|
A paper written by Electrical and Computer Engineering Ph.D. student Chenxi Wen and his advisor, ISR-affiliated Associate Professor Timothy Horiuchi (ECE), was the runner-up for the Best Paper Award at the IEEE Biomedical Circuits and Systems 2018 (BioCAS) conference this fall.
“Power-Law Compression Expands the Dynamic Range of a Neuromorphic Echolocation System” demonstrates the use of square-root and cube-root compression with a neuromorphic VLSI neuron to expand the range of distances over which interaural level differences (ILD) can be used to estimate echo direction in a sonar system based on echolocating bats.
Bats and other mammals use ILD to estimate the direction of high-frequency sounds. To compute the ILD of a sound, independent of overall loudness, excitatory and inhibitory synaptic conductances (encoding the left and right amplitudes) are hypothesized to compete in the neurons of the lateral superior olive. This neural model can also accept power-law compressed amplitudes that can allow a much larger range of input signal levels, a common issue in neural coding.
Although compression of incoming signals improves the input dynamic range of neural circuits, most sensory systems also have active mechanisms to increase the dynamic range further. In the echolocating bat, for example, the loudness of the outgoing pulse is modulated according to the distance of an attended object and the middle-ear muscle provides a time-dependent attenuation that can partially counteract the signal attenuation with range.
The ubiquity of logarithmic compression in biological systems is widely understood to manage the wide dynamic range of sensory signals and to exploit the computational advantages that logarithms provide, particularly for comparisons by subtraction mechanisms. When divisive mechanisms are brought to bear, however, it seems that power-law compression can be more appropriate.
Wen was able to attend the conference in Cleveland in part thanks to funding from the ISR Graduate Student Travel Award. This award, established in 2017, is given to deserving graduate students to help defray the travel costs of attending a conference to present their research. Since its inception, awards have been granted to 10 deserving graduate students.
Related Articles:
Sterbing-D'Angelo interviewed by Forbes magazine
Bats' touch sensor cells enable precision flight
Alumna Kirsten Bohn's bat song research is Science cover story
Oct. 9, 9 pm: Moss lab featured in National Geographic's "Brain Games"
Moss research uncovers bats' systematic 'active sensing' strategies
Heavy media coverage for bat wing hair research findings
Telluride newspaper writes about Neuromorphic Cognition Engineering Workshop
Tiny hairs on bats? wings act as speedometers
Moss, Horiuchi receive $1.5 million NSF grant for complex settings research
Alum Fumin Zhang elected to IEEE Fellow
November 15, 2018
|
