Conventional confidence measures for assessing the reliability of ASR (automatic speech recognition) output are typically derived from "low-level" information which is obtained during speech recognition decoding. In contrast to these approaches, we propose a novel utterance verification framework which incorporates "high-level" knowledge sources. Specifically, we investigate two application-independent measures: in-domain confidence, the degree of match between the input utterance and the application domain of the back-end system, and discourse coherence, the consistency between consecutive utterances in a dialogue session. A joint confidence score is generated by combining these two measures with an orthodox measure based on GPP (generalized posterior probability). The proposed framework was evaluated on an utterance verification task for spontaneous dialogue performed via a (English/Japanese) speech-to-speech translation system. Incorporating the two proposed measures significantly improved utterance verification accuracy compared to using GPP alone, realizing reductions in CER (confidence error-rate) of 11.4% and 8.1% for the English and Japanese sides, respectively. When negligible ASR errors (that do not affect translation) were ignored, further improvement was achieved for the English side, realizing a reduction in CER of up to 14.6% compared to the GPP case.

We presented a new text-independent/text-prompted speaker recognition method by combining speaker-specific Gaussian Mixture Model (GMM) with syllable-based HMM adapted by MLLR or MAP. The robustness of this speaker recognition method for speaking style's change was evaluated in this paper. The speaker identification experiment using NTT database which consists of sentences data uttered at three speed modes (normal, fast and slow) by 35 Japanese speakers (22 males and 13 females) on five sessions over ten months was conducted. Each speaker uttered only 5 training utterances (about 20 seconds in total). A combination method reduced the identification error rate by about 50%. We obtained the accuracy of 98.8% for text-independent speaker identification for three speaking style modes (normal, fast, slow) by using a short test utterance (about 4 seconds). Especially, we obtained the accuracy of 99.4% for normal speaking mode. This result was superior to conventional methods for the same database. We show that the attractive result was brought from the compensational effect between speaker specific GMM and speaker adapted syllable based HMM.

The increased complexity in system design has brought an explosive growth in functional verification time. Thus, many verification methodologies have been proposed to reduce it. One of them is the co-emulation method in which the hardware accelerator and software simulator work together. This paper presents an effective testbench architecture for accelerated verification and reuse of parts of the testbench in co-emulation. The testbench is divided into a synthesizable part which can be hardware accelerated and a non-synthesizable part which remains on the software simulator. The split blocks of the testbench can be reused in other test environments. Experiments with real world systems show that the proposed verification environment has over 31% higher performance than that of the conventional co-emulation environment.

We show that Sun-Chen-Hwang's three-party key agreement protocols using passwords are insecure against an active adversary. Further, we recommend a small change to the protocols that fixes the security problem.

As the use of electronic voting systems and e-commerce systems increases, the efficient batch verification of digital signatures becomes more and more important. In this paper, we first propose a new method to identify bad signatures in batches efficiently for the case when the batch contains one bad signature. The method can find out the bad signature using smaller number of modular multiplications than the existing method. We also propose an extension to the proposed method to find out two or more bad signatures in a batch instance. Experimental results show that our method yields better performance than the existing method in terms of the number of modular multiplications.

This paper presents a multi-matcher on-line signature verification system which fuses the verification scores in pen-position parameter and pen-movement angle one at total decision. Features of pen-position and pen-movement angle are extracted by the sub-band decomposition using the Discrete Wavelet Transform (DWT). In the pen-position, high frequency sub-band signals are considered as individual features to enhance the difference between a genuine signature and its forgery. On the other hand, low frequency sub-band signals are utilized as features for suppressing the intra-class variation in the pen-movement angle. Verification is achieved by the adaptive signal processing using the extracted features. Verification scores in the pen-position and the pen-movement angle are integrated by using a weighted sum rule to make total decision. Experimental results show that the fusion of pen-position and pen-movement angle can improve verification performance.

The Oblivious Transfer (OT), introduced by Rabin in 1981, has become an important and fundamental cryptography technique. An OT protocol should have two important characteristics: the sender's privacy and the chooser's privacy. The sender is a party who will deliver a secret to the chooser. The chooser is another party who acts as receiver to learn some information about the input from the sender. The chooser learns of certain information concerning the sender's input while the sender is not allowed to know what the chooser has learned. Moreover, the chooser cannot acquire any messages that he/she did not choose. Naor and Pinkas have recently proposed an efficient oblivious transfer protocol (EOT) that implementes 1-out-of-n protocol, but this EOT has a flaw: it cannot withstand "the same message attack." In this paper, we will improve Naor and Pinkas EOT and make it resistant to "the same message attack."

This work proposes a technique to automatically obtain timing constraints for a given timed circuit to operate correctly. A designated set of delay parameters of a circuit are first set to sufficiently large bounds, and verification runs followed by failure analysis are repeated. Each verification run performs timed state space enumeration under the given delay bounds, and produces a failure trace if it exists. The failure trace is analyzed, and sufficient timing constraints to prevent the failure are obtained. Then, the delay bounds are tightened according to the timing constraints by using an ILP (Integer Linear Programming) solver. This process terminates when either some delay bounds under which no failure is detected are found or no new delay bounds to prevent the failures can be obtained. The experimental results using a naive implementation show that the proposed method can efficiently handle asynchronous benchmark circuits and nontrivial GasP circuits.

We can model embedded systems as hybrid systems. Moreover, they are distributed and real-time systems. Therefore, it is important to specify and verify randomness and soft real-time properties. For the purpose of system verification, we formally define probabilistic linear hybrid automaton and its symbolic reachability analysis method. It can describe uncertainties and soft real-time characteristics.

There is increasing evidence from in vivo recordings in monkeys trained to respond to stimuli by making left- or rightward eye movements, that firing rates in certain groups of neurons in oculo-motor areas mimic drift-diffusion processes, rising to a (fixed) threshold prior to movement initiation. This supplements earlier observations of psychologists, that human reaction-time and error-rate data can be fitted by random walk and diffusion models, and has renewed interest in optimal decision-making ideas from information theory and statistical decision theory as a clue to neural mechanisms. We review results from decision theory and stochastic ordinary differential equations, and show how they may be extended and applied to derive explicit parameter dependencies in optimal performance that may be tested on human and animal subjects. We then briefly describe a biophysically-based model of a pool of neurons in locus coeruleus, a brainstem nucleus implicated in widespread norepinephrine release. This neurotransmitter can effect transient gain changes in cortical circuits of the type that the abstract drift-diffusion analysis requires. We also describe how optimal gain schedules can be computed in the presence of time-varying noisy signals. We argue that a rational account of how neural spikes give rise to simple behaviors is beginning to emerge.

Jacobian Adaptation (JA) has been successfully used in Automatic Speech Recognition (ASR) systems to adapt the acoustic models from the training to the testing noise conditions. In this work we present an improvement of JA for speaker verification, where a specific training noise reference is estimated for each speaker model. The new proposal, which will be referred to as Model-dependent Noise Reference Jacobian Adaptation (MNRJA), has consistently outperformed JA in our speaker verification experiments.

This paper proposes a partial order reduction algorithm for timed trace theoretic verification in order to detect both safety failures and timing failures of timed circuits efficiently. This algorithm is based on the framework of timed trace theoretic verification according to the original untimed trace theory. Consequently, its conformance checking supports hierarchical structure when verifying timed circuits. Experimenting with the STARI and DME circuits, the proposed approach shows its effectiveness.

A policy is an execution rule (or constraint) for objects in a system to retain security and integrity of the system. We introduce a simple policy specification language and define its operational semantics. A new NFA construction algorithm that works in linear time is proposed and a model checking method for policy controlled system (PCS) is presented. We conducted verification of a sample PCS for hotel reservation by our automatic verification tool and the experimental results showed the efficiency of the proposed method.

We have launched "Highly-Reliable Embedded Software Development" Project, held as a part of e-Society Project, supported by Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The aim of this project is to enable the industry to produce highly reliable and advanced software by introducing latest software technologies into embedded software development. In this paper, we introduce the overview of the projects and our activities and results so far.

Fair scheduling algorithms have been proposed to tackle the problem of bursty and location-dependent errors in wireless packet networks. Most of those algorithms ensure the fairness property and guarantee the QoS of all sessions in a large-scale cellular network such as GSM or GPRS. In this paper, we propose the Weighted-Sacrificing Fair Queueing (WSFQ) scheduling algorithm for small-area and device-limited wireless networks. WSFQ slows down the growth of queue length in limited-buffer devices, still maintains the properties of fairness, and guarantees the throughputs of the system. Moreover, WSFQ can easily adapt itself to various kinds of traffic load. We also implement a packet-based scheduling algorithm, the Packetized Weighted Sacrificing Fair Queueing (PWSFQ), to approach the WSFQ. WSFQ and PWSFQ are evaluated by comparing with other algorithms by mathematical analysis and simulations.

We investigate strategies to improve the utterance verification performance using a 2-class pattern classification approach, including: utilizing N-best candidate scores, modifying segmentation boundaries, applying background and out-of-vocabulary filler models, incorporating contexts, and minimizing verification errors via discriminative training. A connected-digit database recorded in a noisy, moving car with a hands-free microphone mounted on the sun-visor is used to evaluate the verification performance. The equal error rate (EER) of word verification is employed as the sole performance measure. All factors and their effects on the verification performance are presented in detail. The EER is reduced from 29%, using the standard likelihood ratio test, down to 21.4%, when all features are properly integrated.

This paper presents the selective orthogonal matrix least-squares (SOM-LS) method for representing a multiport network characterized by sampled data with the rational matrix, improving the previous works, and providing new criteria. Recently, it is needed in a circuit design to evaluate physical effects of interconnects and package, and the evaluation is done by numerical electromagnetic analysis or measurement by network analyzer. Here, the SOM-LS method with the criteria will play an important role for generating the macromodels of interconnects and package in circuit simulation level. The accuracy of the macromodels is predictable and controllable, that is, the SOM-LS method fits the rational matrix to the sampled data, selecting the dominant poles of the rational matrix. In examples, simple PCB models are analyzed, where the rational matrices are described by Verilog-A, and some simulations are carried out on a commercial circuit simulator.

In this paper, we propose a transformation function for a user's raw iris data, an "iris code" in iris scanning verification on the server, since the iris code requires to be hidden from even a server administrator. We then show that the user can be properly authenticated on the server, even though the iris code is transformed by the proposed function. The reason is that the function has a characteristic, "The (normalized) Hamming distances between the enrolled iris codes and the verified iris codes are conserved before and after the computation of the function," that is, the normalized Hamming distance in this scheme is equal to that in the existing scheme. We also show that the transformed iris code is sufficiently secure to hide the original iris code, even if a stronger attack model is supposed than the previously described model. That can be explained from the following two reasons. One reason is that nonlinear function, which consists of the three-dimensional rotation about the x-axis and the y-axis with the iris code lengthened bit by bit, and the cyclic shift, does not enable an attacker to conjecture the iris code. The other reason is that the success probabilities for the exhaustive search attack concerning the iris code in the supposed attack models are lower than those of the previously proposed methods and are negligible.

We propose discrimination method of synthetic speech using pitch pattern of speech signal. By applying the proposed synthetic speech discrimination system as pre-process before the conventional HMM speaker verification system, we can improve the safety of conventional speaker verification system against imposture using synthetic speech. The proposed method distinguishes between synthetic speech and natural speech according to the pitch pattern which is distribution of value of normalized short-range autocorrelation function. We performed the experiment of user verification, and confirmed the validity of the proposed method.

We propose efficient secret sharing schemes realizing general access structures. Our proposed schemes are perfect secret sharing schemes and include Shamir's (k, n)-threshold schemes as a special case. Furthermore, we show that a verifiable secret sharing scheme for general access structures is realized by one of the proposed schemes.