This paper describes Japanese-English translation of weather reports by ISOBAR. Recently, there have been developed a lot of automatic translation systems, either operational or experimental. In the domain of meteorological reports, TAUM-METEO (English-French) is the best-known operational system, which is classified as one of the transfer systems. Our translation by ISOBAR is still experimental, but its most remarkable feature is that target language (TL) sentences are generated only from interlingual meaning representations of input source language (SL) sentences. That is, our system has no transfer component. In this approach, one sentence can be translated into plural TL sentences which are in the paraphrase relation. During semantic processing of SL sentences, ISOBAR, unlike Wilks' model, is designed to reject definitely such sentences as are semantically anomalous or contradictory to its knowledge. We discuss several merits and demerits of this approach, and show some experimental results in ISOBAR.

This paper shows an experimental approach to the understanding system of natural language and pictorial patterns. The system is titles as ISOBAR (an Information understanding System Of BAsic weather Report). It can accept both linguistic and pictorial inputs. Also it produces either linguistic or pictorial, or both, output according to the commands which it received. The most remarkable point of the system is that the performance of the system is based on the semantic processing of the input. As ISOBAR's world is limited within the weather report of Japan and Far East Asian areas with associated weather charts, so the semantic background of the system is very narrow, and is quite a specific one. But the methodologies and algorithms in the system will be the first step for the embodiment of the more complicated and more general systems. ISOBAR has two operating modes in principle. One is the accumulation of meteorological information, and the other is its retrieval. According to the experimental outcomes, the performance of the system is almost good except for the picture processing procedures. Finally, the problems for the future research are remarked.