Fabrication of 0.2 to 0.3 µm features is vital for future ultralarge scale integration devices. An area of particular concern is whether optical lithography can delineate such feature sizes, i.e., less than the exposure wavelength. The use of a phase shift mask is one of the most effective means of improving resolution in optical lithography. This technology basically makes use of the interference between light transmitting through adjacent apertures of the mask. Various types of phase shift masks and their imaging characteristics are discussed and compared with conventional normal transmission masks. To apply these masks effectively to practical patterns, a phase shifter pattern design tool and mask repair method must be established. The phase shifting technology offers a potential to fabricate 0.3 µm features by using the current i-line stepper, and 0.2 µm features by using excimer laser stepper.

Breakdown fields and the charges to breakdown (QBD) of oxides increased after UV/Cl2 pre-oxidation cleaning. This is due to decreased residual metal contaminants on silicon surfaces in the bottom of the LOCOS region after wet cleaning. Treatment in NH4OH, H2O2 and H2O prior to UV/Cl2 cleaning suppressed increases in surface roughness and kept leakage currents through the oxides after UV/Cl2 cleaning as low as those after wet cleaning alone. The large junction leakage currents--caused by metal contaminants introduced during dry etching--decreased after UV/Cl2 cleaning which removes the contaminated layer.

Recent progress in high-speed semiconductor devices and integrated circuits (ICs) has outpaced the conventional measuring and testing instruments. With advent of ultrashort-pulse laser technology, the electro-optic sampling (EOS) technique based on the Pockels effect has become the most promising solution way of overcoming the frequency limit, whose bandwidth is approaching a terahertz. This paper reviews recent progress on the research of the EOS technniques for measuring ultrahigh-speed electronic devices and ICs. It describes both the principle of the EOS and the key technologies used for noncontact probing of ICs. Internal-node measurements of state-of-the-art high-speed ICs are also presented.

This paper was written for LSI engineers in order to demonstrate the effect of optical interconnections in LSIs to improve both the speed and power performances of 0.5 and 0.2 µm CMOS microprocessors. The feasibilities and problems regarding new micronsize optoelectronic devices as well as associated electronics are discussed. Actual circuit structures clocks and bus lines used for optical interconnection are discussed. Newly designed optical interconnections and the speed power performances are compared with those of the original electrical interconnection systems.

We propose two types of key management systems that use complementary exponential calculation, in which users in the system divide into groups, and the different modulus numbers are assigned to each group and edges between groups. Key generation information over the modulus numbers is issued to a user by a trusted center. The user who receives the information can generate shared encryption keys between users in the system without using key exchange protocol. In our proposed system, the number of primes is one of the parameters for generating key generation information. The number decreases in inverse proportion to the square of the number of groups compared to the original method. Our proposed technique enabled us to extend the number of users in the system to more than one million, which is not possible with the original method.

We propose a new randomized version of DES in which a key-dependent swapping is used to strengthen DES and DES-like cryptosystems against differential cryptanalysis. This new scheme, called RDES, decreases the probability of success in differential attack by decreasing the characteristic probability. The characteristic is the effect of particular differences in plaintext pairs on the differences in the resultant ciphertext pairs. The characteristic probability for the n-round RDES is 2-n+1 times that for the n-round DES. As for the differential cryptanalysis based on characteristics, the 16-round RDES is as strong as the about 20-round DES. Encryption/decryption speed of n-round RDES is almost the same as that of the n-round DES.

This paper presents two methods for securely realizing caller-authenticated and callee-specified calls over telecommunication networks with terminals that accept IC cards having KPS-based cryptographic functions. In the proposed protocols, users can verify that the partner is the proper owner of a certain ID or a certain pen name. Users' privacy is protected even if they do the caller-authenticated and callee-specified calls and do not pay their telephone charge in advance.

This paper discusses the problems facing spoken dialogue processing and the prospects for future improvements. Research on elemental topics like speech recognition, speech synthesis and language understanding has led to improvements in the accuracy and sophistication of each area of study. First, in order to handle a spoken dialogue, we show the necessity for information exchanges between each area of processing as seen through the analysis of spoken dialogue characteristics. Second, we discuss how to integrate those processes and show that the memory-basad approach to spontaneous speech interpretation offers a solution to the problem of process integration. The key to this is setting up a mental state affected by both speech and linguistic information. Finally, we discuss how those mental states are structured and a method for constructing them.

We define the communication complexity of a perfect zero-knowledge interactive proof (ZKIP) as the expected number of bits communicated to achieve the given error probabilities (of both the completeness and the soundness). While the round complexity of ZKIPs has been studied greatly, no progress has been made for the communication complexity of those. This paper shows a perfect ZKIP whose communication complexity is 11/12 of that of the standard perfect ZKIP for a specific class of Quadratic Residuosity.

In this paper a recursive form of Welch-Berlekamp (W-B) algorithm is provided which is a novel and fast decoding algorithm.

This paper discusses three approaches for combining an efficient LR parser and phoneme-context-dependent HMMs and compares them through continuous speech recognition experiments. In continuous speech recognition, phoneme-context-dependent allophonic models are considered very helpful for enhancing the recognition accuracy. They precisely represent allophonic variations caused by the difference in phoneme-contexts. With grammatical constraints based on a context free grammar (CFG), a generalized LR parser is one of the most efficient parsing algorithms for speech recognition. Therefore, the combination of allophonic models and a generalized LR parser is a powerful scheme enabling accurate and efficient speech recognition. In this paper, three phoneme-context-dependent LR parsing algorithms are proposed, which make it possible to drive allophonic HMMs. The algorithms are outlined as follows: (1) Algorithm for predicting the phonemic context dynamically in the LR parser using a phoneme-context-independent LR table. (2) Algorithm for converting an LR table into a phoneme-context-dependent LR table. (3) Algorithm for converting a CFG into a phoneme-context-dependent CFG. This paper also includes discussion of the results of recognition experiments, and a comparison of performance and efficiency of these three algorithms.

Given an integer N, it is easy to determine whether or not N is prime, because a set of primes is in LPP. Then given a composite number N, is it easy to determine whether or not N is of a specified form? In this paper, we consider a subset of odd composite numbers +1MOD4 (resp. +3MOD4), which is a subset of odd composite numbers consisting of prime factors congruent to 1 (resp. 3) modulo 4, and show that (1) there exists a four move (blackbox simulation) perfect ZKIP for the complement of +1MOD4 without any unproven assumption; (2) there exists a five move (blackbox simulation) perfect ZKIP for +1MOD4 without any unproven assumption; (3) there exists a four move (blackbox simulation) perfect ZKIP for +3MOD4 without any unproven assumption; and (4) there exists a five move (blackbox simulation) statistical ZKIP for the complement of +3MOD4 without any unproven assumption. To the best of our knowledge, these are the first results for a language L that seems to be not random self-reducible but has a constant move blackbox simulation perfect or statistical ZKIP for L and without any unproven assumption.

This paper describes a general interface system for speech input and output and a dialog management system, MASCOTS, which is a component of the interface system. The authors designed this interface system, paying attention to its generality; that is, it is not dependent on the problem-solving system it is connected to. The previous version of MASCOTS dealt with the dialog processing only for the speech input based on the SR-plans. We extend MASCOTS to cover the speech output to the user. The revised version of MASCOTS, named MASCOTS II, makes use of topic information given by the topic packet network (TPN) which models the topic transitions in dialogs. Input and output messages are described with the concept representation based on the case structure. For the speech input, prediction of user's utterance is focused and enhanced by using the TPN. The TPN compensates for the shortages of the SR-plan and improves the accuracy of prediction as to stimulus utterances of the user. As the dialog processing in the speech output, MASCOTS II extracts emphatic words and restores missing words to the output message if necessary, e.g., in order to notify the results of speech recognition. The basic mechanisms of the SR-plan and the TPN are shared between the speech input and output processes in MASCOTS II.

In this paper, we investigate the discrepancy between a serial version and a parallel version of zero-knowledge protocols, and clarify the information "leaked" in the parallel version, which is not zero-knowledge unlike the case of the serial version. We consider two sides: one negative and the other positive in the parallel version of zero-knowledge protocols, especially of the Fiat-Shamir scheme.

An adaptive fuzzy network (AFN) is described that can be used to implement most of fuzzy logic functions. We introduce a learning algorithm largely borrowed from backpropagation algorithm and train the AFN system for several typical fuzzy problems. Simulations show that an adaptive fuzzy network can be implemented with the proposed network and algorithm, which would be impractical for a conventional fuzzy system.

Stability property of an optically injection-locked semiconductor laser taking account of gain saturation is discussed. Numerical analysis shows that stable locking region is broadened due to gain saturation. This is because of rapid damping of relaxation oscillation due to gain saturation. It is also found that stable locking region is also broadened with increasing injection current since damping of relaxation oscillation becomes strong with increasing injection current. Numerical calculations of lasing spectrum show that the magnitude of sidepeaks appeared at harmonics of relaxation oscillation frequency under unstable locking condition are suppressed due to gain saturation.

SOI technology has been developed for not only future ULSI, but also intelligent power ICs and sensors. In this paper the SOI fabrication process with wafer bonding and polish-stopper technologies, and its advantages for future ULSI are shown. And high crystal quality of SOI films fabricated with this method, and high speed performance of SOI devices and circuits, are shown from the data of 256 kb full CMOS SRAM chips. Moreover it is shown from the fabrication data of 4 Mb full CMOS SRAM cells that this technology has a large flexibility on device structure design. These results mean that our technology has great advantages for reduction of cell size and improvement of circuit performance.

This paper describes high-temperature operation of nMOSFET on bonded SOI. A long-channel nMOSFET is fabricated on bonded SOI (Si layer thickness 0.3 µm), SOS (Si layer thickness 0.3 µm), and bulk Si, Bonded SOI is produced using pulse-field-assisited bonding and resistivity-sensitive etching. The high-temperature operation of bonded SOI nMOSFET is demonstrated and compared with SOS and bulk MOSFETs. The leakage current variation with temperature is signnificantly smaller in bonded SOI and in SOS than in bulk MOSFETs. At high temperatures, the drain current to leakage current ratio is 100 times higher in bonded SOI than in SOS and bulk devices. At 300, a ratio of 104 is obtained for the bonded SOI nMOSFET. The ratio is expected to be even higher if a reduced channel length and ultrathin (less than 0.1 µm) bonded SOI is used.

The intelligent network services will considerably increase the amount of Control and OAM (Operation, Administration and Maintenance) Information (Ic&o) which will be stored in a huge number of distributed databases. Therefore, the management and the organization of databases have become critical issues for securing network performance. This paper studies one of the IN applications that is likely to be an important user of the Ic&o network, namely the Universal Personal Telecommunication (UPT). UPT enables the personal mobility, based on a UPT number related to the user and not a terminal equipment. The Ic&o information of UPT is carried through an ATM based transport network. Taking two fundamental parameters into consideration, namely delay time and the number of users, and two kinds of data location probabilities, this paper studies two basic procedures for finding target data in UPT databases, i.e., chaining and broadcasting. Results show that, when the data location probability is uniform, the broad-casting mode is the faster mode but, on the other hand, the chaining mode allows a larger number of users because the disk access time is less restrictive than in the broadcasting mode. Moreover, this study shows that increasing the number of databases also increases the allowed number of users up to a specific threshold. With a Broadcast Chaining mode, a better compromise between the delay time and the number of allowed users is obtained. If the probability depends on the location of databases (the probability is conversely proportional to the square of the number of searched databases), the results show that the chaining mode is preferable from both the number of users allowed and the delay time viewpoints. Finally, the implementation aspect is discussed.

This is a partly expository paper discussing how point processes with certain "bursty" features can be qualitatively modelled by the Markovian arrival process, a generalization of the Poisson or Bernoulli processes which can be used to obtain algorithmically tractable matrix solutions to a variety of problems in probability models.