Silicon and physical substrates

Neuromorphic sensors

Emit asynchronous changes from a sensor rather than repeatedly transmitting complete sampled frames.

mixed signal siliconestablished research

Working definition

A neuromorphic sensor produces sparse, often asynchronous events when local measurements cross a defined change criterion. Event cameras are a prominent example, but auditory, tactile, and olfactory variants use different transduction and coding. Evaluation must retain threshold, noise, refractory behavior, timestamp precision, bandwidth, calibration, and downstream reconstruction or task costs.

Mechanism

  • Transduce a physical stimulus locally.
  • Compare change against a threshold or state.
  • Emit a timestamped event with sensor coordinates or polarity.

Measurements

  • Temporal resolution
  • Noise and event rate
  • Task accuracy and total pipeline energy

Reproducibility controls

  • Version hardware, software, firmware, and analysis code.
  • Declare dataset, preprocessing, random seeds, and measurement boundary.
  • Report repeated runs, variation, exclusions, and failed trials.

Limits and failure modes

  • Events are not lossless frames.
  • Sparse output depends on scene and thresholds.

Mathematical connection

Formal structure without substrate erasure

measurement methodTime-series models

Asynchronous event-stream model

Model timestamped sensor changes without forcing them into an undocumented frame representation.

Inputs

  • Event timestamps
  • Coordinates and polarity
  • Calibration data

Outputs

  • Rate and timing features
  • Noise estimate
  • Task-ready representation

Limit: Any binning or reconstruction changes the representation and may erase the temporal advantage being evaluated.

Technical and governance sources

  1. [1]NeuroBench: Advancing Neuromorphic Computing Through Collaborative, Fair and Representative Benchmarking · National Institute of Standards and Technology

    Establishes: A community framework separating algorithm and system tracks and defining task, correctness, efficiency, and reporting procedures intended to make neuromorphic results more comparable and reproducible.

    Boundary: A benchmark ranks submitted systems on declared tasks and metrics. It does not prove general intelligence, biological equivalence, safety, usefulness outside the benchmark, or superiority under unreported host and data costs.

Related concepts