341 research outputs found
An overview of the ATLAS High Level Trigger Dataflow and Supervision.
No full textThe ATLAS high-level trigger (HLT) system provides software-based event selection after the initial LVL1 hardware trigger. It is composed of two stages, the LVL2 trigger and the event filter (EF). The LVL2 trigger performs event selection with optimized algorithms using selected data guided by Region of Interest pointers provided by the LVL1 trigger. Those events selected by LVL2 are built into complete events, which are passed to the EF for a further stage of event selection and classification using off-line algorithms. Events surviving the EF selection are passed for off-line storage. The two stages of HLT are implemented on processor farms. The concept of distributing the selection process between LVL2 and EF is a key element in the architecture, which allows it to be flexible to changes (luminosity, detector knowledge, background conditions, etc.) Although there are some differences in the requirements between these subsystems there are many commonalities. An overview of the dataflow (event selection) and supervision (control, configuration, monitoring) activities in the HLT is given, highlighting where commonalities between the two subsystems can be exploited and indicating where requirements dictate that implementations differ. An HLT prototype system has been built at CERN. Functional testing is being carried out in order to validate the HLT architecture
An empirical study of semantic similarity in WordNet and Word2Vec
No full textThis thesis performs an empirical analysis of Word2Vec by comparing its output to WordNet, a well-known, human-curated lexical database. It finds that Word2Vec tends to uncover more of certain types of semantic relations than others -- with Word2Vec returning more hypernyms, synonomyns and hyponyms than hyponyms or holonyms. It also shows the probability that neighbors separated by a given cosine distance in Word2Vec are semantically related in WordNet. This result both adds to our understanding of the still-unknown Word2Vec and helps to benchmark new semantic tools built from word vectors
Beyond the Ledger: A Cross-Platform Analysis of Cryptocurrency Dynamics
No full textThe cryptocurrency market has experienced rapid growth and dynamic changes, marked by evolving technologies, fluctuating market dynamics, and significant regulatory events. This thesis, ‘Beyond the Ledger: A Cross-Platform Exploratory Analysis of Cryptocurrency Data Dynamics,’ explores the intricate relationships between cryptocurrencies and online platforms such as Reddit, Discord, and GitHub, which serve as critical channels for community engagement, technical development, and information dissemination; through a multi-coin, multi-platform analysis, this study examines the interconnectedness within the cryptocurrency ecosystem, utilizing Granger causality tests and correlation matrices to identify predictive relationships and patterns across the market. The research delves into significant market events such as the Ethereum Merge and the collapse of FTX, analyzing their impact on cryptocurrency prices, community activity, and technological trends and investigates the roles that different platforms play in fostering community sentiment andtechnological innovation, revealing that Reddit tends to drive market speculation, Discord offers real-time community interaction, and GitHub reflects longer-term development activities. Furthermore, the study highlights the importance of cross-platform dynamics during market shocks, demonstrating how information flows across various online spaces influence market behavior and development priorities. The findings suggest that while Bitcoin remains a market leader with a relatively independent behavior, other cryptocurrencies like Ethereum, Solana, Polygon, Cardano, and Dogecoin exhibit varied levels of interconnectedness and are influenced by different factors, including social media trends, technological advancements, and external economic conditions. This thesis concludes that understanding the complex socio-technical ecology of the cryptocurrency market requires a multifaceted approach that considers not only financial and technological elements but also the broader context of community interactions and market shocks; these results contribute to a better understanding of the cryptocurrency ecosystem’s dynamics and offer valuable insights for investors, developers, and researchers seeking to navigate the volatile and rapidly evolving world of cryptocurrencies. Further research is recommended to refine predictive models and expand the scope of analysis to improve generalizability and enhance the comprehensiveness of market studies.</p
Impacts of the Medicaid Expansion Under the Affordable Care Act on Health Insurance Coverage
No full text
This paper looks at the impacts of the Medicaid expansion of 2014. In particular: What is the long-run effect of Medicaid Expansion during the ACA on uninsured rates in the United States? I will be utilizing a newer methodology, the difference-in-differences staggered adoption model which allows for multiple time periods. This model proves to be more refined and less biased in comparison to the normal diff-in-diff used in this field. I find that, the Medicaid expansion of 2014 caused an average change in uninsured rates of 2.33 percentage points for states that expanded in comparison to what they would have experienced had they not. This drops to 1.87 percentage points if we allow for one year of anticipation (changes in behaviors of citizens prior to implementation) in our data.
</div
Impacts of the Medicaid Expansion Under the Affordable Care Act on Health Insurance Coverage
No full text
This paper looks at the impacts of the Medicaid expansion of 2014. In particular: What is the long-run effect of Medicaid Expansion during the ACA on uninsured rates in the United States? I will be utilizing a newer methodology, the difference-in-differences staggered adoption model which allows for multiple time periods. This model proves to be more refined and less biased in comparison to the normal diff-in-diff used in this field. I find that, the Medicaid expansion of 2014 caused an average change in uninsured rates of 2.33 percentage points for states that expanded in comparison to what they would have experienced had they not. This drops to 1.87 percentage points if we allow for one year of anticipation (changes in behaviors of citizens prior to implementation) in our data.
</div
Studying the Impact of Varying Levels of AI Advice on AI-Assisted Human Decision-Making
No full textThe increasing use of AI systems in high-stakes decision-making necessitates a deeper understanding of how these tools interact with human users. AI-assisted human decision-making aims to leverage the relative strengths of humans and AI to achieve complementary team performance. Complementarity depends on the effective reliance of human decision-makers on AI advice. Recent studies focus on AI explanations and human cognitive processes as ways to improve the appropriate use of AI advice. However, much of this research employs highly accurate AI tools in simplified tasks that fail to capture the challenges faced in the real-world deployment of AI decision support systems. This research aims to address this gap by examining the performance of AI-assisted human decision-making in a complex, three-way, ordinal classification task with inconsistent accuracy of the AI tool. Specifically, this research conducts a two-phase study involving: 1) a classroom field study and 2) an online experiment involving a real-world AI detection tool. This research examines the impact of AI advice of varying accuracy and tool information on classification performance, decision confidence, and agreement with AI recommendations. The findings show that while accurate AI advice can improve human performance, inaccuracies in AI advice can lead to humans making high-confidence errors. Complementarity is difficult to achieve due to the suboptimal reliance of human decision-makers on the AI tool’s advice. Furthermore, tool information, such as tool recommendation confidence and tool warning messages, can be effective in moderating human agreement with the tool and their decision confidence; however, the benefits do not transfer to combined performance. This research contributes to the existing body of knowledge on AI-augmented decision-making by providing empirical insights into the challenges of human-AI collaboration in realistic, educational contexts and the influence of tool information on combined decision-making.</p
Beyond the Ledger: A Cross-Platform Analysis of Cryptocurrency Dynamics
No full textThe cryptocurrency market has experienced rapid growth and dynamic changes, marked by evolving technologies, fluctuating market dynamics, and significant regulatory events. This thesis, ‘Beyond the Ledger: A Cross-Platform Exploratory Analysis of Cryptocurrency Data Dynamics,’ explores the intricate relationships between cryptocurrencies and online platforms such as Reddit, Discord, and GitHub, which serve as critical channels for community engagement, technical development, and information dissemination; through a multi-coin, multi-platform analysis, this study examines the interconnectedness within the cryptocurrency ecosystem, utilizing Granger causality tests and correlation matrices to identify predictive relationships and patterns across the market. The research delves into significant market events such as the Ethereum Merge and the collapse of FTX, analyzing their impact on cryptocurrency prices, community activity, and technological trends and investigates the roles that different platforms play in fostering community sentiment andtechnological innovation, revealing that Reddit tends to drive market speculation, Discord offers real-time community interaction, and GitHub reflects longer-term development activities. Furthermore, the study highlights the importance of cross-platform dynamics during market shocks, demonstrating how information flows across various online spaces influence market behavior and development priorities. The findings suggest that while Bitcoin remains a market leader with a relatively independent behavior, other cryptocurrencies like Ethereum, Solana, Polygon, Cardano, and Dogecoin exhibit varied levels of interconnectedness and are influenced by different factors, including social media trends, technological advancements, and external economic conditions. This thesis concludes that understanding the complex socio-technical ecology of the cryptocurrency market requires a multifaceted approach that considers not only financial and technological elements but also the broader context of community interactions and market shocks; these results contribute to a better understanding of the cryptocurrency ecosystem’s dynamics and offer valuable insights for investors, developers, and researchers seeking to navigate the volatile and rapidly evolving world of cryptocurrencies. Further research is recommended to refine predictive models and expand the scope of analysis to improve generalizability and enhance the comprehensiveness of market studies.</p
Studying the Impact of Varying Levels of AI Advice on AI-Assisted Human Decision-Making
No full textThe increasing use of AI systems in high-stakes decision-making necessitates a deeper understanding of how these tools interact with human users. AI-assisted human decision-making aims to leverage the relative strengths of humans and AI to achieve complementary team performance. Complementarity depends on the effective reliance of human decision-makers on AI advice. Recent studies focus on AI explanations and human cognitive processes as ways to improve the appropriate use of AI advice. However, much of this research employs highly accurate AI tools in simplified tasks that fail to capture the challenges faced in the real-world deployment of AI decision support systems. This research aims to address this gap by examining the performance of AI-assisted human decision-making in a complex, three-way, ordinal classification task with inconsistent accuracy of the AI tool. Specifically, this research conducts a two-phase study involving: 1) a classroom field study and 2) an online experiment involving a real-world AI detection tool. This research examines the impact of AI advice of varying accuracy and tool information on classification performance, decision confidence, and agreement with AI recommendations. The findings show that while accurate AI advice can improve human performance, inaccuracies in AI advice can lead to humans making high-confidence errors. Complementarity is difficult to achieve due to the suboptimal reliance of human decision-makers on the AI tool’s advice. Furthermore, tool information, such as tool recommendation confidence and tool warning messages, can be effective in moderating human agreement with the tool and their decision confidence; however, the benefits do not transfer to combined performance. This research contributes to the existing body of knowledge on AI-augmented decision-making by providing empirical insights into the challenges of human-AI collaboration in realistic, educational contexts and the influence of tool information on combined decision-making.</p
Effect Handler Oriented Programming for Data Processing Applications
No full textEffect handler oriented programming or EHOP for short, is a new programming paradigm aiming to achieve separation of concerns in code which will lead to modular, readable and maintainable code. Since EHOP is significantly new, it is important to assess and compare it against traditional, commonly used paradigms in order to see if a wider adoption of EHOP would prove beneficial to computer science. In this research, EHOP was compared with traditional paradigms under the context of data processing applications. An Excel-like command line application called “MiniExcel” was implemented from scratch. Moreover, “Hierarchical EHOP”, a new structural pattern for EHOP was defined which enforces rules between concepts and produces a readable code structure. The main conclusions of this research can be summarized by the following statements. EHOP produces more modular, readable and maintainable code compared to traditional paradigms. Implementing additional concepts and updates to code is seamless using EHOP, yet the lack of development in EHOP’s ecosystem raises frustrating errors and requires the developer to implement libraries that are usually built-in for languages that support traditional paradigms. Functional programming produces faster running code, but EHOP is more memory efficient. Therefore, for applications that interact with users EHOP is the better choice and for applications that only execute code functional programming is more suitable.https://github.com/alibasaran/EHOP-Excel/ The codebase containing the implemented application as well as scripts that were used in programming paradigm analysis.CSE3000 Research ProjectComputer Science and Engineerin
Performative Metafiction: Lemony Snicket, Daniel Handler and The End of <i>A Series of Unfortunate Events</i>
No full textIn "Performative Metafiction: Lemony Snicket, Daniel Handler, and The End of A Series of Unfortunate Events," Sara Austin looks at the metafictional aspect of Lemony Snicket's A Series of Unfortunate Events, with particular emphasis on the series' final volume, The End. She explores the occasional uneasy relationship between the series narrator and "author," Lemony Snicket, and the actual author, Daniel Handler. Handler's entirely pseudonymous role in the publishing process creates a tension within the series' narrative authority, raising issues that, often, adults do not trust children to understand. The popularity of the series, particularly in the United States, belies assumptions that children will neither understand nor enjoy books that raise more questions about the plot and characters than they answer, or that utterly fail to offer the "happily-ever-after" convention that so dominates the worlds of children's publishing
- …
