Cooperation of categorical and behavioral learning in a practical solution to the abstraction problem

Real robots should be able to adapt autonomously to various environments in order to go on executing their tasks without breaking down. They achieve this by learning how to abstract only useful information from a huge amount of information in the environment while executing their tasks. This paper proposes a new architecture which performs categorical learning and behavioral learning in parallel with task execution. We call the architectureSituation Transition Network System (STNS). In categorical learning, it makes a flexible state representation and modifies it according to the results of behaviors. Behavioral learning is reinforcement learning on the state representation. Simulation results have shown that this architecture is able to learn efficiently and adapt to unexpected changes of the environment autonomously. Atsushi Ueno, Ph.D.: He is a research associate in the Artificial Intelligence Laboratory at the Graduate School of Information Science at the Nara Institute of Science and Technology (NAIST). He received the B.E., the M.E., and the Ph.D. degrees in aeronautics and astronautics from the University of Tokyo in 1991, 1993, and 1997 respectively. His research interest is robot learning and autonomous systems. He is a member of Japan Association for Artificial Intelligence (JSAI). Hideaki Takeda, Ph.D.: He is an associate professor in the Artificial Intelligence Laboratory at the Graduate School of Information Science at the Nara Institute of Science and Technology (NAIST). He received his Ph.D. in precision machinery engineering from the University of Tokyo in 1991. He has conducted research on a theory of intelligent computer-aided design systems, in particular experimental study and logical formalization of engineering design. He is also interested in multiagent architectures and ontologies for knowledge base systems.     

Information gathering and searching approaches on the Web

The information accessible through the Internet is increasing explosively as the Web is getting more and more widespread. In this situation, the Web is indispensable information resource for both of information gathering and information searching. Though traditional information retrieval techniques have been applied to information gathering and searching in the Web, they are insufficient for this new form of information source. Fortunately some Al techniques can be straightforwardly applicable to such tasks in the Web, and many researchers are trying this approach. In this paper, we attempt to describe the current state of information gathering and searching technologies in the Web, and the application of AI techniques in the fields. Then we point out limitations of these traditional and AI approaches and introduce two aapproaches: navigation planning and a Mondou search engine for overcoming them. The navigation planning system tries to collect systematic knowledge, rather than Web pages, which are only pieces of knowledge. The Mondou search engine copes with the problems of the query expansion/modification based on the techniques of text/web mining and information visualization. Seiji Yamada, Dr. Eng.: He received the B.S., M.S. and Ph.S. degrees in control engineering and artificial intelligence from Osaka University, Osaka, Japan, in 1984, 1986 and 1989, respectively. From 1989 to 1991, he served as a Research Associate in the Department of Control Engineering at Osaka University. From 1991 to 1996, he served as a Lecturer in the Institute of Scientific and Industrial Research at Osaka University. In 1996, he joined the Department of Computational Intelligence and Systems Science at Tokyo Institute of Technology, Yokohama, Japan, as an Associate Professor. His research interests include artificial intelligence, planning, machine learning for a robotics, intelligent information retrieval in the WWW, human computer interaction, He is a member of AAAI, IEEE, JSAI, RSJ and IEICE. Hiroyuki Kawano, Dr.Eng.: He is an Associate Professor at the Department of Systems Science, Graduate School of Informatics, Kyoto University, Japan. He obtained his B.Eng. and M.Eng. degrees in Applied Mathematics and Physics, and his Dr.Eng. degree in Applied Systems Science from Kyoto University. His research interests are in advanced database technologies, such as data mining, data warehousing, knowledge discovery and web search engine (Mondou). He has served on the program committees of several conferences in the areas of Data Base Systems, and technical committes of advanced information systems.     

Further investigation on the thermal decomposition of aluminium dialkylamides

Both dimethylamino- (I) and diethylamino-alane dimers (II), [(R₂N)₂AlH]₂ (I, R=Me; II, R=Et) decompose above 800° C under a few hundred Pa of hydrogen to dark greyish, hard (Vickers hardness larger than 2000), oxidation-resistive and oxidation-protective deposits which are tentatively identified as Al₅C₃N. The deposits are stable to moisture and diluted hydrochloric acid in contrast to those obtained below 800° C, but they easily dissolve in sodium hydroxide solution at room temperature evolving a gas. The deposits on a stainless steel substrate adhere strongly to the substrate and remain so on rapid heating and cooling. The electrical resistivity of the deposits is in the range 10² to 10⁴ cm.     

Efficient blind dereverberation and echo cancellation based on independent component analysis for actual acoustic signals

This letter presents a new algorithm for blind dereverberation and echo cancellation based on independent component analysis (ICA) for actual acoustic signals. We focus on frequency domain ICA (FD-ICA) because its computational cost and speed of learning convergence are sufficiently reasonable for practical applications such as hands-free speech recognition. In applying conventional FD-ICA as a preprocessing of automatic speech recognition in noisy environments, one of the most critical problems is how to cope with reverberations. To extract a clean signal from the reverberant observation, we model the separation process in the short-time Fourier transform domain and apply the multiple input/output inverse-filtering theorem (MINT) to the FD-ICA separation model. A naive implementation of this method is computationally expensive, because its time complexity is the second order of reverberation time. Therefore, the main issue in dereverberation is to reduce the high computational cost of ICA. In this letter, we reduce the computational complexity to the linear order of the reverberation time by using two techniques: (1) a separation model based on the independence of delayed observed signals with MINT and (2) spatial sphering for preprocessing. Experiments show that the computational cost grows in proportion to the linear order of the reverberation time and that our method improves the word correctness of automatic speech recognition by 10 to 20 points in a RT??= 670 ms reverberant environment.     

Effects of heated seats in vehicles on thermal comfort during the initial warm-up period

Eight subjects participated in a subjective experiment of eight conditions to investigate the effects of heated seats in vehicles on skin temperature, thermal sensation and thermal comfort during the initial warm-up period. The experimental conditions were designed as a combination of air temperature in the test room (5, 10, 15, or 20 °C) and heated seat (on/off). The heated seat was effective for improving thermal comfort during the initial warm-up period when air temperature was lower than 15 °C. Use of heated seats prevented decreases in or increased toe skin temperature. Heated seats also increased foot thermal sensation at 15 and 20 °C. Optimal thermal sensation in contact with the seat was higher when air temperature was lower. Optimal skin temperature in contact with the seat back was higher than that with the seat cushion. Moreover, these optimal skin temperatures were higher when air temperature was lower.     

Integrable multimode interference distributed Bragg reflector laser all-optical flip-flops

A novel fully integrable all-optical flip-flop has been created using distributed Bragg reflector multimode interference bistable laser diodes. The single metal-organic vapor phase epitaxy regrowth offset quantum well structure active/passive integrated flip-flop lases in single mode at 1554 nm and is compatible with standard fabrication methods of InP photonic integrated circuits (PICs). All-optical reset switching through cross-gain saturation was demonstrated over a 52-nm range, at 1522-1574 nm. An optical set has been achieved with -3 dBm and optical reset with less than -5-dBm external light injection. The flip-flop will be useful for integration in PICs of future photonic systems for self-routing and optical memories or buffers.     

A 63-mW H.264/MPEG-4 audio/visual codec LSI with module-wise dynamic Voltage/frequency scaling

A single-chip H.264 and MPEG-4 audio-visual LSI for mobile applications including terrestrial digital broadcasting system (ISDB-T / DVB-H) with a module-wise, dynamic voltage/frequency scaling architecture is presented for the first time. This LSI can keep operating even during the voltage/frequency transition, so there is no performance overhead. It is realized through a dynamic deskewing system and an on-chip voltage regulator with slew rate control. By the combination with traditional low power techniques such as embedded DRAM and clock gating, it consumes only 63 mW in decoding QVGA H.264 video at 15 frames/sec and MPEG-4 AAC LC audio simultaneously.     

高性能粉末冶金材料与其制造工艺的开发

随着常规粉末冶金产业发展，像锻轧粉末冶金铝合金、超塑性高温合金、机械合金化氧化物弥散强化（ODS）合金之类的先进粉末冶金工艺，引起了人们的注意。受先进粉末冶金工艺启发，开发了适应日本汽车与家电产业的新材料与生产工艺，开发了适合于钢铁结构的粉末冶金Al-Si-X合金。为了改进经济性与近终形（NNS）成形性，开发了无包套粉末挤压工艺与精压式粉末锻造工艺。基于这些工艺，还开发了比单一结构材料具有较多功能的新材料，诸如轻量散热器（heat sink）与高性能发动机气缸内衬。