An Efficient Hybrid Music Recommender System Using an Incrementally Trainable Probabilistic Generative Model

This paper presents a hybrid music recommender system that ranks musical pieces while efficiently maintaining collaborative and content-based data, i.e., rating scores given by users and acoustic features of audio signals. This hybrid approach overcomes the conventional tradeoff between recommendation accuracy and variety of recommended artists. Collaborative filtering, which is used on e-commerce sites, cannot recommend nonbrated pieces and provides a narrow variety of artists. Content-based filtering does not have satisfactory accuracy because it is based on the heuristics that the user's favorite pieces will have similar musical content despite there being exceptions. To attain a higher recommendation accuracy along with a wider variety of artists, we use a probabilistic generative model that unifies the collaborative and content-based data in a principled way. This model can explain the generative mechanism of the observed data in the probability theory. The probability distribution over users, pieces, and features is decomposed into three conditionally independent ones by introducing latent variables. This decomposition enables us to efficiently and incrementally adapt the model for increasing numbers of users and rating scores. We evaluated our system by using audio signals of commercial CDs and their corresponding rating scores obtained from an e-commerce site. The results revealed that our system accurately recommended pieces including nonrated ones from a wide variety of artists and maintained a high degree of accuracy even when new users and rating scores were added.     

Pitch-Dependent Identification of Musical Instrument Sounds

This paper describes a musical instrument identification method that takes into consideration the pitch dependency of timbres of musical instruments. The difficulty in musical instrument identification resides in the pitch dependency of musical instrument sounds, that is, acoustic features of most musical instruments vary according to the pitch (fundamental frequency, F0). To cope with this difficulty, we propose an F0-dependent multivariate normal distribution, where each element of the mean vector is represented by a function of F0. Our method first extracts 129 features (e.g., the spectral centroid, the gradient of the straight line approximating the power envelope) from a musical instrument sound and then reduces the dimensionality of the feature space into 18 dimension. In the 18-dimensional feature space, it calculates an F0-dependent mean function and an F0-normalized covariance, and finally applies the Bayes decision rule. Experimental results of identifying 6,247 solo tones of 19 musical instruments shows that the proposed method improved the recognition rate from 75.73% to 79.73%. This research was partially supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Grant-in-Aid for Scientific Research (A), No.15200015, and Informatics Research Center for Development of Knowledge Society Infrastructure (COE program of MEXT, Japan). Tetsuro Kitahara received the B.S. from Tokyo University of Science in 2002 and the M.S. from Kyoto University in 2004. He is currently a Ph.D. course student at Graduate School of Informatics, Kyoto University. Since 2005, he has been a Research Fellow of the Japan Society for the Promotion of Science. His research interests include music informatics. He recieved IPSJ 65th National Convention Student Award in 2003, IPSJ 66th National Convention Student Award and TELECOM System Technology Award for Student in 2004, and IPSJ 67th National Convention Best Paper Award for Young Researcher in 2005. He is a student member of IPSJ, IEICE, JSAI, ASJ, and JSMPC. Masataka Goto received his Doctor of Engineering degree in Electronics, Information and Communication Engineering from Waseda University, Japan, in 1998. He then joined the Electrotechnical Laboratory (ETL; reorganized as the National Institute of Advanced Industrial Science and Technology (AIST) in 2001), where he has been engaged as a researcher ever since. He served concurrently as a researcher in Precursory Research for Embryonic Science and Technology (PRESTO), Japan Science and Technology Corporation (JST) from 2000 to 2003, and an associate professor of the Department of Intelligent Interaction Technologies, Graduate School of Systems and Information Engineering, University of Tsukuba since 2005. His research interests include music information processing and spoken language processing. Dr. Goto received seventeen awards including the IPSJ Best Paper Award and IPSJ Yamashita SIG Research Awards (MUS and SLP) from the Information Processing Society of Japan (IPSJ), Awaya Prize for Outstanding Presentation and Award for Outstanding Poster Presentation from the Acoustical Society of Japan (ASJ), Award for Best Presentation from the Japanese Society for Music Perception and Cognition (JSMPC), WISS 2000 Best Paper Award and Best Presentation Award, and Interaction 2003 Best Paper Award. He is a member of the IPSJ, ASJ, JSMPC, Institute of Electronics, Information and Communication Engineers (IEICE), and International Speech Communication Association (ISCA). Hiroshi G. Okuno received the B.A. and Ph.D from the University of Tokyo in 1972 and 1996, respectively. He worked for Nippon Telegraph and Telephone, Kitano Symbiotic Systems Project, and Tokyo University of Science. He is currently a professor at the Department of Intelligence Technology and Science, Graduate School of Informatics, Kyoto University. He was a visiting scholar at Stanford University, and a visiting associate professor at the University of Tokyo. He has done research in programming languages, parallel processing, and reasoning mechanism in AI, and he is currently engaged in computational auditory scene analysis, music scene analysis and robot audition. He received the best paper awards from the Japanese Society for Artificial Intelligence and the International Society for Applied Intelligence, in 1991 and 2001, respectively. He edited with David Rosenthal “Computational Auditory Scene Analysis” from Lawrence Erlbaum Associates in 1998 and with Taiichi Yuasa “Advanced Lisp Technology” from Taylor and Francis Inc. in 2002. He is a member of IPSJ, JSAI, JSSST, JSCS, ACM, AAAI, ASA, and IEEE.     

Time-resolved diffuse optical tomographic imaging for the provision of both anatomical and functional information about biological tissue

We present in vivo images of near-infrared (NIR) diffuse optical tomography (DOT) of human lower legs and forearm to validate the dual functions of a time-resolved (TR) NIR DOT in clinical diagnosis, i.e., to provide anatomical and functional information simultaneously. The NIR DOT system is composed of time-correlated single-photon-counting channels, and the image reconstruction algorithm is based on the modified generalized pulsed spectral technique, which effectively incorporates the TR data with reasonable computation time. The reconstructed scattering images of both the lower legs and the forearm revealed their anatomies, in which the bones were clearly distinguished from the muscles. In the absorption images, some of the blood vessels were observable. In the functional imaging, a subject was requested to do handgripping exercise to stimulate physiological changes in the forearm tissue. The images of oxyhemoglobin, deoxyhemoglobin, and total hemoglobin concentration changes in the forearm were obtained from the differential images of the absorption at three wavelengths between the exercise and the rest states, which were reconstructed with a differential imaging scheme. These images showed increases in both blood volume and oxyhemoglobin concentration in the arteries and simultaneously showed hypoxia in the corresponding muscles. All the results have demonstrated the capability of TR NIR DOT by reconstruction of the absolute images of the scattering and the absorption with a high spatial resolution that finally provided both the anatomical and functional information inside bulky biological tissues.     

Birefringence control of silica waveguides on Si and itsapplication to a polarization-beam splitter/switch

A novel polarization-beam splitter/switch with a Mach-Zehnder interferometer configuration was fabricated using a silica-based planar lightwave circuit (PLC) on a silicon substrate. The polarization-beam splitter/switch was realized by accurately controlling the waveguide birefringence and the phase state by laser trimming two kinds of stress-applying amorphous silicon film with widths of 50 μm and 95 μm. Fiber-waveguide-fiber insertion loss of 0.5 dB and crosstalk attenuations of over 25.6 dB were attained     

Genotoxicity of nano/microparticles in in vitro micronuclei, in vivo comet and mutation assay systems

Background

Recently, manufactured nano/microparticles such as fullerenes (C₆₀), carbon black (CB) and ceramic fiber are being widely used because of their desirable properties in industrial, medical and cosmetic fields. However, there are few data on these particles in mammalian mutagenesis and carcinogenesis. To examine genotoxic effects by C₆₀, CB and kaolin, an in vitro micronuclei (MN) test was conducted with human lung cancer cell line, A549 cells. In addition, DNA damage and mutations were analyzed by in vivo assay systems using male C57BL/6J or gpt delta transgenic mice which were intratracheally instilled with single or multiple doses of 0.2 mg per animal of particles.

Results

In in vitro genotoxic analysis, increased MN frequencies were observed in A549 cells treated with C₆₀, CB and kaolin in a dose-dependent manner. These three nano/microparticles also induced DNA damage in the lungs of C57BL/6J mice measured by comet assay. Moreover, single or multiple instillations of C₆₀ and kaolin, increased either or both of gpt and Spi^- mutant frequencies in the lungs of gpt delta transgenic mice. Mutation spectra analysis showed transversions were predominant, and more than 60% of the base substitutions occurred at G:C base pairs in the gpt genes. The G:C to C:G transversion was commonly increased by these particle instillations.

Conclusion

Manufactured nano/microparticles, CB, C₆₀ and kaolin, were shown to be genotoxic in in vitro and in vivo assay systems.     

Sample preparation of GaN-based materials on a sapphire substrate for STEM analysis

In this work, a detailed TEM sample preparation recipe based on a wedge polishing technique for GaN-based materials is presented. The obtained samples have atomically flat surfaces without any obvious surface damages such as the formation of amorphous layers. A composition estimation of Al(x)Ga(1-x)N from Z-contrast STEM imaging is carried out using these samples. The results are in good accord with the nominal composition.     

Contribution of Invariant Natural Killer T Cells to the Clearance of Pseudomonas aeruginosa from Skin Wounds

Chronic infections are considered one of the most severe problems in skin wounds, and bacteria are present in over 90% of chronic wounds. Pseudomonas aeruginosa is frequently isolated from chronic wounds and is thought to be a cause of delayed wound healing. Invariant natural killer T (iNKT) cells, unique lymphocytes with a potent regulatory ability in various inflammatory responses, accelerate the wound healing process. In the present study, we investigated the contribution of iNKT cells in the host defense against P. aeruginosa inoculation at the wound sites. We analyzed the re-epithelialization, bacterial load, accumulation of leukocytes, and production of cytokines and antimicrobial peptides. In iNKT cell–deficient (Jα18KO) mice, re-epithelialization was significantly decreased, and the number of live colonies was significantly increased, when compared with those in wild-type (WT) mice on day 7. IL-17A, and IL-22 production was significantly lower in Jα18KO mice than in WT mice on day 5. Furthermore, the administration of α-galactosylceramide (α-GalCer), a specific activator of iNKT cells, led to enhanced host protection, as shown by reduced bacterial load, and to increased production of IL-22, IL-23, and S100A9 compared that of with WT mice. These results suggest that iNKT cells promote P. aeruginosa clearance during skin wound healing.     

ZnO Nanoplatelets with Controlled Thickness: Atomic Insight into Facet-Specific Bimodal Ligand Binding Using DNP NMR

Colloidal nanoplatelets (NPLs) and nanosheets with controlled thickness have recently emerged as an exciting new class of quantum-sized nanomaterials with substantially distinct optical properties compared to 0D quantum dots. Zn-based NPLs are an attractive heavy-metal-free alternative to the so far most widespread cadmium chalcogenide colloidal 2D semiconductor nanostructures, but their synthesis remains challenging to achieve. The authors describe herein, to the best of their knowledge, the first synthesis of highly stable ZnO NPLs with the atomically precise thickness, which for the smallest NPLs is 3.2 nm (corresponding to 12 ZnO layers). Furthermore, by means of dynamic nuclear polarization-enhanced solid-state ¹⁵N NMR, the original role of the benzamidine ligands in stabilizing the surface of these nanomaterials is revealed, which can bind to both the polar and non-polar ZnO facets, acting either as X- or L-type ligands, respectively. This bimodal stabilization allows obtaining hexagonal NPLs for which the surface energy of the facets is modulated by the presence of the ligands. Thus, in-depth study of the interactions at the organic–inorganic interfaces provides a deeper understanding of the ligand–surface interface and should facilitate the future chemistry of stable-by-design nano-objects.     

Design,Synthesis, and Biological Evaluation of Lysine Demethylase 5 C Degraders

Lysine demethylase 5 C (KDM5C) controls epigenetic gene expression and is attracting great interest in the field of chemical epigenetics. KDM5C has emerged as a therapeutic target for anti-prostate cancer agents, and recently we identified triazole 1 as an inhibitor of KDM5C. Compound 1 exhibited highly potent KDM5C-inhibitory activity in in vitro enzyme assays, but did not show strong anticancer effects. Therefore, a different approach is needed for the development of anticancer agents targeting KDM5C. Here, we attempted to identify KDM5C degraders by focusing on a protein-knockdown strategy. Compound 3 b , which was designed based on compound 1 , degraded KDM5C and inhibited the growth of prostate cancer PC-3 cells more strongly than compound 1 . These findings suggest that KDM5C degraders are more effective as anticancer agents than compounds that only inhibit the catalytic activity of KDM5C.