466 research outputs found
The case for tradable remedies in WTO dispute settlement
No full textIn response to concerns over the efficacy of the WTO dispute settlement system, especially in regard to its use by developing countries, Mexico has tabled a proposal to introduce tradable remedies within the Dispute Settlement Understanding. The idea is that a country that has won cause before the WTO, and who is facing non-implementation by the author of the illegal act but feels that its own capacity to exercise its right to impose countermeasures is unlikely to lead to compliance, can auction off that right. The attractiveness of this idea is that it offers an additional possibility to injured WTO members to get something from the dispute settlement mechanism without putting into question the legal nature of the existing contract, that is, the predominantly decentralized system of enforcement in the WTO. Examining all disputes brought to the WTO since its inception, the authors find some support for Mexico's perception that developing countries face a practical problem when they attempt to carry through with effective retaliation within the WTO system. And based on the formal results of Bagwell, Mavroidis, and Staiger (2003), they describe arguments that lend some support to the efficacy of Mexico's proposed solution from the perspective of formal economic theory.Information Technology,International Terrorism&Counterterrorism,General Technology,Economic Theory&Research,Environmental Economics&Policies,Trade and Services,Economic Theory&Research,International Terrorism&Counterterrorism,Information Technology,World Trade Organization
Message passing neural networks for molecular property prediction
No full textThis electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Thesis: M. Eng. in Computer Science and Engineering, Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2019Cataloged from student-submitted PDF version of thesis.Includes bibliographical references (pages 81-84).Developing new drugs relies heavily on understanding the various molecular properties of potential drug candidates. While experimental assays performed in the lab are the best source of information about molecular properties, these assays are slow and expensive. For this reason, there has been great interest in the potential of machine learning models to predict molecular properties without the need for experimental assays. However, recent literature has not yet clearly determined which machine learning models are optimal for molecular property prediction. In this thesis, I apply the Direct Message Passing Neural Network (D-MPNN) from [47, 48] to 19 publicly available property prediction datasets, and I demonstrate that it consistently outperforms prior machine learning models. Additionally, I introduce several optimizations to the D-MPNN which further enhance its performance and lead to new state-of-the-art results.by Kyle Swanson.M. Eng. in Computer Science and EngineeringM.Eng.inComputerScienceandEngineering Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Scienc
Sightfulness
No full textThis electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.Thesis: S.M., Massachusetts Institute of Technology, Department of Architecture, 2019Cataloged from PDF version of thesis.Includes bibliographical references.Architecture has historically been situated within the world of mixed reality. Brunelleschi's Baptistery in Florence introduced illusory space inserting depth into a two-dimensional plane in the fifteenth century. Centuries later, the Bauhaus challenged the hegemony of perspectival space in favor of axonometric projection, subverting the perceived experience of architecture for its formal attributes. Today our perception of the world is mediated by social media, photo manipulation software, and screen resolution. These technologies conspire not for a fidelity to reality but to promote social and political agendas-truth is not truth. Conversely, rapid improvements in camera technology attempts an objective unification of what qualifies as "clear" and "crisp" photography to document the world while counter-cultural technologies, including hallucinogenic drugs, seek clarification of a different kind. Sightfulness is a guided trip through media phase changes. It fluidly moves between photo realistic representations of common elements in space to impossible infinite multiplications, inversions, distortions and moments of pure abstraction. It leverages advances in computational software that allow a blurring between the virtual and the real. It disorients the viewer through unpredictable transitions, destabilizing their conception of the world that they are inhabiting.by Kyle Branchesi.S.M.S.M. Massachusetts Institute of Technology, Department of Architectur
Electrodynamic cavitation with nanoscale control
No full textThe design and implementation of a new technique using an atomic force micro-scope (AFM) has allowed new and interesting studies to be performed with nanoscale spatial resolution. A new form of cavitation known as electrodynamic cavitation (EDC) was characterized and used to generate large structures on the surface of Si wafers. Oxide layers grown on Si/SiO2/H 2O and Si/SiO2/Au/H2O interfaces reached maximum heights of 130 nm and 690 nm, respectively. These structures represent a full order of magnitude increase in height over oxides grown in air under similar voltages and time durations. EDC was also used to perform in situ sharpening of a conducting AFM tip in high purity water with decreases in the radius of curvature of the AFM tips by 40% to 100%
Repeated application of a novel creatine cream improves muscular peak and average power in male subjects
No full textUsing a multicenter, randomized controlled trial, (N = 123, age 23 [+ or -] 4 years) we sought to determine whether administration of a novel, topical creatine supplement could improve muscular performance after acute and repeated (7-day) exposure. To study the acute performance enhancing effects of the supplement, subjects completed 5 sets of 15 maximal concentric single-leg knee extensions with and without the application of a low- (low dose [LD]-3.5 ml) or high-dose (high dose [HD]-7 ml) topical creatine cream. After a wash-out period, subjects had one leg randomized to receive either the creatine or placebo cream, with further randomization into an oral creatine or placebo supplement group. Subjects completed 5 sets of 15 maximal concentric single leg knee extensions before and after the supplementation protocol. After acute application, no significant differences in peak power (LD: 252 [+ or -] 93 W, HD: 261 [+ or -] 100 W, p = 0.21), average power (LD: 172 [+ or -] 65 W, HD: 177 [+ or -] 69 W, p = 0.78), or fatigue index (LD: 13.4 [+ or -] 10.6%, HD: 14 [+ or -] 11.9%, p = 0.79) were observed between experimental and placebo creams (peak power: LD: 244 [+ or -] 76 W, HD: 267 [+ or -] 109 W; average power: LD: 168 [+ or -] 57 W, HD: 177 [+ or -] 67 W; fatigue index: LD: 12.4 [+ or -] 9.6%, HD: 12.8 [+ or -] 10.6%) or when controlling for sex. After the 7-day supplementation protocol, a significant increase in average power (creatine: 203 [+ or -] 61-220 [+ or -] 65 W, placebo: 224 [+ or -] 61-214 [+ or -] 61 W) and peak power (creatine: 264 [+ or -] 73-281 [+ or -] 80 W, placebo: 286 [+ or -] 79-271 [+ or -] 73 W) in the leg receiving creatine cream was observed in male subjects. No differences were observed in female subjects. The topical creatine cream did not enhance measures of muscle performance after acute application, but was able to improve peak and average power in male subjects after 7 consecutive days of application
Effects of Lop and Scatter Slash Treatment on Potential Fire Behavior and Soil Erosion Following a Selection Harvest in a Coast Redwood Forest
No full textEven though harvesting timber is one method of reducing fuel continuity and subsequent potential fire behavior, the residual slash can greatly increase the surface fuel loading and subsequent risk of wildfire on harvested sites. Fire behavior following silvicultural treatments to a stand can vary greatly, with both depth and loading playing a significant role (Nives 1989). Surface fuels and subsequent potential fire behavior has been shown to increase in the first year after harvest in coast redwood forests (Sequoia sempervirens (D. Don) Endl.) (Dicus 2003), which threatens not only natural resources but also structures in an ever-increasing wildland-urban interface. Alternatively, logging slash may serve to hold post-harvest soil in place, thereby decreasing soil erosion and stream sedimentation, which may be of greater importance than fire risk in some areas. Logging operations, particularly roads, can significantly increase erosion rates in coastal forests (Amaranthus et al., 1985). Slash may reduce the detrimental effects of harvesting impacts by decreasing raindrop impact and subsequent soil movement. This research examined surface fuel loading, potential fire behavior, and soil erosion following a single-tree selection harvest and subsequent lop and scatter slash treatment in coast redwood stands near Aptos, California. The specific objectives were to (1) quantify fuel loading and potential fire behavior before and for three years after a selective harvest, and (2) determine if residual slash affected surface soil erosion
Feasibility study for Dechlorination of mixed plastic waste streams
No full textThis report explored the feasibility of dechlorination of mixed plastic waste systems using a low-temperature pyrolysis process. Decomposition of plastics into chlorides is of interest due to the damage these can pose to both the process system, products, and the environment. Therefore, dechlorination of mixed plastics waste streams is essential to remove detrimental chlorides before entering the processing stage. Relevant literature presents that chloride is found in PVC in plastic mixtures, which decomposes two stages; a dechlorination step below 330 0C and then a thermal decomposition of the hydrocarbon chain. This is due to the weaker C-Cl binding energy present in the molecule. Hydrochloride gas is produced from PVC as a result of dechlorination at temperatures below 300 0C. DKR 350 is a plastic mixture that is studied in this report, and this plastic waste stream can contain approximately 0.2-5% PVC. The process of dechlorination of plastic wastes is investigated in this report. As a result of understanding the parameters required to remove chlorides from a mixed plastic waste stream, an experimental setup was created to examine the viability of the process. The results were analysed and compared with relevant literature on PVC degradation to assess its efficiency. Overall, the methodology established by this report was unsuccessful in producing a practical result. However, the viability of the dechlorination process in mixed plastics wastes has been shown as a result of the 84.9% conversion of chlorides in the plastic from the initial mixture to other sources. As a result of this positive indication of the viability of the process, further testing was recommended based on lessons learned from the current report. Further, an improved test setup was suggested based on several recommendations explored as a result of findings from the experimental.Mechanical Engineering | Process and Energy Technolog
Post-Soviet ethnic politics and public goods provision
No full textWhat explains the pattern of public goods distribution across ethnic groups in the states of the former Soviet Union? In this dissertation, I seek to demonstrate how the unique pattern of nation-state formation in the former Soviet states interacts with other institutional legacies in a manner that differentiates it from other regions. Rather than the logic of “ethnic diversity deficit” applied in most analyses of the theme, I explain post-Soviet public goods provision through a logic of ethnic domination and its relationship to other salient features of sociopolitical organization, including informal social networks and ethnodemographic configurations. The unintended institutionalization of an enduring ethnic titular / non-titular binary combines with the Soviet legacy of informal social networks of access that are endemic throughout the region and structure state-society relations. Thus, I propose that Soviet institutional legacies determine both the supply and demand sides of public goods and service provision: ethnic titular political domination ensures preferential targeting to titular coethnics, while the continued significance of informal networks of access disproportionately allows elite and non-elite titulars to demand state resources successfully.
I demonstrate the effects of this relationship in three empirical chapters. Analyzing large-N data from Kyrgyzstan, I show that ethnic Kyrgyz titulars are not only more positive than non-titulars in their evaluations of public goods provision in general, but also that more extensive integration into informal social networks exacerbates this intergroup distributive differentiation. My next chapter presents qualitative data collected during fieldwork in Kyrgyzstan and illustrates the causal mechanisms and grounded understandings of the previous chapter’s findings. The third empirical chapter presents an analysis of large-N data from eight post-Soviet countries in which I explore the extent to which individual coethnicity and coregionality with national leaders is predictive of public goods outcomes. As suggested by much existing research, the findings are dependent on the outcome one studies. Coethnicity and coregionality with national leaders is related to education outcomes in an additive fashion. However, there is no relationship coethnicity or coregionality and first year child survival. I find a more complex interactive relationship between the two explanatory variables and child immunization.Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2021-08-01The student, Kyle Estes, accepted the attached license on 2019-07-08 at 10:06.The student, Kyle Estes, submitted this Dissertation for approval on 2019-07-08 at 10:12.This Dissertation was approved for publication on 2019-07-08 at 16:12.DSpace SAF Submission Ingestion Package generated from Vireo submission #14188 on 2019-11-26 at 13:04:36Made available in DSpace on 2019-11-26T20:49:22Z (GMT). No. of bitstreams: 2
ESTES-DISSERTATION-2019.pdf: 2972298 bytes, checksum: e0fe205c9ce999ddfe3542ff24a274a9 (MD5)
LICENSE.txt: 4207 bytes, checksum: 31f9e3f2d76a4cc563bebeb9d5326b7a (MD5)
Previous issue date: 2019-07-08Embargo set by: Seth Robbins for item 112935
Lift date: 2021-11-26T20:49:41Z
Reason: Author requested U of Illinois access only (OA after 2yrs) in Vireo ETD systemU of I Only Restriction Lifted for Item 112935 on 2021-11-27T10:15:20Z
Accelerating Machine Learning Queries with Linear Algebra Query Processing
No full textThe rapid growth of large-scale machine learning (ML) models has led numerous commercial companies to utilize ML models for generating predictive results to help business decision-making. As two primary components in traditional predictive pipelines, data processing, and model predictions often operate in separate execution environments, leading to redundant engineering and computations. Additionally, the diverging mathematical foundations of data processing and machine learning hinder cross-optimizations by combining these two components, thereby overlooking potential opportunities to expedite predictive pipelines. In this paper, we propose an operator fusing method based on GPU-accelerated linear algebraic evaluation of relational queries. Our method leverages linear algebra computation properties to merge operators in machine learning predictions and data processing, significantly accelerating predictive pipelines by up to 317x. We perform a complexity analysis to deliver quantitative insights into the advantages of operator fusion, considering various data and model dimensions. Furthermore, we extensively evaluate matrix multiplication query processing utilizing the widely-used Star Schema Benchmark. Through comprehensive evaluations, we demonstrate the effectiveness and potential of our approach in improving the efficiency of data processing and machine learning workloads on modern hardware.Web Information System
New Approaches for Secure Distance- Bounding
No full textIn this thesis we design and implement three aspects of secure distance-bounding (DB) schemes as a type of authentication scheme that considers distance as an extra verification parameter. By adding this new parameter to authentication schemes, we can prevent certain attacks that are related to distance, such as relay attack. In fact, the attacking scenarios can be much more complex than the simple relay attack, in addition to the classic authentication scheme attacks. In this thesis we consider the most advanced distance-bounding attack scenarios in a variety of authentication models.
We consider three authentication models in order to add the distance as an extra authentication factor: public-key and anonymous DB are the main fields of this thesis that consider strong adversary with access to directional antenna, and we consider One-Shot DB as a one-message authentication scheme. Each of these fields make a chapter of this thesis. Public-Key Distance-Bounding. In a public-key DB scheme, a prover who owns a key pair and is located within a distance bound to a verifier, who has access to the public-key of the prover, tries to convince the verifier that it is authentic and located within the distance bound. We provide a formal model and two protocols with security proofs. Anonymous Distance-Bounding. In an anonymous DB scheme, a prover who owns a registration certificate and is located within a distance bound to a verifier, who only has access to the public parameters of the system, tries to convince the verifier that it is authentic and located within the distance bound without revealing its identity. We provide a formal model and two secure protocols. One-Shot Distance-Bounding. In an one-shot DB scheme, a prover who owns a secret key and is located within a distance bound to a verifier, who has access to the corresponding key of the prover, tries to convince the verifier that it is authentic and located within the distance bound without receiving any message from the verifier. We provide a formal model and a secure protocol
- …
