Spark (mathematics)

About: Spark (mathematics) is a research topic. Over the lifetime, 7304 publications have been published within this topic receiving 63322 citations.

Biomedical Named Entity Recognition at Scale

Abstract: Named entity recognition (NER) is a widely applicable natural language processing task and building block of question answering, topic modeling, information retrieval, etc. In the medical domain, NER plays a crucial role by extracting meaningful chunks from clinical notes and reports, which are then fed to downstream tasks like assertion status detection, entity resolution, relation extraction, and de-identification. Reimplementing a Bi-LSTM-CNN-Char deep learning architecture on top of Apache Spark, we present a single trainable NER model that obtains new state-of-the-art results on seven public biomedical benchmarks without using heavy contextual embeddings like BERT. This includes improving BC4CHEMD to 93.72% (4.1% gain), Species800 to 80.91% (4.6% gain), and JNLPBA to 81.29% (5.2% gain). In addition, this model is freely available within a production-grade code base as part of the open-source Spark NLP library; can scale up for training and inference in any Spark cluster; has GPU support and libraries for popular programming languages such as Python, R, Scala and Java; and can be extended to support other human languages with no code changes.

Stabilized Combustion in a Spark Ignited Engine through a Long Spark Duration

Abstract: An investigation has been done on the relationship between spark ignition characteristics and combustion stability in a gasoline engine The spark discharge parameters examined are the spark current, energy, and duration It is found that lengthening the spark discharge duration is particularly effective in achieving stabilized combustion The analytical results of a constant volume combustion chamber test verify that a longer spark duration promotes flame initiation and makes reliable flame propagation possible The length of the spark duration is regarded as the period from ignition to the onset of combustion pressure rise The spark duration must be longer than the heat release delay The reason is that a long-duration spark shortens the initial combustion period, thereby making it possible to reduce the fluctuations that occur during that period This is particularly important because such fluctuations are known to be the main cause of the cycle-to-cycle fluctuations in engine combustion The spark duration should be three to four times longer than the conventional spark discharge

New spark-test device for material characterization

Abstract: An automated spark test system based on combining field emission and spark measurements, exploiting a discharging capacitor is investigated. In particular, the remaining charge on the capacitor is analytically solved assuming the field emitted current to follow the Fowler Nordheim expression. The latter allows for field emission measurements from pA to A currents, and spark detection by complete discharge of the capacitor. The measurement theory and experiments on Cu and W are discussed.

Cylinder Pressure and Ionization Current Modeling for Spark Ignited Engines

Abstract: Engine management systems (EMS) need feedback on combustion performance to optimally control internal combustion engines. Ion sensing is one of the cheapest and most simple methods for monitoring the combustion event in a spark ignited engine, but still the physical processes behind the formation of the ionization current are not fully understood. The goal here is to investigate models for ionization currents and make a connection to combustion pressure and temperature. A model for the thermal part of an ionization signal is presented that connects the ionization current to cylinder pressure and temperature. One strength of the model is that it after calibration has only two free parameters, burn angle and initial kernel temperature. By fitting the model to a measured ionization signal it is possible to estimate both cylinder pressure and temperature, where the pressure is estimated with good accuracy. The parameterized ionization current model is composed by four parts; a thermal ionization model, a model for formation of nitric oxide, a combustion temperature model and a cylinder pressure function. The pressure function is an empirical function design where the parameters have physical meaning and the function has the main properties of a solution to the cylinder pressure differential equations. The sensitivity of the ionization current model to combustion temperature and content of nitric oxide is investigated to understand the need of sub-model complexity. Two main results are that the pressure model itself well captures the behavior of the cylinder pressure, and that the parameterized ionization current model can be used with an ionization current as input and work as a virtual cylinder pressure sensor and a combustion analysis tool. This ionization current model not only describes the connection between the ionization current and the combustion process, it also offers new possibilities for EMS to control the internal combustion engine.