Transition-Based Deep Input Linearization
نویسندگان
چکیده
Traditional methods for deep NLG adopt pipeline approaches comprising stages such as constructing syntactic input, predicting function words, linearizing the syntactic input and generating the surface forms. Though easier to visualize, pipeline approaches suffer from error propagation. In addition, information available across modules cannot be leveraged by all modules. We construct a transition-based model to jointly perform linearization, function word prediction and morphological generation, which considerably improves upon the accuracy compared to a pipelined baseline system. On a standard deep input linearization shared task, our system achieves the best results reported so far.
منابع مشابه
Improving the performance of wind turbine equipped with DFIG using STATCOM based on input-output feedback linearization controller
Using the FACTS controllers, such as static synchronous compensator (STATCOM), as it provides continuous reactive power, in the grid including wind turbine (WT) equipped with doubly fed induction generator, for improving voltage profile (under normal circumstances) and providing a transition ability from inductor generator transition state has been proposed. In this paper, in order to contr...
متن کاملImproving the performance of wind turbine equipped with DFIG using STATCOM based on input-output feedback linearization controller
Using the FACTS controllers, such as static synchronous compensator (STATCOM), as it provides continuous reactive power, in the grid including wind turbine (WT) equipped with doubly fed induction generator, for improving voltage profile (under normal circumstances) and providing a transition ability from inductor generator transition state has been proposed. In this paper, in order to contr...
متن کاملAdaptive Input-Output Linearization Control of pH Processes
pH control is a challenging problem due to its highly nonlinear nature. In this paper the performances of two different adaptive global linearizing controllers (GLC) are compared. Least squares technique has been used for identifying the titration curve. The first controller is a standard GLC based on material balances of each species. For implementation of this controller a nonlinear state...
متن کاملRobust Control of Encoderless Synchronous Reluctance Motor Drives Based on Adaptive Backstepping and Input-Output Feedback Linearization Techniques
In this paper, the design and implementation of adaptive speed controller for a sensorless synchronous reluctance motor (SynRM) drive system is proposed. A combination of well-known adaptive input-output feedback linearization (AIOFL) and adaptive backstepping (ABS) techniques are used for speed tracking control of SynRM. The AIOFL controller is capable of estimating motor two-axis inductances ...
متن کاملAdaptive Speed Control of Three-Phase Induction Servo-drives Based on Feedback Linearization Theory
In this paper, based on feedback linearization control method and using a special PI (propotational integrator) regulator (IP) in combination with a feed-forward controller, a three-phase induction servo-drive is speed controlled. First, an observer is employed to estimate the rotor d and q axis flux components. Then, two input-output state variables are introduced to control the dynamics of to...
متن کامل