共找到 6 条结果
Whereas the world relies on computer systems for providing public services, there is a lack of academic work that systematically assess the security of government systems. To partially fill this gap, we conducted a security evaluation of publicly available systems from public institutions. We revisited OWASP top-10 and identified multiple vulnerabilities in deployed services by scanning public government networks. Overall, the unprotected services found have inadequate security level, which must be properly discussed and addressed.
Progress in Natural Language Processing (NLP) has been dictated by the rule of more: more data, more computing power and more complexity, best exemplified by the Large Language Models. However, training (or fine-tuning) large dense models for specific applications usually requires significant amounts of computing resources. This \textbf{Ph.D. dissertation} focuses on an under-investi\-gated NLP data engineering technique, whose potential is enormous in the current scenario known as Instance Selection (IS). The IS goal is to reduce the training set size by removing noisy or redundant instances while maintaining the effectiveness of the trained models and reducing the training process cost. We provide a comprehensive and scientifically sound comparison of IS methods applied to an essential NLP task -- Automatic Text Classification (ATC), considering several classification solutions and many datasets. Our findings reveal a significant untapped potential for IS solutions. We also propose two novel IS solutions that are noise-oriented and redundancy-aware, specifically designed for large datasets and transformer architectures. Our final solution achieved an average reduction of 41\% in
This paper develops a general framework for stochastic modeling of goals and other events in football (soccer) matches. The events are modelled as Cox processes (doubly stochastic Poisson processes) where the event intensities may depend on all the modeled events as well as external factors. The model has a strictly concave log-likelihood function which facilitates its fitting to observed data. Besides event times, the model describes the random lengths of stoppage times which can have a strong influence on the final score of a match. The model is illustrated on eight years of data from Campeonato Brasileiro de Futebol Série A. We find that dynamic regressors significantly improve the in-game predictive power of the model. In particular, a) when a team receives a red card, its goal intensity decreases more than 30%; b) the goal rate of a team increases by 10% if it is losing by one goal and by 20% if its losing by two goals; and c) when the goal difference at the end of the second half is less than or equal to one, the stoppage time is on average more than one minute longer than in matches with a difference of two goals.
These lecture notes are written as reference material for the Advanced Course "Hydrodynamical Methods in Last Passage Percolation Models", given at the 28th Coloquio Brasileiro de Matematica at IMPA, Rio de Janeiro, July 2011.
This work addressed the use of the geometric Brownian motion to simulate the prices of shares listed in the Small Caps index of the Brazilian stock exchange B3 (Brazil, Bolsa, Balcão). The data used refer to the price history from January 2016 to December 2018. The price history of 2019 was used to be compared with the simulated prices. The data was imported from the Yahoo Finance database using the Python programming language, and the simulations were performed for each stock individually, and for portfolios formed based on expected returns, risk and the Sharpe Index. The results were better for portfolios with higher returns, lower risks and higher Sharpe Indexes.
ISDB-T International standard is currently adopted by most Latin America countries and is already installed in most TV sets sold in recent years in the region. To support interactive applications in Digital TV receivers, ISDB-T defines the middleware Ginga. Similar to Digital TV, Digital Radio standards also provide the means to carry interactive applications; however, their specifications for interactive applications are usually more restricted than the ones used in Digital TV. Also, interactive applications for Digital TV and Digital Radio are usually incompatible. Motivated by such observations, this report considers the importance of interactive applications for both TV and Radio Broadcasting and the advantages of using the same middleware and languages specification for Digital TV and Radio. More specifically, it establishes the signaling and definitions on how to transport and execute Ginga-NCL and Ginga-HTML5 applications over DRM (Digital Radio Mondiale) transmission. Ministry of Science, Technology, Innovation and Communication of Brazil is carrying trials with Digital Radio Mondiale standard in order to define the reference model of the Brazilian Digital Radio System (Por