Search

Reservoir computing, a machine learning framework used for modeling the brain, can predict temporal data with little observations and minimal computational resources. However, it is difficult to accurately reproduce the long-term target time series because the reservoir system becomes unstable. This predictive capability is required for a wide variety of time-series processing, including predictions of motor timing and chaotic dynamical systems. This study proposes oscillation-driven reservoir computing (ODRC) with feedback, where oscillatory signals are fed into a reservoir network to stabilize the network activity and induce complex reservoir dynamics. The ODRC can reproduce long-term target time series more accurately than conventional reservoir computing methods in a motor timing and chaotic time-series prediction tasks. Furthermore, it generates a time series similar to the target in the unexperienced period, that is, it can learn the abstract generative rules from limited observations. Given these significant improvements made by the simple and computationally inexpensive implementation, the ODRC would serve as a practical model of various time series data. Moreover, we will discuss biological implications of the ODRC, considering it as a model of neural oscillations and their cerebellar processors.

Automatic modulation classification (AMC) is a technology that identifies a modulation scheme without prior signal information and plays a vital role in various applications, including cognitive radio and link adaptation. With the development of deep learning (DL), DL-based AMC methods have emerged, while most of them focus on reducing computational complexity in a centralized structure. This centralized learning-based AMC (CentAMC) violates data privacy in the aspect of direct transmission of client-side raw data. Federated learning-based AMC (FedeAMC) can bypass this issue by exchanging model parameters, but causes large resultant latency and client-side computational load. Moreover, both CentAMC and FedeAMC are vulnerable to large-scale noise occured in the wireless channel between the client and the server. To this end, we develop a novel AMC method based on a split learning (SL) framework, coined SplitAMC, that can achieve high accuracy even in poor channel conditions, while guaranteeing data privacy and low latency. In SplitAMC, each client can benefit from data privacy leakage by exchanging smashed data and its gradient instead of raw data, and has robustness to noise with the help of high scale of smashed data. Numerical evaluations validate that SplitAMC outperforms CentAMC and FedeAMC in terms of accuracy for all SNRs as well as latency., Comment: to be presented at IEEE VTC2023-Spring

In this work, we propose a novel method for modeling numerous speakers, which enables expressing the overall characteristics of speakers in detail like a trained multi-speaker model without additional training on the target speaker's dataset. Although various works with similar purposes have been actively studied, their performance has not yet reached that of trained multi-speaker models due to their fundamental limitations. To overcome previous limitations, we propose effective methods for feature learning and representing target speakers' speech characteristics by discretizing the features and conditioning them to a speech synthesis model. Our method obtained a significantly higher similarity mean opinion score (SMOS) in subjective similarity evaluation than seen speakers of a high-performance multi-speaker model, even with unseen speakers. The proposed method also outperforms a zero-shot method by significant margins. Furthermore, our method shows remarkable performance in generating new artificial speakers. In addition, we demonstrate that the encoded latent features are sufficiently informative to reconstruct an original speaker's speech completely. It implies that our method can be used as a general methodology to encode and reconstruct speakers' characteristics in various tasks., Comment: ICML 2024

Single-stage text-to-speech models have been actively studied recently, and their results have outperformed two-stage pipeline systems. Although the previous single-stage model has made great progress, there is room for improvement in terms of its intermittent unnaturalness, computational efficiency, and strong dependence on phoneme conversion. In this work, we introduce VITS2, a single-stage text-to-speech model that efficiently synthesizes a more natural speech by improving several aspects of the previous work. We propose improved structures and training mechanisms and present that the proposed methods are effective in improving naturalness, similarity of speech characteristics in a multi-speaker model, and efficiency of training and inference. Furthermore, we demonstrate that the strong dependence on phoneme conversion in previous works can be significantly reduced with our method, which allows a fully end-to-end single-stage approach., Comment: Interspeech 2023

Sutton, Szepesv\'{a}ri and Maei introduced the first gradient temporal-difference (GTD) learning algorithms compatible with both linear function approximation and off-policy training. The goal of this paper is (a) to propose some variants of GTDs with extensive comparative analysis and (b) to establish new theoretical analysis frameworks for the GTDs. These variants are based on convex-concave saddle-point interpretations of GTDs, which effectively unify all the GTDs into a single framework, and provide simple stability analysis based on recent results on primal-dual gradient dynamics. Finally, numerical comparative analysis is given to evaluate these approaches.

Binarization of neural network models is considered as one of the promising methods to deploy deep neural network models on resource-constrained environments such as mobile devices. However, Binary Neural Networks (BNNs) tend to suffer from severe accuracy degradation compared to the full-precision counterpart model. Several techniques were proposed to improve the accuracy of BNNs. One of the approaches is to balance the distribution of binary activations so that the amount of information in the binary activations becomes maximum. Based on extensive analysis, in stark contrast to previous work, we argue that unbalanced activation distribution can actually improve the accuracy of BNNs. We also show that adjusting the threshold values of binary activation functions results in the unbalanced distribution of the binary activation, which increases the accuracy of BNN models. Experimental results show that the accuracy of previous BNN models (e.g. XNOR-Net and Bi-Real-Net) can be improved by simply shifting the threshold values of binary activation functions without requiring any other modification., Comment: CVPR 2021, 10 pages, 10 figures

Integrated gradients are widely employed to evaluate the contribution of input features in classification models because it satisfies the axioms for attribution of prediction. This method, however, requires an appropriate baseline for reliable determination of the contributions. We propose a compensated integrated gradients method that does not require a baseline. In fact, the method compensates the attributions calculated by integrated gradients at an arbitrary baseline using Shapley sampling. We prove that the method retrieves reliable attributions if the processes of input features in a classifier are mutually independent, and they are identical like shared weights in convolutional neural networks. Using three electroencephalogram datasets, we experimentally demonstrate that the attributions of the proposed method are more reliable than those of the original integrated gradients, and its computational complexity is much lower than that of Shapley sampling., Comment: Machine Learning for Health (ML4H) Workshop at NeurIPS 2018 arXiv:1811.07216

We present an analytical model that enables throughput evaluation of Opportunistic Spectrum Orthogonal Frequency Division Multiple Access (OS-OFDMA) networks. The core feature of the model, based on a discrete time Markov chain, is the consideration of different channel and subchannel allocation strategies under different Primary and Secondary user types, traffic and priority levels. The analytical model also assesses the impact of different spectrum sensing strategies on the throughput of OS-OFDMA network. The analysis applies to the IEEE 802.22 standard, to evaluate the impact of two-stage spectrum sensing strategy and varying temporal activity of wireless microphones on the IEEE 802.22 throughput. Our study suggests that OS-OFDMA with subchannel notching and channel bonding could provide almost ten times higher throughput compared with the design without those options, when the activity and density of wireless microphones is very high. Furthermore, we confirm that OS-OFDMA implementation without subchannel notching, used in the IEEE 802.22, is able to support real-time and non-real-time quality of service classes, provided that wireless microphones temporal activity is moderate (with approximately one wireless microphone per 3,000 inhabitants with light urban population density and short duty cycles). Finally, two-stage spectrum sensing option improves OS-OFDMA throughput, provided that the length of spectrum sensing at every stage is optimized using our model.

We present an analytical framework to assess the link layer throughput of multichannel Opportunistic Spectrum Access (OSA) ad hoc networks. Specifically, we focus on analyzing various combinations of collaborative spectrum sensing and Medium Access Control (MAC) protocol abstractions. We decompose collaborative spectrum sensing into layers, parametrize each layer, classify existing solutions, and propose a new protocol called Truncated Time Division Multiple Access (TTDMA) that supports efficient distribution of sensing results in "k out of N" fusion rule. In case of multichannel MAC protocols we evaluate two main approaches of control channel design with (i) dedicated and (ii) hopping channel. We propose to augment these protocols with options of handling secondary user (SU) connections preempted by primary user (PU) by (i) connection buffering until PU departure and (ii) connection switching to a vacant PU channel. By comparing and optimizing different design combinations we show that (i) it is generally better to buffer preempted SU connections than to switch them to PU vacant channels and (ii) TTDMA is a promising design option for collaborative spectrum sensing process when k does not change over time., Comment: 43 pages, 14 figures. Includes a concluding discussion on the validity of the analytical model in P. Pawelczak, S. Pollin, H-S. W. So, A. Bahai, R.V. Prasad, R. Hekmat, Performance Analysis of Multichannel Medium Access Control Algorithms for Opportunistic Spectrum Access, IEEE Transactions on Vehicular Technology, vol. 58, no. 6, pp. 3014-3031, Jul. 2009

We will give a necessary condition for local connectedness of the space of Kleinian punctured torus group using Bromgerg’s local coordinate system and provide a sufficient condition for local connectedness on a dense subset of the necessary condition. That is, the collection of the points where the boundary of the space of punctured torus group is not locally connected is a dense subset of the points satisfying the necessary condition.

We will give a necessary condition for local connectedness of the space of Kleinian punctured torus group using Bromgerg’s local coordinate system and provide a sufficient condition for local connectedness on a dense subset of the necessary condition. That is, the collection of the points where the boundary of the space of punctured torus group is not locally connected is a dense subset of the points satisfying the necessary condition.