This paper proposes a novel architecture, Cross Attention Augmented Transducer (CAAT), for simultaneous translation. The framework aims to jointly optimize the policy and translation models. To effectively consider all possible READ-WRITE simultaneou
s translation action paths, we adapt the online automatic speech recognition (ASR) model, RNN-T, but remove the strong monotonic constraint, which is critical for the translation task to consider reordering. To make CAAT work, we introduce a novel latency loss whose expectation can be optimized by a forward-backward algorithm. We implement CAAT with Transformer while the general CAAT architecture can also be implemented with other attention-based encoder-decoder frameworks. Experiments on both speech-to-text (S2T) and text-to-text (T2T) simultaneous translation tasks show that CAAT achieves significantly better latency-quality trade-offs compared to the state-of-the-art simultaneous translation approaches.
The acousto-optic filters that depend on the interaction technology between the
optical and ultrasonic waves are considered a high- performance filters as they have the
largest tuning range compared to other filters types. These filters can be used
in optical
WDM (Wavelength Division Multiplexing) networks.
This work aims to study two types of these filters, one operates according to the
non-collinear type of interaction, whereas the other type operates according to the quasicollinear
type of interaction, and compare between these two filters types from the
practical side. Thus, this work aims to show the tilt angle effect of the acoustic wave on the
filter's design and its efficiency in electing the desired optical wave length, in addition to
compare the frequency curves related to these angles in terms of their sensitivity to the
changes of the incidence angle.
The importance of this work lies in proving the efficiency of the filters operating
according to the quasi-collinear type of interaction and their ability to achieve the high
values of quality factor.
The purpose of this research is to detect, locate, and define the blood vessels
in the arm of any person who has a problem in taking samples of blood for
laboratory testing in order to make it easier, not dangerous nor harmful.
Depending on the pr
operties of the scattering wave from the blood and the
depth of the penetration, we have calculated the frequency which is
necessary to choose the suitable transducer including the Geometric
Dimension as well as the materials which is made from (we have taken the
5MHz Doppler –CW for 1.2 cm depth, 2.37 attenuation ratio and the 8MHz
Doppler –CW for 0.74 cm depth , with the same attenuation ratio).
Depending on the velocity's variation of the blood flow throughout the
Biodynamic studies for important arteries in the upper limb, we have found
the Doppler frequency which occurs when the acoustic wave passes across
the blood red cell. We have designed a suitable electronic instrument which
includes the transmitter circuit, receiver circuit, and the output unit - Audio
graph.