This paper presents a child verb learning model mainly based on syntactic bootstrapping. The model automatically learns 4-5-year-old children's linguistic knowledge of verbs, including subcategorization frames and thematic roles, using a text in dialogue format. Subcategorization frame acquisition of verbs is guided by the assumption of the existence of nine verb prototypes. These verb prototypes are extracted based on syntactic bootstrapping and some psycholinguistic studies. Thematic roles are assigned by syntactic bootstrapping and other psycholinguistic hypotheses. The experiments are performed on the data from the CHILDES database. The results show that the learning model successfully acquires linguistic knowledge of verbs and also suggest that psycholinguistic studies of child verb learning may provide important hints for linguistic knowledge acquisition in natural language processing (NLP).

The decision logic expression method is generalized so that any two-valued logic utilizing conventional AND, OR, and NOT operations can be expressed in terms of conditional AND. Thus, the scheme is suited to describing decision logics in fields such as education, medicine, and management.

In this paper, we show that a priori probabilities of information bits can be incorporated into metrics for syndrome decoding. Then it is confirmed that soft-in/soft-out decoding is also possible for syndrome decoding in the same way as for Viterbi decoding. The derived results again show that the two decoding algorithms are dual to each other.

Algebraic properties of the set of regular states for a given hierarchically structured learning objectives are discussed. The set is proved to be a distributive lattice, if two operations, meet and join, are properly defined on it. The derived algebraic properties of the set are then applied to some of the procedures in the learning diagnosis and treatment (LDT) reported earlier.

Both Viterbi decoding with labels (i.e., the Yamamoto-Itoh scheme) and the soft-output Viterbi algorithm (SOVA) evaluate the metric difference between the maximum-likelihood (ML) path and the discarded path at each level in the trellis. Noting this fact, we show that the former scheme also provides information about the reliability values for decoded information bits.

The structure of bidirectional syndrome decoding for binary rate (n-1)/n convolutional codes is investigated. It is shown that for backward decoding based on the trellis of a syndrome former HT, the syndrome sequence must be generated in time-reversed order using an extra syndrome former H*T, where H* is a generator matrix of the reciprocal dual code of the original code. It is also shown that if the syndrome bits are generated once and only once using HT and H*T, then the corresponding two error sequences have the intersection of n error symbols, where is the memory length of HT.

It is known that Viterbi decoding based on the code trellis and syndrome decoding based on the syndrome trellis (i.e., error trellis) are equivalent. In this paper, we show that Scarce State Transition (SST) Viterbi decoding of convolutional codes is equivalent to syndrome decoding. First, we derive fundamental relations between the hard-decision input to the main decoder and the encoded data for the main decoder. Then using these relations, we show that the code trellis module for the main decoder in an SST Viterbi decoder can be reduced to a syndrome trellis module. This fact shows that SST Viterbi decoding based on the code trellis is equivalent to syndrome decoding based on the syndrome trellis. We also calculate the SST Viterbi decoding metrics for general convolutional codes assuming an AWGN channel model. It is shown that the derived metrics are equal to those of conventional Viterbi decoding. This fact shows that SST Viterbi decoding is equivalent to conventional Viterbi decoding, and consequently to syndrome decoding.

In this paper, we state some noteworthy facts in connection with simplification of the BCJR algorithm using the bidirectional Viterbi algorithm (BIVA). That is, we clarify the necessity of metric correction in the case that the BIVA is applied to reliability estimation, where information symbols uj obey non-uniform probability distributions.

This paper presents an extension of the database query language SQL to include queries against a database with natural language annotations. The proposed scheme is based on Concept Coupling Model, a language model for handling natural language sentence structures. Integration of the language model with the conventional relational data model provides a unified environment for manipulating information sources comprised of relational tables and natural language texts.