Abstract In areas of activity where the notion accountability is strong, adoption artificial intelligence (AI) limited by opacity and lack understanding its behavior, all more so in embedded domain neural networks are compressed executed on microcontrollers. While NIST introduced 2021 several principles allowing AI explainability, this paper introduces a novel scheme, HistoTrust, combining secu...

s for the One Hundred Twentieth Meeting of the Acoustical Societyof America, Journal of the Acoustical Society of America, Vol. 88 (Suppl.1). Ellis, D. P. W. (1993). Hierarchic models of hearing for sound separation andreconstruction. 1993 IEEE Workshop on Applications of Signal Processingto Audio and Acoustics, IEEE, New York, pp. 157–160. Hunt, M. J., Lennig, M., & Mermelstein...

BACKGROUND Subitizing involves recognition mechanisms that allow effortless enumeration of up to four visual objects, however despite ample resolution experimental data suggest that only one pitch can be reliably enumerated. This may be due to the grouping of tones according to harmonic relationships by recognition mechanisms prior to fine pitch processing. Poorer frequency resolution of audito...

In Drosophila melanogaster, antennal hearing organs mediate the detection of conspecific songs. Combining laser Doppler vibrometry, acoustic near-field measurements and anatomical analysis, we have investigated the first steps in Drosophila audition, i.e. the conversion of acoustic energy into mechanical vibrations and the subsequent transmission of vibrations to the auditory receptors in the b...

We are seeking to enable humanoid robots to participate in live human-robot musical ensembles. In ensembles, the musicians must be able to listen to the audio that they and the others are producing in order to identify high-level features of the music (e.g., location in the score, tempo, cues). This information is crucial for the musicians so that they can know what to play next. One requiremen...