The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
We present Marlin, a variable-to-fixed (VF) codec optimized for decoding speed. Marlin builds upon a novel way of constructing VF dictionaries that maximizes efficiency for a given dictionary size. On a lossless image coding experiment, Marlin achieves a compression ratio of 1.94 at 2494MiB/s. Marlin is as fast as state-of-the-art high-throughput codecs (e.g., Snappy, 1.24 at 2643MiB/s), and its compression...
We provide a novel upper bound on the rate-distortion function with multiple decoders with side information, which takes the form of a linear program. This complements a recently-obtained lower bound that also takes the form of a linear program. We show that the upper bound is optimal in several instances of the problem1.
Word units are a popular choice in statistical language modelling. For inflective and agglutinative languages this choice may result in a high out of vocabulary rate. Subword units, such as morphs, provide an interesting alternative to words. These units can be derived in an unsupervised fashion and empirically show lower out of vocabulary rates. This paper proposes a morph-to-word transduction to...
Spoken term detection, especially of out-of-vocabulary (OOV) keywords, benefits from the use of sub-word systems. We experiment with different language-independent approaches to sub-word unit generation, generating both syllable-like and morpheme-like units, and demonstrate how the performance of syllable-like units can be improved by artificially increasing the number of unique units. The effect...
In recent years, so-called, “end-to-end” speech recognition systems have emerged as viable alternatives to traditional ASR frameworks. Keyword search, localizing an orthographic query in a speech corpus, is typically performed by using automatic speech recognition (ASR) to generate an index. Previous work has evaluated the use of end-to-end systems for ASR on well known corpora (WSJ, Switchboard,...
Error concealment (EC) is one of the target applications of inpainting techniques. Some methods combine the estimated lost motion vectors (MVs) with the exemplar-based inpainting technique to recover the lost regions. Due to the erroneous motion vectors that might indicate a moving object as background object and vice versa, these methods are still showing visual artifacts in the recovered regions...
Sudoku is a pencil puzzle. The aim of the solver is to complete the 9×9 grid by filling in a digit in every cell in a certain rule. In this study, we regard the process of solving Sudoku as a process of decoding a codeword from a received word, and show the expected decoding error probability obtained by experiments for erasure channels.
Recently, partitional clustering approaches based on Evolutionary Algorithms (EAs) have shown promising in solving the data clustering problems. However, with the nearest prototype (NP) rule as the method for decoding, most of them are only suitable for clustering datasets with convex (e.g. hyperspherical) clusters. In this paper, we propose an automatic clustering approach using particle swarm optimization...
In this paper, a duality between wiretap and state-dependent channels with non-causal channel state information at the transmitter is established. First, a common achievable scheme is described for a certain class of state-dependent and wiretap channels. Further, state-dependent and wiretap channels for which this scheme is capacity (resp. secrecy capacity) achieving are identified. These channels...
A novel sparse coding framework with unity range codes and the possibility to produce a discriminative dictionary is presented. The framework is, in contrast to many other works, able to handle unsupervised, supervised and semi-supervised settings. Furthermore, codes are constrained to be in unity range, which is beneficial in many scenarios. The paper presents the framework and solvers used to produce...
We propose a max-pooling based loss function for training Long Short-Term Memory (LSTM) networks for small-footprint keyword spotting (KWS), with low CPU, memory, and latency requirements. The max-pooling loss training can be further guided by initializing with a cross-entropy loss trained network. A posterior smoothing based evaluation approach is employed to measure keyword spotting performance...
We investigate the outage and decoding-delay performances for full-duplex (FD) decode-forward (DF) relaying with backward and sliding window decoding. In our analysis, we consider a block fading channel with full channel state information (CSI) availability at receivers and with limited CSI at transmitters. For backward decoding where the destination starts decoding from the last transmission block,...
Spatially-coupled (SC) codes have recently attracted significant attention due to their capability to achieve capacity approaching performance. Significant recent research has been devoted to the asymptotic study of SC codes. Although the asymptotic analysis is relevant, it cannot be directly applied in the finite-length setting due to cycle-free and averaging assumptions. In this paper, we consider...
This paper proposes a Modified Multiple Triangular Inequality Elimination (MMTIE) to further reduce the search number of candidate codevectors. The MMTIE adopts the same original search space as the MTIE scheme with the initial best-matched codevector selected by the Initial Index Code Assignment (IICA), and integrates the intersection rule of the Candidate Codevectors Group (CCG) scheme to further...
A distributed hypothesis testing problem is considered, where the goal is to declare the distribution of two random variables, based on their observations. Defining two error events, the error exponent of Type II is studied under a fixed constraint over the error of type I. A novel approach is presented, based on random binning. The benefits of this approach are demonstrated through an example, compared...
In this paper we investigate the impact of puncturing on a given information set. This study is necessary to design a good rate-compatible puncturing pattern for hybrid automatic repeat request based on incremental redundancy (HARQ-IR) since an information set should be unchanged during retransmission. We first identify that for an information set, there exist the so-called catastrophic puncturing...
We consider the problem of secret key-based authentication under a privacy constraint on the enrolled source data. The user's measurement in the authentication phase is assumed to be controllable via a cost-constrained “action” sequence. We provide a single-letter characterization of the optimal tradeoff among the secret key rate, storage rate, leakage rate, and action cost. The result can be relevant...
The application of lossless data compression methods may significantly decrease the required storage space or throughput. The quality of the compression ratio is influenced by several factors, like the available space and processing resources and the statistical characteristic of the data to be compressed. In this contribution we will perform the boundary analysis of Virtual Dictionary Extension(-VDE),...
In this paper we consider multicomponent network codes for a channel with random linear transformations and packet errors. We present a new code construction, which is a generalization of MZP network codes, and similarly uses rank codes as subcodes. In the new construction we use a greedy algorithm, when selecting the code components. It allows us to increase code cardinality compared with the previously...
We present results on the performance of a network-coding based method for broadcast file delivery in an ad hoc network consisting of half-duplex packet radios. Our technique allows a receiver to exploit the broadcast nature of the wireless medium by overhearing packets sent by a node that is not necessarily one of the receiver's designated relays. Delivery of the file to the remote destinations is...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.