Micheli, Aurora, Gallicchio, Claudio, Tomen, Nergis · Oct 27, 2026
A spiking reservoir tuned to distinguish temporal frequencies
Markdown · JSON · Original source
Aurora Micheli, Claudio Gallicchio and Nergis Tomen introduce HRF-Res, a liquid state machine that replaces Leaky Integrate-and-Fire (LIF) neurons with Resonate-and-Fire (HRF) neurons. They argue that LIF neurons’ low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies. HRF neurons have intrinsic oscillatory dynamics and selectively respond to input frequencies, which the authors say lets HRF-Res capture frequency-specific temporal patterns. The authors report more discriminative representations and higher classification accuracy than LIF-based reservoirs, while retaining the energy efficiency of spiking computation. They report up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir. Across ten involving , sequential images, audio spike trains and neuromorphic vision, they report state-of-the-art accuracy among spiking reservoir methods. These are the paper’s reported results; the abstract does not specify benchmark-by-benchmark results or the conditions behind the energy comparison. How does HRF-Res’s accuracy and energy use vary across the ten benchmarks and comparison conditions?
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
cs.AI · Article
Source details
Original source date: Oct 27, 2026 (2026-10-27).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-11T08:39:06.391Z. First observed: 2026-10-30T06:42:43.061Z. Retrieved: 2026-10-31T05:46:41.656Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The supplied evidence is the abstract; it does not provide -by-benchmark results or describe the conditions behind the reported energy comparison. The abstract reports classification and energy results but does not establish performance in applications beyond the evaluated benchmarks.
Open question
How does HRF-Res’s accuracy and energy use vary across the ten and comparison conditions?
Source-stated
The authors introduce HRF-Res, a liquid state machine using HRF neurons in place of LIF neurons.
We introduce HRF-Res, a liquid state machine in which LIF neurons are replaced by a heterogeneous population of Harmonic Resonate-and-Fire (HRF) neurons.
Source-stated
The authors argue that LIF neurons’ low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies.
whose low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies.
Source-stated
The authors say HRF neurons’ intrinsic oscillatory dynamics and frequency-selective responses let the reservoir capture frequency-specific temporal patterns.
These neurons exhibit intrinsic oscillatory dynamics and selectively respond to input frequencies, enabling the reservoir to capture frequency-specific temporal patterns.
Source-stated
The authors report more discriminative representations and higher classification accuracy than LIF-based reservoirs, without sacrificing the energy efficiency of spiking computation.
The result is more discriminative representations and higher classification accuracy than LIF-based reservoirs, achieved without sacrificing the energy efficiency of spiking computation
Source-stated
The authors report up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir.
with up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir.
Source-stated
The authors report state-of-the-art accuracy among spiking reservoir methods across ten benchmarks spanning four application areas.
Evaluated on ten benchmarks spanning time series, sequential images, audio spike trains, and neuromorphic vision, HRF-Res achieves state-of-the-art accuracy among spiking reservoir methods.
Source and contributor context
Micheli, Aurora, Gallicchio, Claudio, Tomen, Nergis · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Time-series
Observations arranged over time, such as monthly sales; forecasting uses their patterns to estimate later values.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Harmonic
Harmonic develops systems that generate mathematical reasoning and proofs, including Aristotle.
Editorial provenance
Version 1. Publication ID: fa0557f5-911e-4695-81ea-d2ab95ffaf8f.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-11T08:39:06.391Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: c2dd0477771ffaedb562b9d566a3ea0f8c7de71a1b843cfe6047fc747151f28f. Source text hash: cd7e531f359b2a2f12e3879b671b24f69da4a4f32eb5ff0a48935db2e637d120.
Revision and source-check history
Version 1 · 2026-11-11T08:39:06.391Z
A spiking reservoir tuned to distinguish temporal frequencies
Aurora Micheli, Claudio Gallicchio and Nergis Tomen introduce HRF-Res, a liquid state machine that replaces Leaky Integrate-and-Fire (LIF) neurons with Resonate-and-Fire (HRF) neurons. They argue that LIF neurons’ low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies. HRF neurons have intrinsic oscillatory dynamics and selectively respond to input frequencies, which the authors say lets HRF-Res capture frequency-specific temporal patterns. The authors report more discriminative representations and higher classification accuracy than LIF-based reservoirs, while retaining the energy efficiency of spiking computation. They report up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir. Across ten involving , sequential images, audio spike trains and neuromorphic vision, they report state-of-the-art accuracy among spiking reservoir methods. These are the paper’s reported results; the abstract does not specify benchmark-by-benchmark results or the conditions behind the energy comparison. How does HRF-Res’s accuracy and energy use vary across the ten benchmarks and comparison conditions?
Original source
cs.AI · Article
Source details
Original source date: Oct 27, 2026 (2026-10-27).
Date basis: Date basis not provided.
Added to Escape Meridian: 2026-11-11T08:39:06.391Z. First observed: 2026-10-30T06:42:43.061Z. Retrieved: 2026-10-31T05:46:41.656Z.
Limits
The supplied evidence is the abstract; it does not provide -by-benchmark results or describe the conditions behind the reported energy comparison. The abstract reports classification and energy results but does not establish performance in applications beyond the evaluated benchmarks.
Open question
How does HRF-Res’s accuracy and energy use vary across the ten and comparison conditions?
Source-stated
The authors introduce HRF-Res, a liquid state machine using HRF neurons in place of LIF neurons.
We introduce HRF-Res, a liquid state machine in which LIF neurons are replaced by a heterogeneous population of Harmonic Resonate-and-Fire (HRF) neurons.
Source-stated
The authors argue that LIF neurons’ low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies.
whose low-pass dynamics limit their ability to distinguish information encoded at different temporal frequencies.
Source-stated
The authors say HRF neurons’ intrinsic oscillatory dynamics and frequency-selective responses let the reservoir capture frequency-specific temporal patterns.
These neurons exhibit intrinsic oscillatory dynamics and selectively respond to input frequencies, enabling the reservoir to capture frequency-specific temporal patterns.
Source-stated
The authors report more discriminative representations and higher classification accuracy than LIF-based reservoirs, without sacrificing the energy efficiency of spiking computation.
The result is more discriminative representations and higher classification accuracy than LIF-based reservoirs, achieved without sacrificing the energy efficiency of spiking computation
Source-stated
The authors report up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir.
with up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir.
Source-stated
The authors report state-of-the-art accuracy among spiking reservoir methods across ten benchmarks spanning four application areas.
Evaluated on ten benchmarks spanning time series, sequential images, audio spike trains, and neuromorphic vision, HRF-Res achieves state-of-the-art accuracy among spiking reservoir methods.
Source and contributor context
Micheli, Aurora, Gallicchio, Claudio, Tomen, Nergis · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
No suitable media was found by the supported page extraction.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Time-series
Observations arranged over time, such as monthly sales; forecasting uses their patterns to estimate later values.
arXiv
arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Harmonic
Harmonic develops systems that generate mathematical reasoning and proofs, including Aristotle.
Editorial provenance
Version 1. Publication ID: fa0557f5-911e-4695-81ea-d2ab95ffaf8f.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-11T08:39:06.391Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: material_development.
Observation: c2dd0477771ffaedb562b9d566a3ea0f8c7de71a1b843cfe6047fc747151f28f. Source text hash: cd7e531f359b2a2f12e3879b671b24f69da4a4f32eb5ff0a48935db2e637d120.
uncertain · 2026-11-11T08:39:06.391Z · anthropic/claude-opus-5.5
Publication: fa0557f5-911e-4695-81ea-d2ab95ffaf8f
attribution — pass: The paper is correctly attributed to Aurora Micheli, Claudio Gallicchio and Nergis Tomen. The draft frames the results as the authors' reported claims rather than established facts.
Authors:Aurora Micheli, Claudio Gallicchio, Nergis Tomen
summary — uncertain: The draft is faithful but hard for a non-specialist to follow. It leaves central terms unexplained, including spiking reservoir, liquid state machine, LIF, low-pass dynamics and neuromorphic vision. It also drops the abstract's opening rationale: spiking reservoirs pair cheap training with energy-saving computation that runs only when sparse spikes occur. Smallest repair: add one plain sentence drawn from that opening line and add tooltips for the terms. Separately, the question is repeated verbatim at the end of the summary, which is redundant but minor.
Spiking reservoir computing combines the training efficiency of fixed recurrent reservoirs with the energy efficiency of event-driven inference, where computation is triggered by sparse spikes rather than executed densely at every timestep.
date — pass: The originalPublished date of 2026-10-27 matches the v1 submission date. It is not the later feed date of 2026-10-31.
[Submitted on 27 Oct 2026]
support — pass: Every claim matches an abstract excerpt. The draft keeps the 'up to' qualifier on the 19x energy figure and the 'among spiking reservoir methods' scope on the state-of-the-art claim. The limitations correctly note that the abstract gives no per-benchmark detail. 'Four application areas' is a mild relabeling of four data types, not a contradiction.
with up to 19x lower energy consumption than an equivalent non-spiking oscillatory reservoir.
media — pass: The selected media is the verified arXiv abstract page for this paper. It is the direct original source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Harmonic develops systems that generate mathematical reasoning and proofs, including Aristotle.
Explore HarmonicA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceObservations arranged over time, such as monthly sales; forecasting uses their patterns to estimate later values.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceHarmonic develops systems that generate mathematical reasoning and proofs, including Aristotle.
Explore HarmonicA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceObservations arranged over time, such as monthly sales; forecasting uses their patterns to estimate later values.
Definition referencearXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.
Explore arXivA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition referenceA defined set of tests used to compare systems. Results depend on the tasks and measures selected.
Definition reference