Dataset distillation (DD) aims to generate a compact synthetic dataset that enables efficient training of neural networks while maintaining performance comparable to that achieved with the original dataset. However, existing methods often suffer from two main limitations. They either rely on computationally intensive iterative optimization procedures or depend heavily on architecture-specific designs. These issues limit their practicality for large-scale datasets and hinder generalization across different model architectures. To overcome these challenges, recent research has explored the use of diffusion models as an architecture-agnostic approach to dataset distillation, offering improved scalability and generalization for large-scale datasets across diverse model architectures. While diffusion-based dataset distillation methods have shown considerable potential, several challenges remain. Notably, certain approaches exhibit a distributional mismatch between the pre-trained diffusion model and the target dataset, which can adversely affect the fidelity and representativeness of the generated samples. Others require substantial fine-tuning to achieve high fidelity, which negates the benefits of architectural flexibility. In this work, we propose a new diffusion-based dataset distillation framework that effectively preserves the characteristics of the original dataset without requiring any fine-tuning. Our method employs adaptive sampling and repulsion regularization to enhance both the fidelity and diversity of generated samples. As a result, the proposed approach outperforms state-of-the-art distillation methods across a wide range of datasets and model architectures.

Recent research has devoted considerable effort to verifying the intermediate reasoning steps of chain-of-thought (CoT) trajectories using process reward models (PRMs) and other verifier models.However, training a PRM typically requires human annotators to assign reward scores to each reasoning step, which is both costly and time-consuming.Existing automated approaches, such as Monte Carlo (MC) estimation, also demand substantial computational resources due to repeated LLM rollouts.To overcome these limitations, we propose contrastive pointwise mutual information (CPMI), a novel automatic reward labeling method that leverages the model's internal probability to infer step-level supervision while significantly reducing the computational burden of annotating dataset.CPMI quantifies how much a reasoning step increases the mutual information between the step and the correct target answer relative to hard-negative alternatives.This contrastive signal serves as a proxy for the step's contribution to the final solution and yields a reliable reward.The experimental results show that CPMI-based labeling reduces dataset construction time by 84% and token generation by 98% compared to MC estimation, while achieving higher accuracy on process-level evaluations and mathematical reasoning benchmarks.The code is available at https://github.com/nakyungLee20/CPMI.

Pruning is a highly effective method for reducing the size of neural networks with negligible impact on their average performance. However, recent studies have revealed that pruning actually amplifies the bias in the models, leading to decreased performance for underrepresented groups. To address this issue, we first analyze the impact of pruning on the confidence of each sample and introduce Accumulated Confidence (AC). AC is a proxy that facilitates the identification of bias-conflicting and bias-aligned samples without relying on group annotations. We then propose a debiasing algorithm, which is called DEbiasing Network through Pruning (DENP). DENP utilizes AC to mitigate bias within the network. Even without bias information, DENP exhibits remarkable debiasing performance on varying levels of sparsity, effectively mitigating the bias-exacerbating property of pruning and resulting in both sparse and debiased neural networks. Moreover, even when compared with state-of-the-art debiasing baselines under identical conditions, the DENP still achieves the best performance on multiple benchmark datasets, demonstrating its superior debiasing capabilities.

As the use of machine learning models has increased, numerous studies have aimed to enhance fairness. However, research on the intersection of fairness and explainability remains insufficient, leading to potential issues in gaining the trust of actual users. Here, we propose a novel module that constructs a fair latent space, enabling faithful explanation while ensuring fairness. The fair latent space is constructed by disentangling and redistributing labels and sensitive attributes, allowing the generation of counterfactual explanations for each type of information. Our module is attached to a pretrained generative model, transforming its biased latent space into a fair latent space. Additionally, since only the module needs to be trained, there are advantages in terms of time and cost savings, without the need to train the entire generative model. We validate the fair latent space with various fairness metrics and demonstrate that our approach can effectively provide explanations for biased decisions and assurances of fairness.

Recent advancements in quantum computing highlight the need for efficient encoding of classical data into quantum states to ensure robust quantum information processing. Traditional encoding schemes often impose impractical requirements about the knowledge of quantum states and lack adaptability to noisy quantum channels and broader tasks. To address these limitations, we propose a novel end-to-end learnable quantum transcoding scheme explicitly optimized for compactness and robustness in noisy quantum communication scenarios. Our approach integrates neural network-based data compression with Cholesky decomposition-based quantum encoding and bypasses full density matrix reconstruction. Through normalized quantum observables, our method enables efficient tomography and achieves high reconstruction and classification performance even under extreme noise conditions.

Vector Quantization (VQ)-based digital semantic communication is gaining prominence due to its compatibility with existing digital systems. However, it faces a critical trade-off between computational efficiency and the risk of codeword overfitting. To address this challenge, we propose ResUME, a novel digital semantic communication framework designed for both rate-controllability and computational efficiency. ResUME integrates residual vector quantization (RVQ) with unequal modulation to achieve robust and flexible encoding of latent representations. Our framework leverages the hierarchical structure of RVQ, where layers encode information of varying significance. We analyze this property from an information-theoretic perspective to optimize the modulation order. This optimization provides a form of unequal error protection, maximizing mutual information under spectral efficiency constraints. Furthermore, to mitigate the risk of overfitting, we introduce a variational loss function inspired by the information bottleneck principle. Experimental results demonstrate that ResUME significantly outperforms conventional modulation-assisted methods across various AWGN channel conditions. An ablation study further validates the contribution of each component in our proposed loss function.

Distributional reinforcement learning improves performance by capturing environmental stochasticity, but a comprehensive theoretical understanding of its effectiveness remains elusive. In addition, the intractable element of the infinite dimensionality of distributions has been overlooked. In this paper, we present a regret analysis of distributional reinforcement learning with general value function approximation in a finite episodic Markov decision process setting. We first introduce a key notion of Bellman unbiasedness which is essential for exactly learnable and provably efficient distributional updates in an online manner. Among all types of statistical functionals for representing infinite-dimensional return distributions, our theoretical results demonstrate that only moment functionals can exactly capture the statistical information. Secondly, we propose a provably efficient algorithm, SF-LSVI, that achieves a tight regret bound of $\tilde{O}(d_E H^{\frac{3}{2}}\sqrt{K})$ where $H$ is the horizon, $K$ is the number of episodes, and $d_E$ is the eluder dimension of a function class.

To design reward that align with human goals, Reinforcement Learning from Human Feedback (RLHF) has emerged as a prominent technique for learning reward functions from human preferences and optimizing models using reinforcement learning algorithms. However, existing RLHF methods often misinterpret trajectories as being generated by an optimal policy, causing inaccurate likelihood estimation and suboptimal learning. To address this, we propose Policy-labeled Preference Learning (PPL) within the Direct Preference Optimization (DPO) framework, which resolves these likelihood mismatch problems by modeling human preferences with regret, reflecting the efficiency of executed policies. Additionally, we introduce a contrastive KL regularization term derived from regret-based principles to enhance sequential contrastive learning. Experiments in high-dimensional continuous control environments demonstrate PPL's significant improvements in offline RLHF performance and its effectiveness in online settings.

Despite their impressive generative capabilities, text-to-image diffusion models often memorize and replicate training data, prompting serious concerns over privacy and copyright. Recent work has attributed this memorization to an attraction basin—a region where applying classifier-free guidance (CFG) steers the denoising trajectory toward memorized outputs—and has proposed deferring CFG application until the denoising trajectory escapes this basin. However, such delays often result in non-memorized images that are poorly aligned with the input prompts, highlighting the need to promote earlier escape so that CFG can be applied sooner in the denoising process. In this work, we show that the initial noise sample plays a crucial role in determining when this escape occurs. We empirically observe that different initial samples lead to varying escape times. Building on this insight, we propose two mitigation strategies that adjust the initial noise—either collectively or individually—to find and utilize initial samples that encourage earlier basin escape. These approaches significantly reduce memorization while preserving image-text alignment.

Fine-tuning pre-trained neural network models has become a widely adopted approach across various domains. However, it can lead to the distortion of pre-trained feature extractors that already possess strong generalization capabilities. Mitigating feature distortion during adaptation to new target domains is crucial. Recent studies have shown promising results in handling feature distortion by aligning the head layer on in-distribution datasets before performing fine-tuning. Nonetheless, a significant limitation arises from the treatment of batch normalization layers during fine-tuning, leading to suboptimal performance. In this paper, we propose Domain-Aware Fine-Tuning (DAFT), a novel approach that incorporates batch normalization conversion and the integration of linear probing and fine-tuning. Our batch normalization conversion method effectively mitigates feature distortion by reducing modifications to the neural network during fine-tuning. Additionally, we introduce the integration of linear probing and fine-tuning to optimize the head layer with gradual adaptation of the feature extractor. By leveraging batch normalization layers and integrating linear probing and fine-tuning, our DAFT significantly mitigates feature distortion and achieves improved model performance on both in-distribution and out-of-distribution datasets. Extensive experiments demonstrate that our method outperforms other baseline methods, demonstrating its effectiveness in not only improving performance but also mitigating feature distortion.

Models trained with empirical risk minimization (ERM) are prone to be biased towards spurious correlations between target labels and bias attributes, which leads to poor performance on data groups lacking spurious correlations. It is particularly challenging to address this problem when access to bias labels is not permitted. To mitigate the effect of spurious correlations without bias labels, we first introduce a novel training objective designed to robustly enhance model performance across all data samples, irrespective of the presence of spurious correlations. From this objective, we then derive a debiasing method, Disagreement Probability based Resampling for debiasing (DPR), which does not require bias labels. DPR leverages the disagreement between the target label and the prediction of a biased model to identify bias-conflicting samples—those without spurious correlations—and upsamples them according to the disagreement probability. Empirical evaluations on multiple benchmarks demonstrate that DPR achieves state-of-the-art performance over existing baselines that do not use bias labels. Furthermore, we provide a theoretical analysis that details how DPR reduces dependency on spurious correlations.

Distributed computing has emerged as a promising solution for accelerating machine learning training processes on large-scale datasets by leveraging the parallel processing capabilities of multiple workers. However, there remain two major issues that still need to be addressed: i) Byzantine attacks from malicious workers, and ii) the effect of slow workers, commonly referred to as stragglers. In this paper, we address both issues concurrently by introducing Group-wise Verifiable Coded Computing (GVCC), a novel approach that combines coding techniques and group-wise verification to enhance robustness against Byzantine attacks and resilience to straggler effects in distributed computing. The key idea of GVCC is to verify a group of computation results from workers at a time, while providing resilience to stragglers through encoding tasks assigned to workers with Group-wise Verifiable Codes. We evaluate the performance of GVCC through experiments conducted on Amazon EC2 clouds and the results show that GVCC outperforms the existing methods in terms of overall processing time and verification time while maintaining the verification performance. This study highlights the potential of GVCC as an effective solution for overcoming the challenges of Byzantine attacks and stragglers in distributed computing for executing matrix multiplication.

Due to the expensive costs of collecting labels in multi-label classification datasets, partially annotated multi-label classification has become an emerging field in computer vision. One baseline approach to this task is to assume unobserved labels as negative labels, but this assumption induces label noise as a form of false negative. To understand the negative impact caused by false negative labels, we study how these labels affect the model's explanation. We observe that the explanation of two models, trained with full and partial labels each, highlights similar regions but with different scaling, where the latter tends to have lower attribution scores. Based on these findings, we propose to boost the attribution scores of the model trained with partial labels to make its explanation resemble that of the model trained with full labels. Even with the conceptually simple approach, the multi-label classification performance improves by a large margin in three different datasets on a single positive label setting and one on a large-scale partial label setting. Code is available at https://github.com/youngwk/BridgeGapExplanationPAMC.

Distributional reinforcement learning algorithms have attempted to utilize estimated uncertainty for exploration, such as optimism in the face of uncertainty. However, using the estimated variance for optimistic exploration may cause biased data collection and hinder convergence or performance. In this paper, we present a novel distributional reinforcement learning that selects actions by randomizing risk criterion without losing the risk-neutral objective. We provide a perturbed distributional Bellman optimality operator by distorting the risk measure. Also,we prove the convergence and optimality of the proposed method with the weaker contraction property. Our theoretical results support that the proposed method does not fall into biased exploration and is guaranteed to converge to an optimal return. Finally, we empirically show that our method outperforms other existing distribution-based algorithms in various environments including Atari 55 games.

Alleviating overestimation bias is a critical challenge for deep reinforcement learning to achieve successful performance on more complex tasks or offline datasets containing out-of-distribution data. In order to overcome overestimation bias, ensemble methods for Q-learning have been investigated to exploit the diversity of multiple Q-functions. Since network initialization has been the predominant approach to promote diversity in Q-functions, heuristically designed diversity injection methods have been studied in the literature. However, previous studies have not attempted to approach guaranteed independence over an ensemble from a theoretical perspective. By introducing a novel regularization loss for Q-ensemble independence based on random matrix theory, we propose spiked Wishart Q-ensemble independence regularization (SPQR) for reinforcement learning. Specifically, we modify the intractable hypothesis testing criterion for the Q-ensemble independence into a tractable KL divergence between the spectral distribution of the Q-ensemble and the target Wigner's semicircle distribution. We implement SPQR in several online and offline ensemble Q-learning algorithms. In the experiments, SPQR outperforms the baseline algorithms in both online and offline RL benchmarks.

One of the objectives of continual learning is to prevent catastrophic forgetting in learning multiple tasks sequentially, and the existing solutions have been driven by the conceptualization of the plasticity-stability dilemma. However, the convergence of continual learning for each sequential task is less studied so far. In this paper, we provide a convergence analysis of memory-based continual learning with stochastic gradient descent and empirical evidence that training current tasks causes the cumulative degradation of previous tasks. We propose an adaptive method for nonconvex continual learning (NCCL), which adjusts step sizes of both previous and current tasks with the gradients. The proposed method can achieve the same convergence rate as the SGD method when the catastrophic forgetting term which we define in the paper is suppressed at each iteration. Further, we demonstrate that the proposed algorithm improves the performance of continual learning over existing methods for several image classification tasks.

In this paper, we consider coded computation for matrix multiplication tasks in distributed computing to mitigate straggler effects. We assume that the stragglers’ computation results can be leveraged at the master by assigning multiple sub-tasks to the workers. We propose a new coded computation scheme, namely Chebyshev coded fully private matrix multiplication (CFP), to preserve the privacy of a master in a scenario where a master wants to obtain a matrix multiplication result from the libraries which are shared by the workers, while concealing both of the two indices of the desired matrices from each worker. The key idea of CFP is to introduce Chebyshev polynomials, which have commutative property, in queries sent to workers to allocate sub-tasks. We also extend CFP to keep the privacy of a master from colluding workers. In conclusion, we show that CFP can preserve the privacy of a master from each worker and efficiently mitigate straggler effects compared to existing schemes.

In this paper, we propose a Doppler compensation method in LEO-MIMO communication, where two LEO satellites are used as amplify-and-forward (AF) relays, and line-of-sight (LOS) propagation dominates the communication channel. We first analyze Doppler effects in uplink and downlink, and propose a dual-hop AF relay channel model including Doppler effects. Also we show that Doppler effects in our scenario can be easily compensated at the LEO satellites, without the need of intersatellite link (ISL) communication.

Weakly supervised multi-label classification (WSML) task, which is to learn a multi-label classification using partially observed labels per image, is becoming increasingly important due to its huge annotation cost. In this work, we first regard unobserved labels as negative labels, casting the WSML task into noisy multi-label classification. From this point of view, we empirically observe that memorization effect, which was first discovered in a noisy multi-class setting, also occurs in a multi-label setting. That is, the model first learns the representation of clean labels, and then starts memorizing noisy labels. Based on this finding, we propose novel methods for WSML which reject or correct the large loss samples to prevent model from memorizing the noisy label. Without heavy and complex components, our proposed methods outperform previous state-of-the-art WSML methods on several partial label settings including Pascal VOC 2012, MS COCO, NUSWIDE, CUB, and OpenImages V3 datasets. Various analysis also show that our methodology actually works well, validating that treating large loss properly matters in a weakly supervised multi-label classification. Our code is available at https://github.com/snucml/LargeLossMatters.

Coded computing has proved its efficiency in handling a straggler issue in distributed computing framework. However, in a coded distributed computing framework, there may exist Byzantine workers who send the wrong computation results to a master to contaminate the overall computation output. Therefore, it is essential to identify Byzantine workers from their computation results in coded computing. In this paper, we consider Byzantine attack identification problem in coded computing for distributed matrix multiplication tasks. We propose locally testable codes which facilitate the efficient Byzantine attack identification, and suggest a hierarchical group testing method for Byzantine attack identification. We show that our scheme requires smaller number of tests than the conventional group testing methods for the existing coded computing schemes.

Coded computing has proved its efficiency in handling a straggler issue in distributed computing framework. It uses error correcting codes to mitigate the effect of the stragglers. However, in a coded distributed computing framework, there may exist Byzantine workers who send the wrong computation results to a master in order to contaminate the overall computation output. Therefore, it is essential to identify Byzantine workers from their computation results in coded computing. In this paper, we consider Byzantine attack identification problem in coded computing for distributed matrix multiplication tasks. We propose a new coding scheme which facilitates the efficient Byzantine attack identification, namely locally testable codes. We also suggest a hierarchical group testing method for Byzantine attack identification. We claim the required number of tests for group testing in our scheme, and show that it requires smaller number of tests than the conventional group testing method for the existing coded computing schemes.

Distributed computing has been a prominent solution to efficiently process massive datasets in parallel. However, the existence of stragglers is one of the major concerns that slows down the overall speed of distributed computing. To deal with this problem, we consider a distributed matrix multiplication scenario where a master assigns multiple tasks to each worker to exploit stragglers’ computing ability (which is typically wasted in conventional distributed computing). We propose Chebyshev polynomial codes, which can achieve order-wise improvement in encoding complexity at the master and communication load in distributed matrix multiplication using task entanglement. The key idea of task entanglement is to reduce the number of encoded matrices for multiple tasks assigned to each worker by intertwining encoded matrices. We experimentally demonstrate that, in cloud environments, Chebyshev polynomial codes can provide significant reduction in overall processing time in distributed computing for matrix multiplication, which is a key computational component in modern deep learning.

In this paper, we consider coded computation for matrix multiplication tasks in distributed computing, which can mitigate the effect of slow workers, called stragglers, by a coding approach. We assume that the stragglers' computation results can be leveraged at the master by assigning multiple sub-tasks to the workers. In this scenario, we propose a new coded computation scheme to preserve the data privacy and security from the non-colluding workers. We also prove that the data privacy and security constraints are satisfied in our scheme in an information-theoretic sense.

Wireless power transfer is a promising solution for wireless networks composed of Internet-of-Things (IoT) devices that may suffer from insufficient battery capacity. A wireless powered communication network (WPCN) is a framework to design energy-constrained networks such as IoT networks. In this article, we consider WPCNs consisting of a hybrid access point (H-AP) and energy harvesting users, all equipped with multiple antennas. A H-AP transfers energy to the users using energy beamforming in the downlink, and the users transmit information using the harvested energy in the uplink, where nonorthogonal multiple access (NOMA) transmission is employed. For the uplink NOMA transmission, users are grouped into multiple clusters by cluster-specific beamforming. In particular, signal alignment is exploited for the beamforming so that the channels of users in a cluster are aligned in the same direction. By signal alignment, the number of messages decoded by successive interference cancellation (SIC) is reduced, which can be effective at lowering the decoding complexity and SIC error propagation. Due to the difficulty of jointly optimizing cluster-specific beamforming and time/energy resources for sum-throughput maximization, we determine the beamforming relying on signal alignment first, and then the resources are optimized for given beamforming. To be more specific, we propose a novel iterative algorithm for cluster-specific beamforming design followed by the sum-throughput maximization algorithm. Numerical results show the sum-throughput performance of the proposed scheme and its robustness toward SIC error propagation compared to existing schemes.

This article considers a power allocation problem in energy harvesting downlink non-orthogonal multiple access (NOMA) systems in which a transmitter sends desired messages to their respective receivers by using harvested energy. To tackle this problem, we make use of a reinforcement learning approach based on a shallow neural network structure. We prove that the optimal power allocation policy and the optimal action-value function depend monotonically on some of their input variables and the shallow neural network structure is designed based on properties revealed in the proof. Different from inefficient deep learning methods that tend to require tremendous computational resources, this structure is capable of fully capturing the characteristics of the desired function with a single hidden layer. The optimized structure also allows learning agents to be robust and highly reliable in learning about randomly occurring data. Furthermore, we provide comprehensive experimental results in harsh environments where various arbitrary factors are assumed in order to demonstrate the robustness of the proposed learning approach compared with deep neural networks without proper grounds. It is also shown that the proposed learning process converges to a policy that outperforms existing power allocation algorithms.

Radar is a key component of autonomous driving. We can capture the target range and velocity using radar signal that is transmitted and reflected by a target. However, the noise floor increases when interference signals exist, which severely affects the detectability of target objects. For these reasons, many previous studies have proposed methods for cancelling interference or reconstructing original signals. However, little work has been done using deep learning algorithm on interference mitigation. In this paper, we propose a new method using deep learning. We improve the performance of the existing deep learning algorithm using attention mechanism. We applied our algorithm to the OFDM radar environment as well as the existing frequency modulated continuous wave(FMCW) radar. The experiment show our deep learning based method outperforms existing methods. The implementation of our paper is available at https://github.com/jwmun/Radar.

This paper deals with the power allocation problem for achieving the upper bound of sum-rate region in energy harvesting downlink channels. We prove that the optimal power allocation policy that maximizes the sum-rate is an increasing function for harvested energy, channel gains, and remaining battery, regardless of the number of users in the downlink channels. We use this proof as a mathematical basis for the construction of a shallow neural network that can fully reflect the increasing property of the optimal policy. This scheme helps us to avoid using big neural networks which requires huge computational resources and causes overfitting. Through experiments, we reveal the inefficiencies and risks of deep neural network that are not optimized enough for the desired policy, and shows that our approach learns a robust policy even with the severe randomness of environments.

In distributed computing system for the master-worker framework, an erasure code is able to mitigate the effects of slow workers, also called stragglers. The distributed computing system combined with coding is referred to as coded computation. For a matrix multiplication, we consider a variation of coded computation that ensures the master's privacy from the workers, which is referred to as private coded matrix multiplication. In the private coded matrix multiplication, the master needs to compute a matrix multiplication on its own matrix and one of the matrices in a library exclusively shared by the external workers. After the master recovers the matrix multiplication through coded matrix multiplication, the workers should not know which matrix in the library was desired by the master, which implies that the master's privacy is ensured. Our problem is a special case of linear private computation, where a linear combination of matrices in the library should be concealed. We propose a private coded matrix multiplication scheme, based on the conventional coded matrix multiplication scheme. In terms of computation time and communication load, we compare our proposed scheme with a conventional robust private information retrieval scheme and private computation schemes.

Massive machine-type communications (mMTC) or massive Internet of things (IoT) is one of the key application scenarios of fifth generation (5G) and beyond cellular networks. Since initial access procedures of legacy systems are not suitable for massive connectivity due to increased collision probability and prohibitive overhead, it is of great interest to design an efficient grant-free access protocol. In this paper, we propose an uncoordinated access protocol for massive connectivity, which leverages a massive number of antennas at the base station (BS), which is expected to be widely deployed in cellular networks. With the proposed scheme consisting of a sparse frame structure and receiver processing based on sparse dictionary learning, a massive number of IoT devices can transmit data without any prior scheduling process. Unlike existing schemes that necessitate significant overhead for preamble signals, the overhead required for the proposed scheme is negligible, which is attractive in terms of resource utilization and transmit power consumption. Numerical results verify that the proposed scheme can be utilized to facilitate massive connectivity in realistic cellular-based IoT scenarios.

In this paper, we consider a power allocation optimization technique for a time-varying fading broadcast channel in energy harvesting communication systems, in which a transmitter with a rechargeable battery transmits messages to receivers using the harvested energy. We first prove that the optimal online power allocation policy for the sum rate maximization of the transmitter is an increasing function of harvested energy, remaining battery, and each user's channel gain. We then construct an appropriate neural network by relying on increasing behavior of the optimal policy. This two-step approach, by using an effective function approximation as well as providing a fundamental guideline for neural network design, can prevent us from wasting the representational capacity of neural networks. On the basis of the neural network, we apply the policy gradient method to solve the power allocation problem. To validate the performance of our approach, we compare it with the closed-form the optimal policy in a partially observable Markov problem. Through further experiments, it is observed that our online solution achieves a performance close to the theoretical upper bound of the performance in a time-varying fading broadcast channel.

We introduce a variation of coded computation that protects the data security and the master's privacy against the workers, which is referred to as private secure coded computation. In private secure coded computation, the master needs to compute a function of its own dataset and one of the datasets in a library exclusively shared by the external workers. After recovering the desired function, the workers should not know which dataset in the library was desired by the master or obtain any information about the master's own data. We propose a private secure coded computation scheme for matrix multiplication, namely private secure polynomial codes, based on private polynomial codes for private coded computation. By simulation, we show that the private secure polynomial codes achieves better computation time than private polynomial codes modified for private secure coded computation.

In this paper, we consider a secure distributed computing scenario in which a master wants to perform matrix multiplication of confidential inputs with multiple workers in parallel. In such a setting, a master does not want to reveal information about the two input matrices to the workers in an information-theoretic sense. We propose a secure distributed computing scheme that can efficiently cope with straggling effects by applying polynomial codes on sub-tasks assigned to workers. The achievable recovery threshold, i.e., the number of workers that a master needs to wait for to get the final product, of our proposed scheme is revealed to be order-optimal to the number of workers. Moreover, we derive the achievable recovery threshold of the proposed scheme is within a constant multiplicative factor from information-theoretic lower bound. As a byproduct, we extend our strategy to secure distributed computing for convolution tasks on confidential data.

In this paper, we consider a power allocation problem for energy harvesting communication systems, where a transmitter wants to send the desired messages to the receiver with the harvested energy in its rechargeable battery. We propose a new power allocation strategy based on deep reinforcement learning technique to maximize the expected total transmitted data for a given random energy arrival and random channel process. The key idea of our scheme is to lead the transmitter, rather than learning the undesirable power allocation policies, by an action-bounding technique using only causal knowledge of the energy and channel processes. This technique helps traditional reinforcement learning algorithms to work more accurately in the systems, and increases the performance of the learning algorithms. Moreover, we show that the proposed scheme achieves better performance with respect to the expected total transmitted data compared to existing power allocation strategies.

In automotive systems, a radar is a key component of autonomous driving. Using transmit and reflected radar signal by a target, we can capture the target range and velocity. However, when interference signals exist, noise floor increases and it severely affects the detectability of target object. For these reasons, previous studies have been proposed to cancel interference or reconstruct original signals. However, the conventional signal processing methods for canceling the interference or reconstructing the transmits signal are difficult tasks, and also have many restrictions. In this work, we propose a novel approach to mitigate interference using deep learning. The proposed method provides high performance in various interference conditions and has low processing time. Moreover, we show that our proposed method achieves better performance compared to existing signal processing methods.

In this paper, we investigate a private information retrieval (PIR) problem for secure distributed storage systems in the presence of an eavesdropper. We design the secure distributed database and the corresponding PIR scheme, which protect not only user privacy (concealing the index of the desired message) from the databases, but also data security (concealing the messages themselves) from an eavesdropper. In our proposed scheme, we use a secret sharing scheme in storing the messages for data security at each of the databases. We consider two different scenarios on whether the databases are aware of the index sets of the secret shares stored in other databases. The key idea in designing an efficient PIR procedure is to exploit the secret shares of undesired messages as a side information by means of storing the secret shares at multiple databases. In particular, it is shown that the rates of the proposed PIR schemes are within a constant multiplicative factor from the derived upper-bound on the capacity of PIR problem.

When a node in a distributed storage system fails, it needs to be promptly repaired to maintain system integrity. While typical erasure codes can provide a significant storage advantage over replication, they suffer from poor repair efficiency. Locally repairable codes (LRCs) tackle this issue by reducing the number of nodes participating in the repair process (locality), at the cost of reduced minimum distance. In this paper, we study the tradeoff between locality and minimum distance of LRCs with local codes that have arbitrary distance requirements. Unlike existing methods, where both the locality and the local distance requirements imposed on every node are identical, we allow the requirements to vary arbitrarily from node to node. Such a property can be an advantage for distributed storage systems with non-homogeneous characteristics. We present Singleton-type distance upper bounds and also provide an optimal code construction with respect to these bounds. In addition, the feasible rate region is characterized by dimension upper bounds that do not depend on the distance.

A Gaussian multiple-input multiple-output wiretap channel in which the eavesdropper and legitimate receiver are equipped with arbitrary numbers of antennas and the transmitter has two antennas is studied in this paper. It is shown that the secrecy capacity of this channel can be achieved by linear precoding. The optimal precoding and power allocation schemes achieving the secrecy capacity are developed subsequently, and the secrecy capacity is compared with the generalized singular value decomposition (GSVD)-based precoding, which is the best previously proposed precoding for this problem. Numerical results show that substantial gain can be obtained in secrecy rate between the proposed and GSVD-based precodings.

A new relay-aided non-orthogonal multiple access (NOMA) technique is proposed for multi-cell uplink cellular networks in which each cell supports K single-antenna users by its respective base station (BS) equipped with N(4Z ≪K) antennas. Cooperative relaying transmission is used to accommodate more than one user per orthogonal resource block in the context of interference-limited cellular networks. With the proposed relayaided NOMA, an Alamouti structure of the desired symbol can further be generated at each BS free of interference, which gives rise to diversity gain of two. The proposed scheme does not require any channel state information (CSI) at users. Further, only limited CS! is required at relays and BSs, which can greatly reduce the control overhead.

There is much discussion in industry and academia about possible technical solutions to address the growth in demand for wireless broadband. Massive multiple-input multiple-output (MIMO) systems are one of the most popular solutions to addressing this broadband demand in fifth generation (5G) cellular systems. Massive MIMO systems employ tens or hundreds of antennas at the base station to enable advanced multiuser MIMO communications. To reap the massive MIMO throughput gain, coherent transmission exploiting accurate channel state information at the transmitter is required. While it is expected that many 5G systems will employ frequency division duplexing (FDD), channel sounding for FDD systems requires a large pilot overhead, which usually scales proportionally to the number of transmit antennas. To resolve this problem, a compressed sensing (CS)-aided channel estimation scheme is proposed, which exploits the observation that the channel statistics change slowly in time. By utilizing a conventional least squares approach and a CS technique simultaneously, the proposed scheme reduces the pilot overhead. Simulation results show that the proposed scheme can estimate the channel with a reduced pilot overhead even when conventional CS cannot be applied.

This paper studies full-duplex (FD) cellular networks in which a base station (BS) operated in FD mode with multiple antennas supports multiple uplink and downlink users simultaneously in the same wireless channel. Two typical FD cellular scenarios are considered, one with half-duplex (HD) users and the other with FD users along with the FD BS. For both the cases, a novel constructive method is developed for finding a closed-form interference alignment (IA) solution, namedcyclic IA. The core idea behind this approach is to construct a set of loop-equations enabling IA in acyclicmanner, so that beamforming vectors are sequentially determined by solving an eigenvalue problem. It is shown analytically that the proposed cyclic IA can achieve theoptimalsum degrees-of-freedom (DoF) when the number of user antennas is large enough to meet the derived conditions. In particular, it is shown that the proposed scheme achieves a twofold DoF gain compared with conventional HD cellular networks even in the presence of inter-link interference, provided the number of users becomes large enough compared with the ratio of the number of BSs and user antennas. Simulation results demonstrate that not only are the analytical DoF results valid, but under a practical multi-cell scenario, the proposed cyclic IA offers significant throughput gains depending on the cell radius.

We study the symmetric degrees-of-freedom (DoF) of partially connected interference networks under linear coding strategies without channel state information at the transmitters beyond topology. We assume that the receivers are equipped with reconfigurable antennas that can switch among their preset modes. In such a network setting, we characterize the class of network topologies in which half linear symmetric DoF is achievable. Moreover, we derive two general upper bounds on the linear symmetric DoF for arbitrary network topologies. We also demonstrate the tightness of our bounds for the class of network topologies in which all the transmitters in the network have at most two co-interferers.

Cooperative regenerating codes are a kind of erasure codes, which are optimal in terms of minimizing the repair bandwidth. An (n, k, d, r)-cooperative regenerating code has n storage nodes, where k arbitrary nodes are enough to reconstruct original data, and r failed nodes can be repaired cooperatively with the help of d arbitrary surviving nodes. In the regenerating-code framework, there exists a tradeoff between the storage capacity of each node α and the repair bandwidth y, but the problem of specifying the optimal storage-bandwidth tradeoff of the exact-repair cooperative regenerating codes remains open. A key contribution of this paper is that an outer bound on the storage-bandwidth tradeoff of exact-repair linear cooperative regenerating codes is proposed. This result can be regarded as a generalization of the outer bound proposed by Prakash et al., which specifies the optimal tradeoff of exact-repair regenerating codes for the case of d = k = n - 1. The proposed outer bound suggests the (α, γ) pairs that no exact-repair codes can achieve but only functional-repair codes can. By observing the size of the set of such (α, γ) pairs, the performance of the proposed outer bound is evaluated under various parameter settings.

In this paper, we characterize the degrees of freedom (DoF) for K-user M × 1 multiple-input single-output interference channels with reconfigurable antennas, which have N-preset modes at the receivers, assuming linear coding strategies in the absence of channel state information at the transmitters, i.e., blind interference alignment. Our linear DoF converse builds on the lemma that if a set of transmit symbols is aligned at their common unintended receivers, those symbols must have independent signal subspace at their corresponding receivers. This lemma arises from the inherent feature that channel state's changing patterns of the links towards the same receiver are always identical, assuming that the coherence time of the channel is long enough. We derive an upper bound for the linear sum DoF, and propose an achievable scheme that exactly achieves the linear sum DoF upper bound when both of the n*/M = R1and MK/n* = R2are integers, where n* denotes the optimal number of preset modes out of N preset modes. For the other cases, where either R1or R2is not an integer, we only give some guidelines how the interfering signals are aligned at the receivers to achieve the upper bound. As an extension, we also show the linear sum DoF upper bound for downlink/uplink cellular networks.

In this paper, we characterize the linear degrees of freedom (DoF) of a cellular network in which the base station (BS) operates in a full-duplex (FD) mode and the users operate in a half-duplex mode. We assume that the BS and the users are equipped with reconfigurable antennas which can be switched between their preset modes. We consider two practical scenarios for different assumptions on channel state information at the transmit sides (CSIT), referred to as no CSIT and partial CSIT models. To derive the inner-bounds for two scenarios, we propose a new achievable scheme which enables interference alignment between uplink and downlink interference signals at each user via preset mode switching of reconfigurable antennas. The key concept of our scheme is to align the interference signals of uplink transmission at the downlink users, through the identical preset mode pattern over the multiple of downlink transmission periods and silence periods of the BS. We also develop an outer-bound on the linear sum DoF of the cellular network for the no CSIT model, which matches up with the inner-bound. Moreover, we also provide a natural variant of the proposed scheme when considering residual self-interference at the FD BS, which can alleviate the shortcoming of the existing self-interference cancellation techniques.

In this paper, we propose a novel relay-aided successive aligned interference cancellation (RaSAIC) method for a 2 × K X channel with L full-duplex relays. The key idea behind the RaSAIC is to apply a cooperative relays' precoding technique that aligns interference signals at unintended receivers while creating an Alamouti structure for the desired signals. With the proposed cooperative relay transmission technique, it is shown that both the optimal degrees of freedom gain of 2K/K+1 and the diversity gain of two are achievable for constant channels when the numbers of relays' antennas are enough to satisfy the derived feasibility conditions without requiring channel state information (CSI) at transmitters but with global and local CSI at relays and receivers, respectively.

In millimeter wave (mmWave) communication systems, large antenna arrays are used to compensate high path loss. While the large array provides high beamforming gain, it also poses a challenge in channel estimation. Since mmWave channels are likely to be sparse in angular domain, the channel estimation can be converted into a sparse recovery problem, and compressed sensing (CS) can be leveraged for the channel estimation. However, conventional non-adaptive CS algorithms show poor recovery performance with low signal-to-noise ratio (SNR), which is common before beamforming in mmWave channels. Although recently developed adaptive CS schemes perform better in a low SNR regime, their excessive feedback requirement hinders practical usage. In this paper, we propose a two-stage CS scheme that requires one-time feedback and is robust to noise, which can be regarded as a compromise between the two approaches. Sufficient conditions for the support recovery with the proposed scheme are characterized, and its effectiveness is also shown numerically.

(n, k, d, r)-cooperative regenerating codes are a kind of erasure codes, where r failed nodes can be repaired cooperatively with the help of arbitrary d surviving nodes. In this paper, we consider exact-repair cooperative regenerating codes whose parameters satisfy k = d = n - r. There exists a tradeoff between the storage capacity of each node α and the repair bandwidth γ in regenerating codes, but the optimal storage-bandwidth tradeoff of the exact-repair cooperative regenerating codes has not been fully specified. We propose an outer bound on the storage-bandwidth tradeoff for the case of k = d = n - r. This result can be regarded as a generalization of the outer bound proposed by Prakash et al. that specifies the optimal tradeoff of exact-repair regenerating codes for the case of k = d = n - 1. Although the proposed outer bound is not always tighter than the cutset bound of the functional repair model, in the cases where n is large and r is small, the proposed outer bound suggests the region on the α-γ plane that no exact-repair codes can achieve, but functional-repair codes can.

This paper proposes a novel multi-antenna relay-aided interference management technique that can use imperfect channel knowledge for interference relay channels. Using the proposed method, it is shown that KM/K+M-1 degrees of freedom (DoF) are achievable in a K-user multiple-input-single-output interference relay channel when the relay has M antennas with a certain type of limited channel knowledge. By leveraging this result, it is demonstrated that the interference-free DoF of K are asymptotically achieved as M approaches infinity. One major implication of these results is that even under this limited channel knowledge, the use of massive antennas at the relay is sufficient to recover the optimal DoF for relay-aided interference networks with perfect channel knowledge.

We study the symmetric degrees-of-freedom (DoF) of partially connected interference networks under linear coding strategies without channel state information at the transmitters beyond topology. We assume that the receivers are equipped with reconfigurable antennas that can switch among their preset modes. In such a network setting, we characterize the class of network topologies in which half linear symmetric DoF is achievable. Moreover, we derive a general upper bound on the linear symmetric DoF for arbitrary network topologies. We also show that this upper bound is tight if the transmitters have at most two co-interferers.

The storage-bandwidth tradeoff of linear and exact-repair regenerating codes is considered. An outer bound on the storage-bandwidth tradeoff is proposed for arbitrary (n, k, d) case. n, k, and d denote typical code parameters of regenerating codes. The proposed outer bound becomes tighter than other existing outer bounds as k gets closer to n. In addition, for the case of d = n − 1, it is shown that the proposed outer bound asymptotically meets the inner bound proposed by Tian et al., which implies the proposed bound converges to the optimal storage-bandwidth tradeoff of exact-repair linear regenerating codes in high rates (k/n ≅ 1).

We consider M x 2 X channels with M+1 antennas at each node. The proposed scheme achieves both the optimal degrees of freedom (DoF) of 2M for the considered channel and the diversity order of 2 simultaneously. The key to achieving them is to design beamformers using interference alignment (IA) scheme with Alamouti codes. Furthermore, the proposed scheme requires 2 symbol extensions. We assume that each transmitter has local channel state information at transmitter (CSIT) and each receiver has full channel state information at receiver (CSIR).

In this paper, we propose a new retrospective interference alignment for two-cell multiple-input multiple-output (MIMO) interfering multiple access channels (IMAC) with the delayed channel state information at the transmitters (CSIT). It is shown that having delayed CSIT can strictly increase the sum-DoF compared to the case of no CSIT. The key idea is to align multiple interfering signals from adjacent cells onto a small dimensional subspace over time by fully exploiting the previously received signals as side information with outdated CSIT in a distributed manner. Remarkably, we show that the retrospective interference alignment can achieve the optimal sum-DoF in the context of two-cell two-user scenario by providing a new outer bound.

We develop a blind interference alignment (BIA) through staggered antenna switching scheme for more realistic channel assumption. Contrary to the general assumption that the coherence time of the channel is long enough to perform BIA, when the coherence time is not long enough, channel coefficients stay constant for limited block length. Therefore, we propose a dynamic supersymbol design algorithm which can construct a supersymbol with limited block length which is determined by the coherence time of the channel. We demonstrate that the supersymbol length can be reduced significantly by aligning interferences in a hierarchical manner referred to as hierarchical BIA. Furthermore, we also show that the proposed dynamic supersymbol design algorithm achieves higher degrees of freedom than the conventional method with a given coherence time.

Due to considerable increases in user mobility and frame length through aggregation, the wireless channel remains no longer time-invariant during the (aggregated) frame transmission time. However, the existing IEEE 802.11 standards still define the channel estimation to be performed only once at the preamble for coherent OFDM receivers, and the same channel information to be used throughout the entire (aggregated) frame processing. Our experimental results reveal that this baseline channel estimation approach seriously deteriorates the WiFi performance, especially for pedestrian mobile users and the recently adopted frame aggregation scheme. In this paper, we propose Channel-Aware Symbol Error Reduction (ChASER), a new practical channel estimation and tracking scheme for WiFi receivers. ChASER utilizes the re-encoding and re-modulation of the received data symbol to keep up with the wireless channel dynamics at the granularity of OFDM symbols. Our extensive, trace-driven link-level simulation shows significant performance gains over a wide range of channel conditions based on the real wireless channel traces collected by the off-the-shelf WiFi device. In addition, the feasibility of its low-complexity and standard compliance is demonstrated by Microsoft's Software Radio (Sora) prototype implementation and experimentation. To our knowledge, ChASER is the first IEEE 802.11n-compatible channel tracking algorithm since other approaches addressing the time-varying channel conditions over a single (aggregated) frame duration require costly modifications of the IEEE 802.11n standard.

We propose a blind interference alignment (BIA) through staggered antenna switching scheme with no ideal channel assumption. Contrary to the ideal assumption that channels remain constant during BIA symbol extension period, when the coherence time of the channel is relatively short, channel coefficients may change during a given symbol extension period. To perform BIA perfectly with realistic channel assumption, we propose a grouping based supersymbol structure for K-user interference channels which can adjust a supersymbol length to given coherence time. It is proved that the supersymbol length could be reduced significantly by an appropriate grouping. Furthermore, it is also shown that the grouping based supersymbol achieves higher degrees of freedom than the conventional method with given coherence time.

In this paper, we propose a new retrospective interference alignment for two-cell multiple-input multiple-output (MIMO) interfering multiple access channels (IMAC) with M receive antennas per base station (BS) and K users per cell each with N transmit antennas. The key idea of the retrospective IA lies in interference-purification during phase 1 and 2, and interference-repetition during phase 3. On the basis of the proposed method, we show that the achievable sum degrees of freedom (sum-DoF) is equal to 2M ([2M/N]+ 1) / (2 [2M/N] + 1) if K M <; NK and min{2KN, M} if NK ≤ M with completely delayed channel state information at the transmitters (CSIT). It is shown that having delayed CSIT can strictly increase the sum-DoF compared to the case with no CSIT for cellular networks, especially for uplink scenarios.

A novel interference alignment technique combined with interference cancellation is proposed. A new scenario of single-antenna 2-user X channel with a multiple-antenna relay is considered. The proposed interference alignment and cancellation scheme does not require any wireline links between receivers. In the proposed scheme, interference signals are not decoded. Instead, interference signals are just aligned, and the aligned signal is cancelled out to extract the desired signal. We call the proposed scheme an aligned interference cancellation technique. It is known that DoF of 4/2 is achievable for the two-user X channel using the interference alignment. In the proposed scheme, relay precoding matrices are designed to accomplish two purposes. One is to cancel out the aligned interference signals perfectly, and the other is to enable an equivalent received signal model of the Alamouti's codes. The proposed relay-aided interference alignment and cancellation technique not only makes interference alignment feasible, but also improves performance. The interference signals are aligned at a specific time slot, and the precoding matrices of the relay are designed to cancel the aligned interference. In simulations, it is verified that the diversity gain of 2 is achieved with the degrees-of-freedom (DoF) of 4/3.

Spectrum sensing is a critical component in cognitive radio. Meanwhile, robust principal component analysis (rPCA) can decompose a matrix into low-rank and sparse matrices. In general, the covariance matrix of a correlated signal is low-rank and the covariance matrix of white noise is diagonal, which can be regarded as sparse. This fact implies that rPCA can be used as a powerful tool for spectrum sensing. A novel spectrum sensing technique which utilizes the characteristics of covariance matrices and rPCA is proposed in this paper. The proposed scheme is also compared to existing schemes based on sample covariance matrices by simulations.

The cognitive radio (CR) system in which multiple secondary users (SU) search for spectrum opportunities generated by the absence of primary users (PU) is considered in this paper. The occupancy of a CR channel is modeled as a Markov chain, and it is assumed that the Markov chain has only two states: idle or busy. Since parameters of the Markov chain are unknown to SUs \emph{a priori} and the states transit independently of the sensing and utilization of SUs, this problem can be considered as a kind of RMAB (restless multi-armed bandit) problem. We propose an efficient channel allocation algorithm for SUs, which is constructed through combination of multiple single-user MAB policies. When a performance of the proposed algorithm is measured by regret which is defined as the total reward difference from the ideal Bayesian policy in which the stationary probability is known to SUs, the order of regret growth of the proposed algorithm seems to be negatively decreasing, giving a better performance than any other existing policy under the 2-state Markov chain case. In order to estimate the performance of the proposed algorithm appropriately, we introduce a new definition of the regret, which uses a belief vector based Bayesian policy as the ideal policy. We observe experimentally that the order of the newly defined regret in the proposed algorithm is similar to logarithmic order under certain conditions.

In this paper, we derive the exact probability distribution of post-scheduling signal-to-interference-plus-noise ratio (SINR) considering both user feedback and scheduling. We also develop an optimized adaptive modulation scheme in orthogonal random beamforming systems with M transmit antennas and K single-antenna users. The exact probability distributions of each user's feedback SINR and the exact postscheduling SINR are derived rigorously by direct integration and multinomial distribution. It is also shown that the derived cumulative distribution function (CDF) of the post-scheduling SINR happens to be identical to the the existing approximate CDF for SINR higher than 0 dB. The closed form expressions of system performance, such as average spectral efficiency (ASE) and average bit error ratio (A-BER), are derived using the CDF of the post-scheduling SINR. The optimal SINR thresholds that maximize the ASE with a target A-BER constraint are solved using the derived closed form CDF and a Lagrange multiplier. Key contributions of this paper include the derivation of the exact CDF of post-scheduling SINR by direct integration, and its application to an optimized adaptive modulation based on a Lagrange multiplier. Simulations show the correspondence between theoretical and empirical CDF's, and the performance improvement of the proposed adaptive modulation method in terms of ASE.

Adaptive modulation is a link adaptation technique that dynamically adjusts modulation level and code rate to the time-varying channel conditions in order to obtain considerably high data rates. In this paper, we propose a new adaptive modulation algorithm in which an modulation and coding scheme level is selected by the received SNR plus the offset value obtained from the transmission results. Simulations, which are done in the IEEE 802.11n environment, show that the proposed algorithm performs better than the well-known adaptive modulation algorithms such as auto rate fallback and general SNR-based techniques. Particularly, the proposed algorithm has more robust throughput performance while satisfying the target packet error ratio for fast fading channels.

This paper presents a programmable capacitive sensing circuit for multi-touch applications, especially enabling mutual capacitance sensing in a large-sized touch panel (over 10-inches). To control the offset of the capacitive sensor, the circuit, based on a switched-capacitor integrator, internally includes a cancelling capacitor array to discharge a feedback capacitor and prevent excessive accumulation. Also, the feedback capacitor array can make the sensitivity of the sensor programmable. With programmable capacitors, this design can make the sensor more compatible with various panels and use a lower feedback capacitor than that of a conventional switched-capacitor. The capacitive interface circuit for a large-sized touch panel has 1728 channels (54×32), which simultaneously sense the capacitance variation during each line driving for mutual capacitance sensing. In addition, charge integration is repeated to increase the signal-to-noise ratio (SNR). The performance of the proposed topology is evaluated mainly with a 10.1-inch LCD touch panel. The measured SNR of the sensor is 23dB at a 60Hz frame rate with twenty repetitions of charge integration. The prototype IC is implemented in a 0.35-um CMOS technology.

In this paper, we propose a physical layer network coding technique combined with asynchronous code division multiple access (CDMA). We consider a scenario where there are multiple pair of nodes with a single relay, and each pair of nodes exchange information asynchronously. With the assumption of asynchronous transmission, there is a trade-off between the number of nodes and the inter-user interference because the spreading codes are not orthogonal to each other. The BER and the throughput of the system are analyzed, and simulations are performed to verify the results. It is shown that physical layer (analog) network coding improves throughput compared to the conventional routing protocol with large Eb/N0, low number of users, and large spreading factor.

In conventional multi-user MIMO systems, Shannon capacity has been used mostly as the performance measure for user selection. In this paper, we propose two new MIMO scheduling techniques based on BER instead of capacity as the performance measure for rate maximization with target BER. One of the key contributions of this paper is to use BER instead of capacity as the user selection metric, and another is the novel power allocation techniques considering user scheduling for the rate maximization strategy. Simulation results show that the proposed BER based algorithms produce higher rate compared to the conventional capacity based user selection algorithm with much lower computational complexity as well as the schemes meet target BER.

It has been well known that channel aware systems (closed-loop) have better performance than channel unknown systems (open-loop). In order to know channel in the transmitter, receivers should send own channel information the transmitter through reverse channel. However, the channel information should be fed back in Frequency Division Duplexing (FDD) additionally since forward link channel and reverse link channel are different each other. In FDD, the channel information is usually quantized as a codebook and each receiver feedback the best codeword index within the codebook by any metric generally. The codebook has the characteristic that as codebook size is larger system performance increases but computational complexity to find the best codeword also increases exponentially in conventional codebook structure. In this paper, we propose some codebook design schemes which have hierarchical structure to overcome the complexity problem as well as achieves near-optimal performance.

The multiuser MIMO (MU-MIMO) systems tend to have inter-user interference in practical systems with limited feedback. Practical MU-MIMO systems with limited feedback can be divided into two classes in general. One is zero-forcing beamforming (ZFBF), and the other is per user unitary rate control (PU2RC) which is an extended version of random beamforming (RBF). To improve performance, ZFBF needs more feedback bits, and PU2RC needs a larger number of users. We propose a new hybrid MU-MISO system which has the advantages of the two MU-MIMO schemes simultaneously, and analyze the sum-rate performance and the quantization error of the hybrid scheme. The analysis shows how the number of feedback bits, the number of users, and the signal power affect the sum-rate. The throughput scaling laws are also derived in the high and the medium SNR regimes. Simulation results show that the proposed hybrid MU-MISO system performs better than either of the two MU-MIMO systems with the same number of feedback bits. Another advantage is that it has lower complexity for codeword search since it has a hierarchical codebook structure.

Multiuser MIMO (MU-MIMO) systems have advantages over single-user MIMO systems in terms of system performance. In MU-MIMO systems, inter-user interference needs to be dealt with especially when linear processing is used. The block diagonalization (BD) method is one of techniques that are widely used to eliminate the inter-user interference. In a cellular system where there are many users, the subset of users which maximizes the system performance should be selected since the base station cannot support all the users in the cell. In this paper, we propose a low complexity MU-MIMO scheduling scheme using BD with chordal distance. For a large number of users, the optimal scheduling technique needs an exhaustive search, which is impractical. One of the key ideas of this paper is to use chordal distance as a measure of orthogonality between different users. Simulation results show the proposed algorithm has throughput close to the optimal scheduling scheme with lower complexity than existing low complexity scheduling algorithms.

Multi-antenna downlink system with limited feedback can provide enormous capacity gains. In practical, two systems can be considered. One is zero-forcing beamforming (ZFBF), and the other is per user unitary rate control (PU2RC) which has been proposed for emerging cellular standards. In order to have good performance, ZFBF requires many feedback bits, and PU2RC requires many users. Recently, the authors proposed a hybrid multi-user MISO system which has merits of those two schemes simultaneously. In this paper, we analyze the sum-rate performance and quantization error of the hybrid scheme. The throughput scaling laws are also derived in high and medium SNR regimes.

This paper analyzes the optimal degrees of freedom (DoF) for K-user M × N MIMO interference channels with constant channel coefficients. In this paper, we use the methodology of interpreting the interference alignment problem as a multivariate polynomial system. The properness of the system enables us to determine the feasibility of interference alignment. Using this methodology, we propose an algorithm to obtain the optimal DoF, and compute the closed-form optimal DoF. We show that K min(M,N) is optimal DoF if δ ≥ min(M,N), where δ = ⌊M+N over K+1⌋. If δ < min(M,N), K M+N over K+1 is optimal DoF when M+N over K+1 is an integer, and ⌈K M+N over K+1⌉ − 1 is the optimal DoF when M+N over K+1 is not an integer. A key contribution of this paper is that it provides a closed-form optimal DoF for M × N MIMO interference channels when the DoF of each user can be different.

In a multiuser MIMO cellular system where there are many candidate users, it is critical to select a user group which maximizes the overall throughput of the system. However, the optimal scheduling strategy is computationally prohibitive when the total number of users is large. In this paper, we propose a determinant based user selection algorithm which reduces the search complexity without much performance degradation. In order to reduce pilot overhead by precoding matrix for multiuser MIMO, we also propose a new pilot scheme with only one set of pilot, and the new pilot scheme is combined with the proposed scheduling algorithm. Simulation results show that the proposed scheduling algorithm with a new pilot scheme reduces computational complexity and pilot overhead with negligible performance degradation compared to the exhaustive scheduling with a conventional pilot scheme which uses two sets of pilots.

The research on driverless cars has been making much progress lately. In this paper, we propose a new traffic control system without traffic lights at an intersection. We assume a system with fully autonomous driverless cars, and infrastructure to avoid collision completely. When automobiles approach an intersection, they communicates with the access point in both random access mode and polling mode, and the movement of the automobiles will be coordinated by the infrastructure (access point). Traffic congestion is very difficult to predict and deal with because it is a function of many unknown factors such as number of cars, weather, road constructions, accidents, etc. The proposed algorithm is designed for urban road networks to ease the congestion, and make it more predictable at the same time. A key idea of this paper is that IEEE 802.11 DCF/PCF mechanisms are used to control traffic flow for driverless cars when there are no traffic lights at an intersection. The algorithm utilizes the concept of contention/contention-free period of IEEE 802.11 to find a balance between efficiency of traffic flow and fairness between users.

In cellular systems, a basestation and mobile stations need to be synchronized before data exchange. Since the basestation clock reference is more accurate, a mobile station typically derives its clock reference from the basestation. But the carrier frequency offset due to Doppler shift may have harmful effects on the local clock derivation. This letter proposes a joint clock and frequency synchronization technique between a basestation and a mobile station, which is effective even with Doppler shift. We derive the joint estimation algorithm by analyzing the phase and the amplitude distortion caused by the sampling frequency offset and the carrier frequency offset. Simulation results showing the effectiveness of the proposed algorithm will also be presented.

In this letter, we analyze the post-scheduling SINR mismatch of a multiuser orthogonal random beamforming (ORBF) system. The mismatch probability is derived for a Gaussian broadcast channel with M transmit antennas and K single-antenna users. Numerical results are shown to be identical to theoretical results. The probability analysis enables us to calculate the minimum number of users to satisfy a target mismatch probability. Another way to use the mismatch probability is to control CQI mismatch handling mechanisms.

We consider a power allocation problem for three-node cooperative diversity systems using a relay node. Specifically, we focus on an amplify-and-forward (AF) protocol due to the simple operation at the relay node which makes the mathematical analysis tractable. This protocol consists of two phases. The source transmits the signal to both relay and destination in the first phase. In the second phase, the relay amplifies the received signal, and transmits to the destination. Depending on the source participation in the second phase, the protocol can be categorized into nonorthogonal AF (NAF) or orthogonal AF (OAF). In this paper, we consider the power allocation problem of the NAF protocol in a more general scenario than in previous literatures. We then show that there are only two modes of transmission which can be used if the optimal power allocation process is carried out. They correspond to the direct transmission and OAF with optimal power allocation. That is, the solution for optimal power allocation in the systems with NAF as a relay protocol excludes the nonorthogonal transmission. Instead, the selection between the direct transmission and OAF is optimal, i.e., using NAF protocol all the time is not optimal.

In this paper, we analyze channel quality indicator (CQI) mismatch probabilities of multiuser orthogonal random beamforming (ORBF) systems. First, we derive the CQI mismatch probability of the ORBF system with M transmit antennas and K users equipped with one receive antenna. Using that result, we also derive the upper and lower bounds of the per user unitary and rate control (PU2RC) system with 2 transmit antennas and K users equipped with one receive antenna, which is being adopted in the IEEE 802.16m standard. Finally, we compare theoretical results with numerical results from simulation.

We consider a multiple-input multiple-output (MIMO) system assisted by multiple non-regenerative single-antenna relays. This system where the antennas of all the relays are distributed geographically is fundamentally different from a single relay case with multiple antennas. The optimal power allocation for a relay with multiple antennas has already been known. In that solution, the power is distributed among the eigenmodes which are computed by the singular value decompositions of the first and the second hop channels. However, no closed form solution is known for the proposed system. In this paper, we propose a near optimal inter-relay power allocation method that combines the relay selection with the water-filling type closed form solution based on an asymptotic approach. In fact, the asymptotic approach itself is effective only for a specific antenna configuration, but we demonstrate through numerical simulations that if we combine this approach with the relay selection technique, the performance is comparable to the optimal case by exhaustive search even for a general antenna configuration.

This paper presents closed-form expressions of the distribution of signal-to-interference-plus-noise ratios (SINRs) and the average bit error rate (BER) for an orthogonal random beamforming system in multiple-antenna Gaussian broadcast channels. First, we derive the statistical distribution of the SINR of the scheduled transmit signals and confirm that the approximate form of the cumulative distribution function (cdf) of the SINR is equivalent to the formula in \cite{sharif} and \cite{letaief}. We then obtain a closed-form expression for the total average BER of the system by using the approximate cdf. Our theoretical results are compared to the Monte Carlo simulation results.

It has been shown that multiuser MIMO systems have better capacity performance than single-user MIMO systems. For a large number of users, the transmitter selects a subset of users with good channels to maximize the system performance. Shannon capacity has been used most of the time as the performance measure for multiuser MIMO systems. In this paper, we consider BER as an alternative performance metric, and propose two new user selection algorithms with block diagonalization (BD), which eliminates inter-user interference in a multiuser MIMO system. The first approach is BER minimization with fixed rate, and the second approach is rate maximization with fixed target BER. Simulation results show that the proposed algorithms produce lower BER and/or higher rate when they are compared to the conventional capacity based multiuser MIMO user selection algorithm with comparable computational complexity.

Multiuser MIMO (MU-MIMO) system with limited feedback can be divided into two classes generally. One is zero-forcing beamforming (ZFBF), and the other is per user unitary rate control (PU2RC) which is an extended technique of random beamforming (RBF). However, in order to perform well, ZFBF requires many feedback bits, and PU2RC requires many users. These drawbacks of the existing MU-MIMO systems limit system performance in real environment. In this paper, we propose a new MU-MISO system which has merits of two MU-MIMO schemes simultaneously. Simulation results show the new MU-MISO system has good performance with simple feedback and low complexity user selection.

We propose an improved minimum mean square error (MMSE) vertical Bell Labs layered space-time (V-BLAST) detection technique, called a soft input, soft output, and soft feedback (SIOF) V-BLAST detector, for turbo multi-input multioutput (turbo-MIMO) systems. We derive a symbol estimator by minimizing the power of the interference plus noise, given a priori probabilities of undetected layer symbols and a posteriori probabilities for past detected layer symbols. For a low-complexity implementation, an approximate SIOF algorithm is presented, which allows for a time-invariant realization of the symbol ordering and an MMSE filtering process. Another implementation, referred to as the iterative SIOF algorithm is introduced, which decides on symbol detection order based on a posteriori symbol probabilities to improve the detection performance. Simulations performed on a space-time bit-interleaved coded modulation (STBICM) architecture over quasi-static MIMO fading channels demonstrate that the SIOF V-BLAST detector provides performance gains over previous turbo-BLAST detectors, most notably when more transmit antennas are used.

Multiple-input multiple-output (MIMO) systems assisted by multiple relays with single antenna are considered. The signal transmission consists of two phases. In the first phase, the source node broadcasts the vector symbols to all relays, then all relays forward the symbols multiplied by gain parameters to the destination simultaneously. Unlike the case of full cooperation among relays such as a single relay with multiple antennas, in this case there may be no closed form optimal solution for the relay gains with respect to utility and/or cost functions due to the distributed relays. Therefore we propose an alternative approach in which we use the relay selection scheme along with conventional methods to find the relay gains. As a utility or cost function, we choose the determinant and trace of mean square error (MSE) matrix which are related to capacity and MMSE respectively. Further we combine this approach with a greedy search algorithm in order to reduce computational complexity. We provide simulation results to validate that the proposed methods incur a negligible performance loss compared to the optimal solution.

For a single relay channel, we study the average throughput of two different amplify-and-forward (AF) protocols, which are orthogonal-AF (OAF) and nonorthogonal-AF (NAF). The NAF protocol has been proposed to overcome a significant loss of performance of OAF in the high spectral efficiency region, and it was also shown that NAF performs better than OAF in terms of the diversity-multiplexing tradeoff. However, the results have been evaluated at the asymptotically high signal to noise ratio (SNR), and the power allocation problem between the source and the relay was neglected. We examine which protocol has a better performance in a practical system operating at a finite SNR. We also study where a relay should be located with consideration of the power allocation. In this paper, we show that the performance depends on both SNR and power allocation ratio, which indicates OAF may perform better in a certain environment.

This paper presents a new and simple decoding algorithm for layered space time block codes such as the two independent Alamouti's codes which are also called the double space-time transmit diversity (DSTTD) system. By using group interference suppression and successive interference cancellation, we can treat DSTTD as two independent space-time block codes (STBC). We can then decode both of these STBC's through a simple maximum likelihood (ML) detector with null space-based interference cancellation. We also compare the proposed interference cancellation (IC) scheme with the conventional MMSE IC scheme. The performance of the proposed IC scheme is comparable to that of the MMSE IC scheme while the complexity reduction factor of the proposed scheme can be up to 5 compared to the MMSE IC scheme.

In this paper, an improved minimum mean square error (MMSE) soft feedback detector, called the soft input, soft output, and soft feedback (SIOF) symbol detector, is proposed for turbo multi-input multi-output (TURBO-MIMO) systems. The SIOF symbol detector is derived by minimizing the power of interference plus noise, given a priori probabilities of yet undetected layers and a posteriori probabilities of detected layers. As a result, soft feedback interference cancellation based on a posteriori information is derived, yielding symbol detection robust to error propagation effects. Furthermore, a low complexity implementation using approximate detection ordering and linear filtering is introduced. Simulations performed for block fading channels show that the SIOF symbol detector exhibits performance gains over the existing TURBO-BLAST algorithm [3].

This paper presents a new and simple decoding algorithm for layered space time block codes such as the two independent Alamouti's codes which are also called the double space-time transmit diversity (DSTTD) system. By using group interference suppression and successive interference cancellation, we can treat DSTTD as two independent space-time block codes (STBC). We can then decode both of these STBC's through a simple maximum likelihood (ML) detector with null space-based interference cancellation. We also compare the proposed interference cancellation (IC) scheme with the conventional MMSE IC scheme. The performance of the proposed IC scheme is comparable to that of the MMSE IC scheme while the complexity of the proposed scheme is lower than that of the MMSE IC scheme.

In multiuser MIMO systems, the base station schedules transmissions to a group of users simultaneously. Since the data transmitted to each user are different, in order to avoid the inter-user interference, a transmit preprocessing technique which decomposes the multiuser MIMO downlink channel into multiple parallel independent single-user MIMO channels can be used. When the number of users is larger than the maximum that the system can support simultaneously, the base station selects a subset of users who have the best instantaneous channel quality to maximize the system throughput. Since the exhaustive search for the optimal user set is computationally prohibitive, a low complexity scheduling algorithm which aims to maximize the capacity upper bound is proposed. Simulation results show that the proposed scheduling algorithm achieves comparable total throughput as the optimal algorithm with much lower complexity.

One of the key ideas in this paper is to merge frequency offset compensation into the FFT matrix. The new inverse transform matrix is used in lieu of regular FFT. In uplink of OFDMA, different users (subcarriers) may have different frequency offsets, which causes severe performance degradation. It is also difficult to estimate/compensate these frequency offsets in uplink case. Even if the frequency offsets are correctly estimated, the compensation of one subcarrier (or user) may worsen the inter-carrier interference from the frequency offsets of other subcarriers (users). A single MMSE-based transform is used to perform both inverse Fourier transform and frequency offset compensation. Compared to the conventional methods such as circular convolution, this approach has lower computational complexity with comparable performance.

In this paper, we propose low complexity LDPC decoding algorithms with variable number of edges. The main idea of this paper is to use the bit nodes selectively instead of using all the connected bit nodes. The proposed algorithms choose the bit node selectively when calculating the check node update equation. First, we study the relationship between the number of bit nodes for updating check node and the performance of the BP decoding algorithm. Two selective algorithms are proposed accordingly. The first algorithm (algorithm 1) does not use the bit nodes whose values are highly reliable when calculating the check node update equation. In the second algorithm (algorithm 2), we increase the number of bit nodes to update check node each time we have a decoding failure. These two algorithms have lower complexity with small performance degradation compared to the LLR-BP based decoding algorithm. These algorithms reduce the computational complexity at high SNR more than they do at low SNR.

We propose a practical SVD-based scheduling technique for MIMO broadcast channel. Multiuser diversity is exploited by scheduling transmissions simultaneously to a group of users who have the best instantaneous channel quality. In order to avoid the inter-user interference, we use a transmit preprocessing technique which decomposes the multiuser MIMO downlink channel into multiple parallel independent single-user MIMO channels. Simulation results show that the proposed suboptimal multiuser scheduling algorithm achieve almost the same bit error rate performance as the optimal scheduling algorithm.

This letter deals with the log likelihood ratio computation for the turbo decoder operating on the forward link of a CDMA system. It is shown that the log likelihood ratio contains a time-varying factor, which is the power ratio between the data and the pilot channels. Two alternative techniques for estimating the power ratio are compared along with an ideal scheme.

We present a multiresolution approach for video indexing and feature matching of subband-coded video databases. Four different scene-change detectors were tested; scene-change detection is applied only on the lowest subband for computational efficiency. Two kinds of scene changes, abrupt and smoothly accumulated, mark the beginning of new scene segments. The index for each scene segment is the pair of the histograms of two representative frames, the first and the last frame of the scene. Using the approach of query by example, the index-matching algorithm takes a multiresolution approach by hierarchically comparing histograms at different resolutions. The search algorithm for the match between example query and its target scene segment starts from the coarsest resolution and moves to the next finer resolution until the finest resolution is reached. Experimental results are presented, and the proposed indexing technique appears to be promising for its speed and its inherent hierarchical search procedure.

This paper presents a new motion-compensated subband video-coding technique using variable reference-frame positions for motion estimation. The work builds on temporal segmentation for determining the reference-frame positions and incorporates multiresolution motion estimation in the subband domain. The key advantage of the new algorithm is the complexity reduction in motion estimation that uses the multiresolution property of subband decomposition. Temporal segmentation is performed with the lowest spatial subband, based on the detection of two types of scene change. The frames of input video are split into seven spatial subbands, and the motion vectors for each subband are generated by a hierarchical motion-estimation algorithm. Blockwise differential pulse code modulation (DPCM) and a uniform quantizer are used for the lowest subband of the first frame only, and the subbands of all other cases are coded by PCM with a dead zone quantizer. Simulation results show that our scene-adaptive scheme compares favorably with a fixed interpolation structure.

A new multispectral image compression technique based on the Karhunen-Loeve transform (KLT) and the discrete cosine transform (DCT) is proposed. The quadtree for determining the transform block size and the quantizer for encoding the transform coefficients are jointly optimized in a rate-distortion sense. The problem is solved by a Lagrange multiplier approach. After a quadtree is determined by this approach, a one-dimensional (1-D) KLT is applied to the spectral axis for each block before the DCT is applied on the spatial domain. The eigenvectors of the autocovariance matrix, the quantization scale, and the quantized transform coefficients for each block are the output of the encoder. The overhead information required in this scheme is the bits for the quadtree, KLT, and quantizer representation.

An optimal frequency domain bit allocation technique is presented. A Lagrange multiplier based approach is used for the rate-distortion optimization of the quantization matrix in MPEG-2 encoding. Since the optimal quantization matrix is heavily dependent on the choice of the quantizer, the quantizer is also jointly optimized in order to alleviate the interaction between the two. The complexity of the algorithm is analyzed, and it is shown to be directly proportional to the number of quantization matrices and quantization scales. Experimental results comparing the new technique with a conventional MPEG-2 encoding are also presented.

A new approach to finding the best block size for quadtree-based motion estimation is developed and compared to a conventional fixed block size algorithm. The optimal balance between the motion vector bits and the discrete cosine transform (DCT) coefficient bits is achieved by a fast tree optimization technique based on Lagrange multiplier.

A new bit allocation technique between motion vectors and DCT coefficients for MPEG-2 encoding is presented. In MPEG-2 encoding, the differential between two neighboring motion vectors is losslessly encoded by a Huffman table. In order to adjust the bit allocation for the motion vectors, a motion vector smoothing approach is taken. The motion vectors are smoothed by a low pass filter when the bit allocation for the motion vector has to be reduced. An algorithm based on the Lagrange multiplier is used to achieve the optimal balance between motion vector smoothing and DCT coefficients quantization.

We present an algorithm for joint optimization of anchor frame separation and bit allocation for motion-compensated video coders. The anchor frame separation is optimized in the sense that the distortion is minimized under a bit budget constraint. At the same time, the quantization for each frame in a group of pictures is also optimized in an operational rate distortion sense. The optimal anchor frame separation does depend on the quantization of each frame so that the two optimization problems cannot be separated. A Lagrange multiplier approach can be used to obtain the optimal solution if we assume that the rate-distortion curve is convex. Heuristic algorithms based on simulated annealing and greedy trellis selection are also presented to reduce the computational complexity.

A new compression technique using luminance compensation with adaptive block size for multispectral or volumetric images is proposed. The luminance compensation which is a function approximation technique is updated block by block for efficient coding. Existing compression techniques do not exploit the property that the spatial features of neighboring images are almost identical in multispectral or medical volumetric images. In other words, the image content of each image usually has almost the same spatial shape even when the pixel values are different. The luminance compensation based on least squares takes advantage of the relationship of almost identical spatial feature. A Lagrange multiplier technique is also used to optimize the quadtree for the adaptive block size.

In this paper, we present a new approach to find the best quadtree for variable block size (VBS) motion estimation for video coding. In this approach, the quantizer which encodes the prediction residual is also jointly optimized in a rate-distortion sense. The desired balance of the bit allocation between the motion vector coding and the residual error coding is achieved by a Lagrange multiplier approach for the joint optimization problem. A fast tree optimization technique is used to reduce the computational complexity of the algorithm so that it can be run in realtime in a appropriate hardware.

We present a new subband video coding algorithm with temporally adaptive motion interpolation. In the approach proposed, the reference frames for motion estimation are adaptively selected using temporal segmentation in the lowest spatial subband. Variable target bit allocation for each picture type in a group of pictures is used to allow variable number of reference frames with the constraint of constant output bit rate. Block-wise DPCM, PCM, and run-length coding combined with truncated Huffman coding are used to encode the quantized data in the subbands. Simulation results of the adaptive scheme compare favorably with those of a non-adaptive scheme.>

In this paper, we present an algorithm for joint optimisation of motion compensation structure and bit allocation for motion compensated video coders. In this scheme, the motion compensation structure is optimized in the sense that the distortion is minimized under a bit budget constraint. At the same time, the quantisation for each frame in the group of pictures (GOP) is also optimized in an operational rate distortion sense.>

In this paper we present a novel technique to dynamically adapt motion interpolation structures by temporal segmentation. The number of reference frames and the intervals between them are adjusted according to the temporal variation of the input video. Bit-rate control for this dynamic group of pictures (GOP) structure is achieved by taking advantage of temporal masking in human vision. Constant picture quality can be obtained by variable-bit-rate coding using this approach. Further improvement can be made when the intervals between reference frames are chosen by minimizing a measure of the coding difficulty of a GOP. Advantages for low bit-rate coding and implications for variable-bit-rate coding are discussed. Simulations on test video are presented for various GOP structures and temporal segmentation methods, and the results compare favorably with those for conventional fixed GOP structures.

In this paper we present a novel technique to dynamically adapt motion interpolation structures by temporal segmentation. The interval between two reference frames is adjusted according to the temporal variation of the input video. The difficulty of bit rate control for this dynamic group of pictures (GOP) structure is resolved by taking advantage of temporal masking in human vision. Six different frame types are used for efficient bit rate control, and telescopic search is used for fast motion estimation because frame distances between reference frames are dynamically varying. Constant picture quality can be obtained by variable bit rate coding using this approach and the statistical bit rate behavior of the coder is discussed. Advantages for low bit rate coding and storage media applications and implications for HDTV coding are discussed. Simulations on test video including HDTV sequences are presented for various GOP structures and different bit rates, and the results compare favorably with those for conventional fixed GOP structures.