Swarm music
Particle swarm optimization (PSO) is an approach to creating an intelligent computer algorithm for solving problems. It is based on the feeding patterns of a swarm of insects searching for food. However, with a little adaptation one might imagine a flock of seagulls searching for a tune. Indeed, researchers in China have used PSO to generate pentatonic music that follows the traditions of Chinese folk music with a technological twist. Their multi-melody space PSO algorithm (MSPA) searches for a solution to the problem of generating a pleasing melody within the confines of a given musical key and mode. The algorithm tests the “fitness” of the generated melody.
Zheng, X., Wang L., Li, D., Shen, L., Gao, Y., Guo, W. and Wang, Y. (2017) ‘Algorithm composition of Chinese folk music based on swarm intelligence‘, Int. J. Computing Science and Mathematics, Vol. 8, No. 5, pp.437-446.
Foul language
The vernacular of any society inevitably incorporates expletives and taboo words associated with bodily functions, body parts, the reproductive act and other areas of life. The online world of social media is no exception. However, as with the offline world of foul language, most societies prefer that children do not hear such words. Researchers from India are now working on a classification system that can detect taboo words in text conversations and might be used to check the content of such conversations for appropriateness without any undue invasion of the youngster’s privacy.
Kawate, S. and Patil, K. (2017) ‘Analysis of foul language usage in social media text conversation‘, Int. J. Social Media and Interactive Learning Environments, Vol. 5, No. 3, pp.227-251.
Music was my first love
What are you listening to? And, more to the point, what would you like to listen to next? Music recommendation system famously get it wrong when it comes to genre classification and make suggestions as to what the listener may wish to hear next. Now, researchers have developed a two-step approach to identifying the musical genre of a given audio file. They first use Voronoi audio similarity (VAS) and extract content-based features from the audio signal by splitting the song into time segments. Some clever mathematics based on volume and apparent tempo cross-checked against songs from a million-song database can be used to train the system. Tests on a subset selected the appropriate genre 4 out of every 5 times (78% accuracy).
Kalapatapu, P., Tejas, N.N., Dalmia, S., Gupta, P., Inguva, B., and Malapati, A. (2017) ‘A novel similarity measure: Voronoi audio similarity for genre classification‘ Int. J. Intell. Sys. Technol. Applicat., Vol. 16, No, 4, 309-318.
Is there no right or wrong?
Are quantitative techniques relevant to the development of financial accounting practice? That’s the question posed by researchers in Brazil. At the bottom line they ask, “Is it better to be roughly right or exactly wrong?” Their focus is the rules of International Financial Reporting Standards (IFRSs), which allow or request fair value measurement. Fundamentally, the answers to such questions will affect how past financial records affect one’s predictions of future financial performance. “Although financial statements are made of economic events that occurred in the past, part of these events are likely to affect the cash flows to and from the entity in the future. For this reason, there are two essential systems concerning timing in accounting, cash basis and accrual basis,” the team reports. They then offer a prospective answer to their question.
Flores, E. and Braunbeck, G.O. (2017) ‘What is better: to be roughly right or exactly wrong? The role of quantitative methods in financial accounting‘, Int. J. Multivariate Data Analysis, Vol. 1, No. 2, pp.162-172.