Equalization with Neural Network Circuitry for High-Speed Signal Link

Chu, Yunhui; Chen, Fan; Lang, John Temple; Lee, Beomtaek

doi:10.1109/isemc.2019.8825275

Cited by 6 publications

(4 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The forget gate, f t decides which information needs attention and which can be ignored. The input gate, i t decides what information is relevant to update in the current Compute ℰ ′′ from ℰ ′ using (15) and fit G with ℋ and ℰ ′′ .…”

Section: Returnmentioning

confidence: 99%

“…For example, [8] uses a short transient simulation to train a polynomial chaos surrogate model and the surrogate model is then used to estimate the jitter and eye diagram of the output signal, while [9] uses Bayesian optimization to perform a worst-case analysis of the eye diagram. In addition, neural network based techniques have also been developed such as in the application of multilayer perceptron (MLP) neural network for the modeling of the eye diagram [10]- [13], jitter [14], channel equalization [15], and physical parameters such as resistance, conductance, inductance, and capacitance [16]. However, MLP lacks the ability to capture sequential information.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Transient Simulations of High-Speed Channels Using CNN-LSTM With an Adaptive Successive Halving Algorithm for Automated Hyperparameter Optimizations

Goay

Ahmad

Goh³

2021

IEEE Access

View full text Add to dashboard Cite

Transient simulations of high-speed channels can be very time intensive. Recurrent neural network (RNN) based methods can be used to speed up the process by training a RNN model on a relatively short bit sequence, and then using a multi-steps rolling forecast method to predict subsequent bits. However, the performance of the RNN model is highly affected by its hyperparameters. We propose an algorithm named adaptive successive halving automated hyperparameter optimization (ASH-HPO) which combines successive halving, Bayesian optimization (BO), and progressive sampling to tune the hyperparameters of the RNN models. Modifications are proposed to the successive halving and progressive sampling algorithms for better efficiency on time series data. The ASH-HPO algorithm trains on smaller dataset subsets initially, then expands the training dataset progressively and adaptively adds or removes models along the process. In this paper, we use the ASH-HPO algorithm to optimize the hyperparameters of convolutional neural networks (CNNs), long short-term memory (LSTM) networks, and CNN-LSTM networks. We demonstrate the effectiveness of the ASH-HPO algorithm using a PCIe Gen 2 channel, a PCIe Gen 5 channel, and a PAM4 differential channel. We also investigate the effects of several settings and tunable variables of the ASH-HPO algorithm on its convergence speed. As a benchmark, we compared the ASH-HPO algorithm to three state-of-the-art HPO methods: BO, successive halving, and hyperband. The results show that the ASH-HPO algorithm converges faster than the other HPO methods on transient simulation problems.INDEX TERMS Automated hyperparameter optimization, convolutional neural network (CNN), highspeed channel, long short-term memory (LSTM) network, progressive sampling, transient simulation.

show abstract

Section: Returnmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Transient Simulations of High-Speed Channels Using CNN-LSTM With an Adaptive Successive Halving Algorithm for Automated Hyperparameter Optimizations

Goay

Ahmad

Goh³

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Recent work of Chu et al [20] introduced a neural network circuitry equalizer (NNE) for high-speed signal link construction. An artificial neural network (ANN) is integrated into the continuous-time linear EQ (CTLE) as the replacement of the DFE block (block (b) in Fig.…”

Section: Introductionmentioning

confidence: 99%

Long Short-Term Memory Neural Equalizer

Wang

et al. 2022

SSRN Journal

View full text Add to dashboard Cite

A trainable neural equalizer based on Long Short-Term Memory (LSTM) neural network architecture is proposed in this paper to recover the channel output signal. The current widely used solution for the transmission line signal recovering is generally realized through DFE or FFE-DFE combination. The novel learning-based equalizer is suitable for highly non-linear signal restoration thanks to its recurrent design. The effectiveness of the LSTM equalizer is shown through an ADS simulation channel signal equalization task including a quantitative and qualitative comparison with an FFE-DFE combination. The LSTM neural network shows good equalization results compared to that of the FFE-DFE combination. The advantage of a trainable LSTM equalizer lies in its ability to learn its parameters in a flexible manner, to tackle complex scenario without any hardware modification. This can reduce the equalizer implantation cost for variant transmission channels and bring additional portability in practical applications.

show abstract

“…ERY high-speed parallel data transmission is facing increasing challenges due to higher data rate requirements and routing density. Complex equalization techniques for the transmitter and receiver side are being developed to tackle the signal degradation effects related to resistive loss, current leakage, impedance mismatches, and inter-symbol interference (ISI) [1]- [4]. On the crosstalk side, higher speed requirements translate into larger electromagnetic interference coming from neighboring wires, together with thinner geometries with higher inductive effects.…”

Section: Introductionmentioning

confidence: 99%