Mitigating Imprecise Timing in Spiking Neural Networks through Offset-Aware Training
Bibliographic record
Abstract
Neuromorphic accelerators, whether asynchronous meshes or globally-asynchronous, locally synchronous (GALS) fabrics, usually devote die area to barrier synchronizers, finely tuned clock trees, and metastability filters to keep spikes in lock-step. How much of that circuitry is truly indispensable? We quantify the intrinsic jitter tolerance of spiking neural networks (SNNs) and show how to push this tolerance further. We introduce a clock offset injector, a single GPU tensor op that time-shifts each core’s spikes by an arbitrary delay, and allows any PyTorch SNN to be stress-tested or hardened without code changes. Vision (N-MNIST) and audio (Spiking Heidelberg Digits) models, trained with and without our proposed offset-aware augmentation, are evaluated across nine skew levels (0-50% of the sample window) and up to eight cores. Baseline N-MNIST stays within ±1 percentage points (pp) until skew reaches 18 ms (6.1%). By accounting for temporally misaligned spikes during training time, we extend survival to 46 ms. Timing-critical SHD collapses beyond 23 ms (9%) but regains 6-9 pp after augmentation. These numbers map out safe envelopes for clock simplification: vision-only workloads can tolerate ≈ 3 ms per hop, while mixed vision-audio workloads must hold skew below 2 ms, or depend on timing-offset aware training. This ultimately shows how to enable lower-effort clocking and using software-level mitigation when needed.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".