My research work is primarily motivated by these exciting trends in mobile networks. We believe WiFi has the unique potential to form a community based infrastructure that provides a truly scalable, efficient and ubiquitous... ]]> Tao Jin http://iris.lib.neu.edu/comp_sci_diss/33 http://iris.lib.neu.edu/comp_sci_diss/33 Thu, 09 May 2013 07:30:50 PDT The Holy Grail of parallel programming is to provide good speedup while hiding or avoiding the pitfalls of concurrency. But some level in the tower of abstraction must face facts: parallel processors execute code concurrently, and the interplay between concurrent code, synchronization, and the memory subsystem is a major determiner of performance. Effective parallel programming must ultimately be supported by scalable concurrent algorithms—algorithms that tolerate (or even embrace) concurrency for the sake of scaling with available parallelism. This dissertation makes several contributions to the understanding and expression of such algorithms:

• It shows how to understand scalable algorithms in terms... ]]> Aaron Joseph Turon http://iris.lib.neu.edu/comp_sci_diss/32 http://iris.lib.neu.edu/comp_sci_diss/32 Wed, 08 May 2013 13:22:29 PDT In the coming years, remote health monitoring is an area that is expected to grow significantly. Systems designed to follow-up with patients at home can be used not only to reduce visits to the doctor but also to augment the face-to-face interactions between patients and physicians. These systems could also provide much-needed care to the millions of people living in rural areas.

  While many researchers are investigating remote sensing technologies, the use of self-report in technological systems for long-term health monitoring remains a relatively understudied area. In this thesis, we investigate two main challenges in building systems designed for the... ]]> Laura Vardoulakis http://iris.lib.neu.edu/comp_sci_diss/31 http://iris.lib.neu.edu/comp_sci_diss/31 Tue, 16 Apr 2013 09:24:27 PDT Wireless communication networks span a wide range of applications such as traditional telephony, internet data access, and cyber-physical infrastructure monitoring. However, the quality of service is still lacking, and problems such as dropped calls, dead spots, and slow network speed are still very common. Those are due to fundamental constraints such as the scarcity of radio-frequency spectrum, signal propagation effects, e.g., channel fading and shadowing, and the small form factor of mobile devices with limited energy capacity and antenna diversity.

  In this thesis, we explore a new cooperative communication model. We investigate communication strategies that exploit the channel and traffic... ]]> Yin Wang http://iris.lib.neu.edu/comp_sci_diss/30 http://iris.lib.neu.edu/comp_sci_diss/30 Wed, 10 Apr 2013 08:56:38 PDT Diffusion processes are important models for many real-world phenomena, such as the spread of disease or rumors. We studied different aspects of diffusion processes in networks, focusing on designing efficient distributed algorithms for positive diffusion processes and good intervention strategies to control harmful diffusions.

  First, we design and analyze various distributed algorithms for diffusion processes. We want to devise efficient distributed algorithms, which are easy to implement, to help the spreading of positive/useful information. We refer to these processes as positive diffusions. Earlier work has studied this for a variety of models, mainly based on static networks. The major point... ]]> Zhifeng Sun http://iris.lib.neu.edu/comp_sci_diss/29 http://iris.lib.neu.edu/comp_sci_diss/29 Tue, 26 Mar 2013 08:25:15 PDT An Embodied Conversation Agent (ECA) is a computer interface designed to simulate human face-to-face conversation with its users, through the production of synthesized or prerecorded speech combined with an humanoid embodiment: a representation, either physical or visual, of a body capable of producing some portion of the nonverbal behaviors associated with speech, such as mouth movements, eye movements, head movements, hand gestures, facial expressions, and body posture. A key research problem in the design and implementation of realistic ECAs is generating the range of verbal and nonverbal behavior present in human conversation with appropriate frequency, timing, and quality.

  ECAs have... ]]> Daniel Jason Schulman http://iris.lib.neu.edu/comp_sci_diss/28 http://iris.lib.neu.edu/comp_sci_diss/28 Mon, 25 Mar 2013 13:47:51 PDT The clock speed of current CPUs and RAM has stopped scaling with Moore's Law. Yet the scale of applications in science and engineering continues to increase. In order to address this scaling of applications, newer NUMA architectures are emerging. These include parallel disks, hybrid CPU-GPU, and many-core CPUs.

  Existing CPU-based algorithms, as well as legacy sequential code, need to be migrated to NUMA in order to stay on the curve of Moore's Law.

  Migrating applications from traditional sequential architectures to NUMA is difficult, because NUMA architectures penalize data access patterns and computation patterns that are often used by programs on... ]]> Vlad Slavici http://iris.lib.neu.edu/comp_sci_diss/27 http://iris.lib.neu.edu/comp_sci_diss/27 Mon, 11 Mar 2013 09:29:32 PDT Reversible debuggers have existed since the early 1970s. However, they are not widely used, with the possible exception of GDB. GDB's target record is useful only when the cause of the bug is close in time to the bug manifestation. When the cause of the bug is far away from the manifestation, one resorts to a series of debugging sessions with the goal of narrowing down the cause of the bug. Thanks to reverse execution, it is possible to jump back and forth to any time of the execution.

  In this dissertation, we present a new reversible debugging platform, based... ]]> Ana-Maria Vişan http://iris.lib.neu.edu/comp_sci_diss/26 http://iris.lib.neu.edu/comp_sci_diss/26 Fri, 08 Mar 2013 12:28:34 PST In a higher-order language, the dominant control-flow mechanism is function call and return. Most higher-order flow analyses do not handle call and return well: they remember only a bounded number of pending calls because they approximate programs as finite-state machines. Call/return mismatch introduces precision-degrading spurious execution paths and increases the analysis time.

  We present flow analyses that provide unbounded call/return matching in a general setting: our analyses apply to typed and untyped languages, with first-class functions, side effects, tail calls and first-class control. This is made possible by several individual techniques. We generalize Sharir and Pnueli's summarization technique to handle... ]]> Dimitrios Vardoulakis http://iris.lib.neu.edu/comp_sci_diss/25 http://iris.lib.neu.edu/comp_sci_diss/25 Tue, 05 Mar 2013 08:25:49 PST Ranking documents in response to user queries is one of the fundamental problems in Information Retrieval. Learning to Rank has emerged as an effective approach for data-driven construction of ranking algorithms. Although many algorithms have been created, the effect of the properties of the training data through which such algorithms are developed has not been systematically studied. The creation of a learning dataset requires great effort since every document in the dataset has to be manually labeled with its degree of relevance. Motivated by the importance of efficiently creating quality datasets, I study 1) the effect of characteristics of the... ]]> Stefan Savev http://iris.lib.neu.edu/comp_sci_diss/24 http://iris.lib.neu.edu/comp_sci_diss/24 Wed, 20 Feb 2013 08:36:42 PST Despite their constant evolution over the last few decades, wireless communication networks still struggle with energy conservation. The problem manifests itself in many applications, in particular wireless sensor networks, where communications do not occur frequently and nodes often remain idle, and also small-scale communication networks, whose nodes need to be minuscule. Such applications can achieve optimal energy-efficiency using passive (battery-less) receivers that wirelessly receive energy and information at the same time. In this thesis, we present techniques to simultaneously deliver energy alongside information during wireless communications.

  First, we present mechanisms to consolidate energy and information transfer in wireless sensor networks.... ]]> Hooman Javaheri http://iris.lib.neu.edu/comp_sci_diss/23 http://iris.lib.neu.edu/comp_sci_diss/23 Thu, 07 Feb 2013 12:22:07 PST The ACL2 theorem prover combines a first-order dialect of LISP with an automated proof engine for first-order logic. While ACL2 is logically quite powerful, it can be difficult to build and maintain large models due to its ad hoc systems for modularity, namespace management, logical encapsulation, and macro expansion. I propose a new language, Refined ACL2, extending ACL2 with expressive component and macro systems designed to accommodate large-scale proof development and flexible logical abstractions. The component system of Refined ACL2 adapts many features of ML's functors and signatures to ACL2. Components support nesting, parameterization, translucent specification, and refinement of abstract... ]]> Carl Eastlund http://iris.lib.neu.edu/comp_sci_diss/22 http://iris.lib.neu.edu/comp_sci_diss/22 Mon, 04 Feb 2013 12:13:26 PST Contracts are a popular mechanism for enhancing the interface of components. In the world of first-order functions, programmers embrace contracts because they write them in a familiar language and easily understand them as a pair of a pre-condition andsame expressiveness to programmers but their meaning subtly differs from the familiar first-order notion. For instance, it is unclear what the behavior of dependent contracts for higher-order functions or of contracts for mutable data should be. As a consequence, it is difficult to design monitoring systems for such higher-order worlds.

  In response to this problem, this dissertation investigates complete monitors, a formal... ]]> Christos Dimoulas http://iris.lib.neu.edu/comp_sci_diss/21 http://iris.lib.neu.edu/comp_sci_diss/21 Fri, 25 Jan 2013 08:57:17 PST Flash storage has rapidly grown over the last few years into a redoubtable competitor to disk drives, due to its lower power consumption, lower latencies and lower cost per IOPs. Along with its attractive characteristics, flash also presents challenges, such as limited write endurance and the inability to overwrite a block in place. To address these issues, a software called the flash translation layer (FTL) maps logical blocks to physical locations on flash. Externally, the FTL presents a normal hard disk interface; internally, it implements address translation, garbage collection and wear-leveling algorithms which spread writes uniformly across the device.

  In... ]]> Simona Boboila http://iris.lib.neu.edu/comp_sci_diss/20 http://iris.lib.neu.edu/comp_sci_diss/20 Mon, 14 Jan 2013 09:40:25 PST Contracts allow programmers to specify the expected behavior and use of program components separately from the code of the components themselves. Since Bertrand Meyer introduced contracts to working programmers via the Eiffel programming language, Eiffel-like contract systems have been designed for many other object-oriented languages. Contract systems are not limited to object-oriented programming; Findler and Felleisen showed how to add contracts to languages with higher-order functions and formalized the notions of contract boundaries and blame.

  Currently, contract systems come with two major omissions: monitoring the invariants of mutable data structures and protecting first-class components, which are used in the construction... ]]> Tharpe Stephen Strickland http://iris.lib.neu.edu/comp_sci_diss/19 http://iris.lib.neu.edu/comp_sci_diss/19 Wed, 09 Jan 2013 12:15:48 PST The problem of building test collections is central to the development of information retrieval systems such as search engines. The primary use of test collections is the evaluation of IR systems. The widely employed "Cranfield paradigm" dictates that the information relevant to a topic be encoded at the level of documents, therefore requiring effectively complete document relevance assessments. As this is no longer practical for modern corpora, numerous problems arise, including scalability, reusability, and applicability.

  We propose a new method for relevance assessment based on relevant information, not relevant documents. Once the relevant information is collected, any document can be... ]]> Shahzad K. Rajput http://iris.lib.neu.edu/comp_sci_diss/18 http://iris.lib.neu.edu/comp_sci_diss/18 Wed, 12 Dec 2012 10:36:36 PST Cloud computing has ushered in a new paradigm with the availability of computing as a service, letting customers share the same physical infrastructure and purchase computing resources on demand (e.g. Amazon EC2 and Windows Azure). The multi-tenancy property of cloud computing offers clients flexibility while creating a unique set of challenges in areas such as reliability and security.

  In this thesis we study one challenge (SecureCloud) and three opportunities (DNSCloud, WebCloud and SamaritanCloud). In SecureCloud we explore how multi-tenancy, or the sharing of resources across users, can lead to the undermining of privacy and security. Taking Amazon EC2 as an... ]]> Fangfei Zhou http://iris.lib.neu.edu/comp_sci_diss/16 http://iris.lib.neu.edu/comp_sci_diss/16 Wed, 31 Oct 2012 14:10:31 PDT When a user submits a query to a search engine, the search engine computes a score for each document according to its relevance to the query, and ranks the documents based on their scores. Due to the complexity of the modern search engine, the score itself is not sufficient for the information retrieval application requiring combining different ranked lists. Inferring the score distributions for relevant and non-relevant documents and estimating the probability of relevance become imperative. In this thesis, we address two major research questions: (1) How to model score distributions in a more accurate manner for relevant and non-relevant... ]]> Keshi Dai http://iris.lib.neu.edu/comp_sci_diss/15 http://iris.lib.neu.edu/comp_sci_diss/15 Tue, 30 Oct 2012 07:01:53 PDT This dissertation uses volatility and spacing to allow one to quantify the dynamics of a wide class of ranked systems. The systems we consider are any set of items, each with an associated score that may change over time. We define volatility as the standard deviation of the score of an item. We define spacing as the distance in score from one item to its neighbor. From these two concepts we construct a model using stochastic differential equa- tions. We measure the model parameters in a variety of ranked systems and use the model to reproduce the salient features observed... ]]> Nicholas Blumm http://iris.lib.neu.edu/comp_sci_diss/14 http://iris.lib.neu.edu/comp_sci_diss/14 Tue, 07 Aug 2012 11:27:10 PDT Millions of diagrams are available online in scientific and technical documents. The knowledge contained in diagrams is a rich resource, but one that has been little exploited, in comparison to text. In this thesis we demonstrate that it is possible to efficiently build compact representations of diagram content. The representations can support a wide variety of diagram-based systems such as retrieval, classification, and the building of knowledge bases that integrate text and diagrams. We demonstrate the strengths of our approach through studies of diagram retrieval as well as supervised and unsupervised machine learning for classification. The techniques are applied to... ]]> Mingyan Shao Information storage and retrieval systems - Charts diagrams etc. Charts diagrams etc. - Classification