL) is a powerful methodology for teaching autonomous agents complex behaviors and skills. A critical component in most RL algorithms is the reward function -- a mathematical function that provides numerical estimates for desirable and undesirable states. Typically, the reward function must be hand-designed by a human expert and, as a result, the scope of a robot's autonomy and ability to safely explore and learn in new and unforeseen environments is constrained by the specifics of the designed reward function. In this thesis, I design and implement a stateful collision anticipation model with powerful predictive capability based upon my research of sequential data modeling and modern recurrent neural networks. I also develop deep reinforcement learning methods whose rewards are generated by self-supervised training and intrinsic signals. The main objective is to work towards the development of resilient robots that can learn to anticipate and avoid damaging interactions by combining visual and proprioceptive cues from internal sensors. The introduced solutions are inspired by pain pathways in humans and animals, because such pathways are known to guide decision-making processes and promote self-preservation. A new "robot dodge ball' benchmark is introduced in order to test the validity of the developed algorithms in dynamic environments.
ContributorsRichardson, Trevor W (Author) / Ben Amor, Heni (Thesis advisor) / Yang, Yezhou (Committee member) / Srivastava, Siddharth (Committee member) / Arizona State University (Publisher)
Created2018
Description
In this thesis, a new approach to learning-based planning is presented where critical regions of an environment with low probability measure are learned from a given set of motion plans. Critical regions are learned using convolutional neural networks (CNN) to improve sampling processes for motion planning (MP).
In addition to an identification network, a new sampling-based motion planner, Learn and Link, is introduced. This planner leverages critical regions to overcome the limitations of uniform sampling while still maintaining guarantees of correctness inherent to sampling-based algorithms. Learn and Link is evaluated against planners from the Open Motion Planning Library (OMPL) on an extensive suite of challenging navigation planning problems. This work shows that critical areas of an environment are learnable, and can be used by Learn and Link to solve MP problems with far less planning time than existing sampling-based planners.
In addition to an identification network, a new sampling-based motion planner, Learn and Link, is introduced. This planner leverages critical regions to overcome the limitations of uniform sampling while still maintaining guarantees of correctness inherent to sampling-based algorithms. Learn and Link is evaluated against planners from the Open Motion Planning Library (OMPL) on an extensive suite of challenging navigation planning problems. This work shows that critical areas of an environment are learnable, and can be used by Learn and Link to solve MP problems with far less planning time than existing sampling-based planners.
ContributorsMolina, Daniel, M.S (Author) / Srivastava, Siddharth (Thesis advisor) / Li, Baoxin (Committee member) / Zhang, Yu (Committee member) / Arizona State University (Publisher)
Created2019
Description
Knowledge Representation (KR) is one of the prominent approaches to Artificial Intelligence (AI) that is concerned with representing knowledge in a form that computer systems can utilize to solve complex problems. Answer Set Programming (ASP), based on the stable model semantics, is a widely-used KR framework that facilitates elegant and efficient representations for many problem domains that require complex reasoning.
However, while ASP is effective on deterministic problem domains, it is not suitable for applications involving quantitative uncertainty, for example, those that require probabilistic reasoning. Furthermore, it is hard to utilize information that can be statistically induced from data with ASP problem modeling.
This dissertation presents the language LP^MLN, which is a probabilistic extension of the stable model semantics with the concept of weighted rules, inspired by Markov Logic. An LP^MLN program defines a probability distribution over "soft" stable models, which may not satisfy all rules, but the more rules with the bigger weights they satisfy, the bigger their probabilities. LP^MLN takes advantage of both ASP and Markov Logic in a single framework, allowing representation of problems that require both logical and probabilistic reasoning in an intuitive and elaboration tolerant way.
This dissertation establishes formal relations between LP^MLN and several other formalisms, discusses inference and weight learning algorithms under LP^MLN, and presents systems implementing the algorithms. LP^MLN systems can be used to compute other languages translatable into LP^MLN.
The advantage of LP^MLN for probabilistic reasoning is illustrated by a probabilistic extension of the action language BC+, called pBC+, defined as a high-level notation of LP^MLN for describing transition systems. Various probabilistic reasoning about transition systems, especially probabilistic diagnosis, can be modeled in pBC+ and computed using LP^MLN systems. pBC+ is further extended with the notion of utility, through a decision-theoretic extension of LP^MLN, and related with Markov Decision Process (MDP) in terms of policy optimization problems. pBC+ can be used to represent (PO)MDP in a succinct and elaboration tolerant way, which enables planning with (PO)MDP algorithms in action domains whose description requires rich KR constructs, such as recursive definitions and indirect effects of actions.
However, while ASP is effective on deterministic problem domains, it is not suitable for applications involving quantitative uncertainty, for example, those that require probabilistic reasoning. Furthermore, it is hard to utilize information that can be statistically induced from data with ASP problem modeling.
This dissertation presents the language LP^MLN, which is a probabilistic extension of the stable model semantics with the concept of weighted rules, inspired by Markov Logic. An LP^MLN program defines a probability distribution over "soft" stable models, which may not satisfy all rules, but the more rules with the bigger weights they satisfy, the bigger their probabilities. LP^MLN takes advantage of both ASP and Markov Logic in a single framework, allowing representation of problems that require both logical and probabilistic reasoning in an intuitive and elaboration tolerant way.
This dissertation establishes formal relations between LP^MLN and several other formalisms, discusses inference and weight learning algorithms under LP^MLN, and presents systems implementing the algorithms. LP^MLN systems can be used to compute other languages translatable into LP^MLN.
The advantage of LP^MLN for probabilistic reasoning is illustrated by a probabilistic extension of the action language BC+, called pBC+, defined as a high-level notation of LP^MLN for describing transition systems. Various probabilistic reasoning about transition systems, especially probabilistic diagnosis, can be modeled in pBC+ and computed using LP^MLN systems. pBC+ is further extended with the notion of utility, through a decision-theoretic extension of LP^MLN, and related with Markov Decision Process (MDP) in terms of policy optimization problems. pBC+ can be used to represent (PO)MDP in a succinct and elaboration tolerant way, which enables planning with (PO)MDP algorithms in action domains whose description requires rich KR constructs, such as recursive definitions and indirect effects of actions.
ContributorsWang, Yi (Author) / Lee, Joohyung (Thesis advisor) / Baral, Chitta (Committee member) / Kambhampati, Subbarao (Committee member) / Natarajan, Sriraam (Committee member) / Srivastava, Siddharth (Committee member) / Arizona State University (Publisher)
Created2019
Description
Recent advancements in external memory based neural networks have shown promise
in solving tasks that require precise storage and retrieval of past information. Re-
searchers have applied these models to a wide range of tasks that have algorithmic
properties but have not applied these models to real-world robotic tasks. In this
thesis, we present memory-augmented neural networks that synthesize robot navigation policies which a) encode long-term temporal dependencies b) make decisions in
partially observed environments and c) quantify the uncertainty inherent in the task.
We extract information about the temporal structure of a task via imitation learning
from human demonstration and evaluate the performance of the models on control
policies for a robot navigation task. Experiments are performed in partially observed
environments in both simulation and the real world
in solving tasks that require precise storage and retrieval of past information. Re-
searchers have applied these models to a wide range of tasks that have algorithmic
properties but have not applied these models to real-world robotic tasks. In this
thesis, we present memory-augmented neural networks that synthesize robot navigation policies which a) encode long-term temporal dependencies b) make decisions in
partially observed environments and c) quantify the uncertainty inherent in the task.
We extract information about the temporal structure of a task via imitation learning
from human demonstration and evaluate the performance of the models on control
policies for a robot navigation task. Experiments are performed in partially observed
environments in both simulation and the real world
ContributorsSrivatsav, Nambi (Author) / Ben Amor, Hani (Thesis advisor) / Srivastava, Siddharth (Committee member) / Tong, Hanghang (Committee member) / Arizona State University (Publisher)
Created2018
Description
With the new independence of adulthood, college students are a group susceptible to adopting unsupported, if not harmful, health practices. A survey of Arizona State University undergraduate students (N=200) was conducted to evaluate supplement use, trust in information sources, and beliefs about supplement regulation. Of those who reported using supplements, college students most frequently received information from friends and family. STEM majors in fields unrelated to health who were taking a supplement were found to be less likely to receive information about the supplement from a medical practitioner than those in health fields or those in non-STEM majors (-26.9%, p=0.018). STEM majors in health-related fields were 15.0% more likely to treat colds and/or cold symptoms with research-supported methods identified from reliable sources, while non-health STEM and non-STEM majors were more likely to take unsupported cold treatments (p=0.010). Surveyed students, regardless of major, also stated they would trust a medical practitioner for supplement advice above other sources (88.0%), and the majority expressed a belief that dietary supplements are approved/regulated by the government (59.8%).
ContributorsPerez, Jacob Tanner (Author) / Hendrickson, Kirstin (Thesis director) / Lefler, Scott (Committee member) / College of Liberal Arts and Sciences (Contributor) / School of Molecular Sciences (Contributor) / Department of Physics (Contributor) / Barrett, The Honors College (Contributor)
Created2018-05
ContributorsChandler, N. Kayla (Author) / Neisewander, Janet (Thesis director) / Sanabria, Federico (Committee member) / Olive, M. Foster (Committee member) / Barrett, The Honors College (Contributor) / College of Liberal Arts and Sciences (Contributor)
Created2013-05
Description
I propose that norms regulate behaviors that negatively impact an individual's survival and reproduction. But because monitoring and enforcing of norms can be costly, individuals should be selective about which norms they police and under what circumstances they should do so. Two studies tested this idea by experimentally activating fitness-relevant motives and having participants answer questions about the policing of norms. The first study examined a norm prescribing respect for status and another proscribing sexual coercion. Results from Study 1 failed to support the hypotheses; activating a status-seeking motive did not have the predicted effects on policing of the respect-status norm nor did activating a mating motive have the predicted effects on policing of the respect-status norm or anti-coercion norm. Study 2 examined two new norms, one prescribing that people stay home when sick and the other proscribing people from having sex with another person's partners. Study 2 also manipulated whether self or others were the target of the policing. Study 2 failed to provide support; a disease avoidance motive failed to have effects on policing of the stay home when sick norm. Individuals in a relationship under a mating motive wanted less policing of others for violation of the mate poaching norm than those in a baseline condition, opposite of the predicted effects.
ContributorsSmith, M. Kristopher (Author) / Neuberg, L. Steven (Thesis director) / Presson, Clark (Committee member) / Hruschka, J. Daniel (Committee member) / Barrett, The Honors College (Contributor) / College of Liberal Arts and Sciences (Contributor)
Created2013-05
Description
Literature in public administration emphasizes a growing dissatisfaction with government on the part of residents. Where there tends to be a lack in the literature is in terms of solutions to this problem. We would like to argue that the engagement process itself has the power to foster a profound attitudinal shift on the part of both residents and government. This paper explores the structural and cultural barriers to satisfactory public engagement both from literature and a combination of policy analysis, semi-structured interviews and participatory observation within the City of Tempe. We then provide recommendations to the City of Tempe on how to overcome these barriers and effect authentic public engagement practices. With these new suggested practices and mindsets, we provide a way that people can have the power to create their own community.
ContributorsRiffle, Morgan (Co-author) / Tchida, Celina (Co-author) / Ingram-Waters, Mary (Thesis director) / Grzanka, Patrick (Committee member) / King, Cheryl (Committee member) / Barrett, The Honors College (Contributor) / College of Liberal Arts and Sciences (Contributor)
Created2013-05
Description
This thesis examines the relationship between unofficial, official, and parallel Islam in Uzbekistan following the end of the Soviet Union. Key touchstone moments in Uzbekistan during the twentieth-century show the history between unofficial and official Islam and the resulting precedents set for Muslims gathering against the government. This historical analysis shows how President Karimov and the Uzbek government view and approach Islam in the country following independence.
ContributorsTieslink, Evan (Author) / Batalden, Stephen (Thesis director) / Kefeli, Agnes (Committee member) / Saikia, Yasmin (Committee member) / Barrett, The Honors College (Contributor) / College of Liberal Arts and Sciences (Contributor) / School of Politics and Global Studies (Contributor) / School of Historical, Philosophical and Religious Studies (Contributor)
Created2013-05
Description
Through this creative project, I executed a Distracted Driving Awareness Campaign at Arizona State University to raise awareness about the dangers of distracted driving, specifically texting while driving. As an Undergraduate Student Government Senator, my priority is the safety and success of students, both in and out of the classroom. By partnering with State Farm and AT&T, we were able to raise awareness about the dangers of distracted driving and collected over 200 pledges from students to never text and drive.
ContributorsHibbs, Jordan Ashley (Author) / Miller, Clark (Thesis director) / Parmentier, Mary Jane (Committee member) / Barrett, The Honors College (Contributor) / College of Liberal Arts and Sciences (Contributor) / School of Politics and Global Studies (Contributor) / Department of Psychology (Contributor) / Graduate College (Contributor)
Created2014-05